Nyaaori/rockbox
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
rockbox/lib/rbcodec/dsp/compressor.c

615 lines
20 KiB
C
Raw Normal View History

Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
/***************************************************************************
* __________ __ ___.
* Open \______ \ ____ ____ | | _\_ |__ _______ ___
* Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ /
* Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < <
* Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \
* \/ \/ \/ \/ \/
* $Id$
*
* Copyright (C) 2009 Jeffrey Goode
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
* KIND, either express or implied.
*
****************************************************************************/
Remove explicit config.h and system.h includes from DSP code. Replace with rbcodecconfig.h and platform.h includes. Remove now- unneeded ones as well. Change-Id: I6111b71e90bf86d9fe272a7916f2d34a5c6dd724
2013-05-27 07:23:33 +00:00
#include "rbcodecconfig.h"
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
#include "fixedpoint.h"
#include "fracmul.h"
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
#include <string.h>
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/* Define LOGF_ENABLE to enable logf output in this file
* #define LOGF_ENABLE
*/
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
#include "logf.h"
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
#include "dsp_proc_entry.h"
Make rbcodec/dsp includes more specific. Change-Id: Idb6af40df26f5b8499a40e8b98602261ef227044
2012-04-29 21:31:30 +00:00
#include "compressor.h"
Enable setting of global output samplerate on certain targets. Replaces the NATIVE_FREQUENCY constant with a configurable frequency. The user may select 48000Hz if the hardware supports it. The default is still 44100Hz and the minimum is 44100Hz. The setting is located in the playback settings, under "Frequency". "Frequency" was duplicated in english.lang for now to avoid having to fix every .lang file for the moment and throwing everything out of sync because of the new play_frequency feature in features.txt. The next cleanup should combine it with the one included for recording and generalize the ID label. If the hardware doesn't support 48000Hz, no setting will be available. On particular hardware where very high rates are practical and desireable, the upper bound can be extended by patching. The PCM mixer can be configured to play at the full hardware frequency range. The DSP core can configure to the hardware minimum up to the maximum playback setting (some buffers must be reserved according to the maximum rate). If only 44100Hz is supported or possible on a given target for playback, using the DSP and mixer at other samperates is possible if the hardware offers them. Change-Id: I6023cf0c0baa8bc6292b6919b4dd3618a6a25622 Reviewed-on: http://gerrit.rockbox.org/479 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-05-23 17:58:51 +00:00
#include "dsp_misc.h"
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
#define UNITY (1L << 24) /* unity gain in S7.24 format */
#define MAX_DLY 960 /* Max number of samples to delay
output (960 = 5ms @ 192 kHz)
*/
#define MAX_CH 4 /* Is there a good malloc() or equal
for rockbox?
*/
#define DLY_TIME 3 /* milliseconds */
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
static struct compressor_settings curr_set; /* Cached settings */
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
static int32_t comp_makeup_gain IBSS_ATTR; /* S7.24 format */
static int32_t comp_curve[66] IBSS_ATTR; /* S7.24 format */
static int32_t release_gain IBSS_ATTR; /* S7.24 format */
static int32_t release_holdoff IBSS_ATTR; /* S7.24 format */
/* 1-pole filter coefficients for exponential attack/release times */
static int32_t rlsca IBSS_ATTR; /* Release 'alpha' */
static int32_t rlscb IBSS_ATTR; /* Release 'beta' */
static int32_t attca IBSS_ATTR; /* Attack 'alpha' */
static int32_t attcb IBSS_ATTR; /* Attack 'beta' */
static int32_t limitca IBSS_ATTR; /* Limiter Attack 'alpha' */
/* 1-pole filter coefficients for sidechain pre-emphasis filters */
static int32_t hp1ca IBSS_ATTR; /* hpf1 'alpha' */
static int32_t hp2ca IBSS_ATTR; /* hpf2 'beta' */
/* 1-pole hp filter state variables for pre-emphasis filters */
static int32_t hpfx1 IBSS_ATTR; /* hpf1 and hpf2 x[n-1] */
static int32_t hp1y1 IBSS_ATTR; /* hpf2 y[n-1] */
static int32_t hp2y1 IBSS_ATTR; /* hpf2 y[n-1] */
/* Delay Line for look-ahead compression */
static int32_t labuf[MAX_CH][MAX_DLY]; /* look-ahead buffer */
static int32_t delay_time;
static int32_t delay_write;
static int32_t delay_read;
/** 1-Pole LP Filter first coefficient computation
* Returns S7.24 format integer used for "a" coefficient
* rc: "RC Time Constant", or time to decay to 1/e
* fs: Sampling Rate
* Interpret attack and release time as an RC time constant
* (time to decay to 1/e)
* 1-pole filters use approximation
* a0 = 1/(fs*rc + 1)
* b1 = 1.0 - a0
* fs = Sampling Rate
* rc = Time to decay to 1/e
* y[n] = a0*x[n] + b1*y[n-1]
*
* According to simulation on Intel hardware
* this algorithm produces < 2% error for rc < ~100ms
* For rc 100ms - 1000ms, error approaches 0%
* For compressor attack/release times, this is more than adequate.
*
* Error was measured against the more rigorous computation:
* a0 = 1.0 - e^(-1.0/(fs*rc))
*/
int32_t get_lpf_coeff(int32_t rc, int32_t fs, int32_t rc_units)
{
int32_t c = fs*rc;
c /= rc_units;
c += 1;
c = UNITY/c;
return c;
}
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** Coefficients to get 10dB change per time period "rc"
* from 1-pole LP filter topology
* This function is better used to match behavior of
* linear release which was implemented prior to implementation
* of exponential attack/release function
*/
int32_t get_att_rls_coeff(int32_t rc, int32_t fs)
{
int32_t c = UNITY/fs;
c *= 1152; /* 1000 * 10/( 20*log10( 1/e ) ) */
c /= rc;
return c;
}
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
/** COMPRESSOR UPDATE
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
* Called via the menu system to configure the compressor process
*/
Enable setting of global output samplerate on certain targets. Replaces the NATIVE_FREQUENCY constant with a configurable frequency. The user may select 48000Hz if the hardware supports it. The default is still 44100Hz and the minimum is 44100Hz. The setting is located in the playback settings, under "Frequency". "Frequency" was duplicated in english.lang for now to avoid having to fix every .lang file for the moment and throwing everything out of sync because of the new play_frequency feature in features.txt. The next cleanup should combine it with the one included for recording and generalize the ID label. If the hardware doesn't support 48000Hz, no setting will be available. On particular hardware where very high rates are practical and desireable, the upper bound can be extended by patching. The PCM mixer can be configured to play at the full hardware frequency range. The DSP core can configure to the hardware minimum up to the maximum playback setting (some buffers must be reserved according to the maximum rate). If only 44100Hz is supported or possible on a given target for playback, using the DSP and mixer at other samperates is possible if the hardware offers them. Change-Id: I6023cf0c0baa8bc6292b6919b4dd3618a6a25622 Reviewed-on: http://gerrit.rockbox.org/479 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-05-23 17:58:51 +00:00
static bool compressor_update(struct dsp_config *dsp,
const struct compressor_settings *settings)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
/* make settings values useful */
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
int threshold = settings->threshold;
bool auto_gain = settings->makeup_gain == 1;
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
static const int comp_ratios[] = { 2, 4, 6, 10, 0 };
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
int ratio = comp_ratios[settings->ratio];
bool soft_knee = settings->knee == 1;
int32_t release = settings->release_time;
int32_t attack = settings->attack_time;
/* Compute Attack and Release Coefficients */
int32_t fs = dsp_get_output_frequency(dsp);
/* Release */
rlsca = get_att_rls_coeff(release, fs);
rlscb = UNITY - rlsca ;
/* Attack */
if(attack > 0)
{
attca = get_att_rls_coeff(attack, fs);
attcb = UNITY - attca ;
}
else {
attca = UNITY;
attcb = 0;
}
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/* Sidechain pre-emphasis filter coefficients */
hp1ca = fs + 0x003C1; /** The "magic" constant is 1/RC. This filter
* cut-off is approximately 237 Hz
*/
hp1ca = UNITY/hp1ca;
hp1ca *= fs;
hp2ca = fs + 0x02065; /* The "magic" constant is 1/RC. This filter
* cut-off is approximately 2.18 kHz
*/
hp2ca = UNITY/hp2ca;
hp2ca *= fs;
bool changed = settings == &curr_set; /* If frequency changes */
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
bool active = threshold < 0;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
if (memcmp(settings, &curr_set, sizeof (curr_set)))
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
/* Compressor settings have changed since last call */
changed = true;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
#if defined(ROCKBOX_HAS_LOGF) && defined(LOGF_ENABLE)
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
if (settings->threshold != curr_set.threshold)
{
logf(" Compressor Threshold: %d dB\tEnabled: %s",
threshold, active ? "Yes" : "No");
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
if (settings->makeup_gain != curr_set.makeup_gain)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
logf(" Compressor Makeup Gain: %s",
auto_gain ? "Auto" : "Off");
}
iBasso DX50/DX90: Major code cleanup and reorganization. Reorganization - Separated iBasso devices from PLATFORM_ANDROID. These are now standlone hosted targets. Most device specific code is in the firmware/target/hosted/ibasso directory. - No dependency on Android SDK, only the Android NDK is needed. 32 bit Android NDK and Android API Level 16. - Separate implementation for each device where feasible. Code cleanup - Rewrite of existing code, from simple reformat to complete reimplementation. - New backlight interface, seperating backlight from touchscreen. - Rewrite of device button handler, removing unneeded code and fixing memory leaks. - New Debug messages interface logging to Android adb logcat (DEBUGF, panicf, logf). - Rewrite of lcd device handler, removing unneeded code and fixing memory leaks. - Rewrite of audiohw device handler/pcm interface, removing unneeded code and fixing memory leaks, enabling 44.1/48kHz pthreaded playback. - Rewrite of power and powermng, proper shutdown, using batterylog results (see http://gerrit.rockbox.org/r/#/c/1047/). - Rewrite of configure (Android NDK) and device specific config. - Rewrite of the Android NDK specific Makefile. Misc - All plugins/games/demos activated. - Update tinyalsa to latest from https://github.com/tinyalsa/tinyalsa. Includes - http://gerrit.rockbox.org/r/#/c/993/ - http://gerrit.rockbox.org/r/#/c/1010/ - http://gerrit.rockbox.org/r/#/c/1035/ Does not include http://gerrit.rockbox.org/r/#/c/1007/ due to new backlight interface and new option for hold switch, touchscreen, physical button interaction. Rockbox needs the iBasso DX50/DX90 loader for startup, see http://gerrit.rockbox.org/r/#/c/1099/ The loader expects Rockbox to be installed in /mnt/sdcard/.rockbox/. If /mnt/sdcard/ is accessed as USB mass storage device, Rockbox will exit gracefully and the loader will restart Rockbox on USB disconnect. Tested on iBasso DX50. Compiled (not tested) for iBasso DX90. Compiled (not tested) for PLATFORM_ANDROID. Change-Id: I5f5e22e68f5b4cf29c28e2b40b2c265f2beb7ab7
2015-02-02 20:44:29 +00:00
if (settings->ratio != curr_set.ratio)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
if (ratio)
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
{ logf(" Compressor Ratio: %d:1", ratio); }
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
else
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
{ logf(" Compressor Ratio: Limit"); }
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
iBasso DX50/DX90: Major code cleanup and reorganization. Reorganization - Separated iBasso devices from PLATFORM_ANDROID. These are now standlone hosted targets. Most device specific code is in the firmware/target/hosted/ibasso directory. - No dependency on Android SDK, only the Android NDK is needed. 32 bit Android NDK and Android API Level 16. - Separate implementation for each device where feasible. Code cleanup - Rewrite of existing code, from simple reformat to complete reimplementation. - New backlight interface, seperating backlight from touchscreen. - Rewrite of device button handler, removing unneeded code and fixing memory leaks. - New Debug messages interface logging to Android adb logcat (DEBUGF, panicf, logf). - Rewrite of lcd device handler, removing unneeded code and fixing memory leaks. - Rewrite of audiohw device handler/pcm interface, removing unneeded code and fixing memory leaks, enabling 44.1/48kHz pthreaded playback. - Rewrite of power and powermng, proper shutdown, using batterylog results (see http://gerrit.rockbox.org/r/#/c/1047/). - Rewrite of configure (Android NDK) and device specific config. - Rewrite of the Android NDK specific Makefile. Misc - All plugins/games/demos activated. - Update tinyalsa to latest from https://github.com/tinyalsa/tinyalsa. Includes - http://gerrit.rockbox.org/r/#/c/993/ - http://gerrit.rockbox.org/r/#/c/1010/ - http://gerrit.rockbox.org/r/#/c/1035/ Does not include http://gerrit.rockbox.org/r/#/c/1007/ due to new backlight interface and new option for hold switch, touchscreen, physical button interaction. Rockbox needs the iBasso DX50/DX90 loader for startup, see http://gerrit.rockbox.org/r/#/c/1099/ The loader expects Rockbox to be installed in /mnt/sdcard/.rockbox/. If /mnt/sdcard/ is accessed as USB mass storage device, Rockbox will exit gracefully and the loader will restart Rockbox on USB disconnect. Tested on iBasso DX50. Compiled (not tested) for iBasso DX90. Compiled (not tested) for PLATFORM_ANDROID. Change-Id: I5f5e22e68f5b4cf29c28e2b40b2c265f2beb7ab7
2015-02-02 20:44:29 +00:00
if (settings->knee != curr_set.knee)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
logf(" Compressor Knee: %s", soft_knee?"Soft":"Hard");
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
iBasso DX50/DX90: Major code cleanup and reorganization. Reorganization - Separated iBasso devices from PLATFORM_ANDROID. These are now standlone hosted targets. Most device specific code is in the firmware/target/hosted/ibasso directory. - No dependency on Android SDK, only the Android NDK is needed. 32 bit Android NDK and Android API Level 16. - Separate implementation for each device where feasible. Code cleanup - Rewrite of existing code, from simple reformat to complete reimplementation. - New backlight interface, seperating backlight from touchscreen. - Rewrite of device button handler, removing unneeded code and fixing memory leaks. - New Debug messages interface logging to Android adb logcat (DEBUGF, panicf, logf). - Rewrite of lcd device handler, removing unneeded code and fixing memory leaks. - Rewrite of audiohw device handler/pcm interface, removing unneeded code and fixing memory leaks, enabling 44.1/48kHz pthreaded playback. - Rewrite of power and powermng, proper shutdown, using batterylog results (see http://gerrit.rockbox.org/r/#/c/1047/). - Rewrite of configure (Android NDK) and device specific config. - Rewrite of the Android NDK specific Makefile. Misc - All plugins/games/demos activated. - Update tinyalsa to latest from https://github.com/tinyalsa/tinyalsa. Includes - http://gerrit.rockbox.org/r/#/c/993/ - http://gerrit.rockbox.org/r/#/c/1010/ - http://gerrit.rockbox.org/r/#/c/1035/ Does not include http://gerrit.rockbox.org/r/#/c/1007/ due to new backlight interface and new option for hold switch, touchscreen, physical button interaction. Rockbox needs the iBasso DX50/DX90 loader for startup, see http://gerrit.rockbox.org/r/#/c/1099/ The loader expects Rockbox to be installed in /mnt/sdcard/.rockbox/. If /mnt/sdcard/ is accessed as USB mass storage device, Rockbox will exit gracefully and the loader will restart Rockbox on USB disconnect. Tested on iBasso DX50. Compiled (not tested) for iBasso DX90. Compiled (not tested) for PLATFORM_ANDROID. Change-Id: I5f5e22e68f5b4cf29c28e2b40b2c265f2beb7ab7
2015-02-02 20:44:29 +00:00
if (settings->release_time != curr_set.release_time)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
logf(" Compressor Release: %d", release);
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
iBasso DX50/DX90: Major code cleanup and reorganization. Reorganization - Separated iBasso devices from PLATFORM_ANDROID. These are now standlone hosted targets. Most device specific code is in the firmware/target/hosted/ibasso directory. - No dependency on Android SDK, only the Android NDK is needed. 32 bit Android NDK and Android API Level 16. - Separate implementation for each device where feasible. Code cleanup - Rewrite of existing code, from simple reformat to complete reimplementation. - New backlight interface, seperating backlight from touchscreen. - Rewrite of device button handler, removing unneeded code and fixing memory leaks. - New Debug messages interface logging to Android adb logcat (DEBUGF, panicf, logf). - Rewrite of lcd device handler, removing unneeded code and fixing memory leaks. - Rewrite of audiohw device handler/pcm interface, removing unneeded code and fixing memory leaks, enabling 44.1/48kHz pthreaded playback. - Rewrite of power and powermng, proper shutdown, using batterylog results (see http://gerrit.rockbox.org/r/#/c/1047/). - Rewrite of configure (Android NDK) and device specific config. - Rewrite of the Android NDK specific Makefile. Misc - All plugins/games/demos activated. - Update tinyalsa to latest from https://github.com/tinyalsa/tinyalsa. Includes - http://gerrit.rockbox.org/r/#/c/993/ - http://gerrit.rockbox.org/r/#/c/1010/ - http://gerrit.rockbox.org/r/#/c/1035/ Does not include http://gerrit.rockbox.org/r/#/c/1007/ due to new backlight interface and new option for hold switch, touchscreen, physical button interaction. Rockbox needs the iBasso DX50/DX90 loader for startup, see http://gerrit.rockbox.org/r/#/c/1099/ The loader expects Rockbox to be installed in /mnt/sdcard/.rockbox/. If /mnt/sdcard/ is accessed as USB mass storage device, Rockbox will exit gracefully and the loader will restart Rockbox on USB disconnect. Tested on iBasso DX50. Compiled (not tested) for iBasso DX90. Compiled (not tested) for PLATFORM_ANDROID. Change-Id: I5f5e22e68f5b4cf29c28e2b40b2c265f2beb7ab7
2015-02-02 20:44:29 +00:00
if (settings->attack_time != curr_set.attack_time)
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
{
logf(" Compressor Attack: %d", attack);
}
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
#endif
curr_set = *settings;
}
if (!changed || !active)
return active;
/* configure variables for compressor operation */
static const int32_t db[] = {
/* positive db equivalents in S15.16 format */
0x000000, 0x241FA4, 0x1E1A5E, 0x1A94C8,
0x181518, 0x1624EA, 0x148F82, 0x1338BD,
0x120FD2, 0x1109EB, 0x101FA4, 0x0F4BB6,
0x0E8A3C, 0x0DD840, 0x0D3377, 0x0C9A0E,
0x0C0A8C, 0x0B83BE, 0x0B04A5, 0x0A8C6C,
0x0A1A5E, 0x09ADE1, 0x094670, 0x08E398,
0x0884F6, 0x082A30, 0x07D2FA, 0x077F0F,
0x072E31, 0x06E02A, 0x0694C8, 0x064BDF,
0x060546, 0x05C0DA, 0x057E78, 0x053E03,
0x04FF5F, 0x04C273, 0x048726, 0x044D64,
0x041518, 0x03DE30, 0x03A89B, 0x037448,
0x03412A, 0x030F32, 0x02DE52, 0x02AE80,
0x027FB0, 0x0251D6, 0x0224EA, 0x01F8E2,
0x01CDB4, 0x01A359, 0x0179C9, 0x0150FC,
0x0128EB, 0x010190, 0x00DAE4, 0x00B4E1,
0x008F82, 0x006AC1, 0x004699, 0x002305};
struct curve_point
{
int32_t db; /* S15.16 format */
int32_t offset; /* S15.16 format */
} db_curve[5];
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** Set up the shape of the compression curve first as decibel values
* db_curve[0] = bottom of knee
* [1] = threshold
* [2] = top of knee
* [3] = 0 db input
* [4] = ~+12db input (2 bits clipping overhead)
*/
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
db_curve[1].db = threshold << 16;
if (soft_knee)
{
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/* bottom of knee is 3dB below the threshold for soft knee */
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
db_curve[0].db = db_curve[1].db - (3 << 16);
/* top of knee is 3dB above the threshold for soft knee */
db_curve[2].db = db_curve[1].db + (3 << 16);
if (ratio)
/* offset = -3db * (ratio - 1) / ratio */
db_curve[2].offset = (int32_t)((long long)(-3 << 16)
* (ratio - 1) / ratio);
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
else
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
/* offset = -3db for hard limit */
db_curve[2].offset = (-3 << 16);
}
else
{
/* bottom of knee is at the threshold for hard knee */
db_curve[0].db = threshold << 16;
/* top of knee is at the threshold for hard knee */
db_curve[2].db = threshold << 16;
db_curve[2].offset = 0;
}
/* Calculate 0db and ~+12db offsets */
db_curve[4].db = 0xC0A8C; /* db of 2 bits clipping */
if (ratio)
{
/* offset = threshold * (ratio - 1) / ratio */
db_curve[3].offset = (int32_t)((long long)(threshold << 16)
* (ratio - 1) / ratio);
db_curve[4].offset = (int32_t)((long long)-db_curve[4].db
* (ratio - 1) / ratio) + db_curve[3].offset;
}
else
{
/* offset = threshold for hard limit */
db_curve[3].offset = (threshold << 16);
db_curve[4].offset = -db_curve[4].db + db_curve[3].offset;
}
/** Now set up the comp_curve table with compression offsets in the
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
* form of gain factors in S7.24 format
* comp_curve[0] is 0 (-infinity db) input
*/
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
comp_curve[0] = UNITY;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** comp_curve[1 to 63] are intermediate compression values
* corresponding to the 6 MSB of the input values of a non-clipped
* signal
*/
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
for (int i = 1; i < 64; i++)
{
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** db constants are stored as positive numbers;
* make them negative here
*/
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
int32_t this_db = -db[i];
/* no compression below the knee */
if (this_db <= db_curve[0].db)
comp_curve[i] = UNITY;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** if soft knee and below top of knee,
* interpolate along soft knee slope
*/
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
else if (soft_knee && (this_db <= db_curve[2].db))
comp_curve[i] = fp_factor(fp_mul(
((this_db - db_curve[0].db) / 6),
db_curve[2].offset, 16), 16) << 8;
/* interpolate along ratio slope above the knee */
else
comp_curve[i] = fp_factor(fp_mul(
fp_div((db_curve[1].db - this_db), db_curve[1].db, 16),
db_curve[3].offset, 16), 16) << 8;
}
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** comp_curve[64] is the compression level of a maximum level,
* non-clipped signal
*/
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
comp_curve[64] = fp_factor(db_curve[3].offset, 16) << 8;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** comp_curve[65] is the compression level of a maximum level,
* clipped signal
*/
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
comp_curve[65] = fp_factor(db_curve[4].offset, 16) << 8;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** if using auto peak, then makeup gain is max offset -
* 3dB headroom
*/
comp_makeup_gain = auto_gain ?
fp_factor(-(db_curve[3].offset) - 0x4AC4, 16) << 8 : UNITY;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
#if defined(ROCKBOX_HAS_LOGF) && defined(LOGF_ENABLE)
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
logf("\n *** Compression Offsets ***");
/* some settings for display only, not used in calculations */
db_curve[0].offset = 0;
db_curve[1].offset = 0;
db_curve[3].db = 0;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
for (int i = 0; i <= 4; i++)
{
logf("Curve[%d]: db: % 6.2f\toffset: % 6.2f", i,
(float)db_curve[i].db / (1 << 16),
(float)db_curve[i].offset / (1 << 16));
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
logf("\nGain factors:");
for (int i = 1; i <= 65; i++)
{
Make rbcodec/dsp includes more specific. Change-Id: Idb6af40df26f5b8499a40e8b98602261ef227044
2012-04-29 21:31:30 +00:00
DEBUGF("%02d: %.6f ", i, (float)comp_curve[i] / UNITY);
iBasso DX50/DX90: Major code cleanup and reorganization. Reorganization - Separated iBasso devices from PLATFORM_ANDROID. These are now standlone hosted targets. Most device specific code is in the firmware/target/hosted/ibasso directory. - No dependency on Android SDK, only the Android NDK is needed. 32 bit Android NDK and Android API Level 16. - Separate implementation for each device where feasible. Code cleanup - Rewrite of existing code, from simple reformat to complete reimplementation. - New backlight interface, seperating backlight from touchscreen. - Rewrite of device button handler, removing unneeded code and fixing memory leaks. - New Debug messages interface logging to Android adb logcat (DEBUGF, panicf, logf). - Rewrite of lcd device handler, removing unneeded code and fixing memory leaks. - Rewrite of audiohw device handler/pcm interface, removing unneeded code and fixing memory leaks, enabling 44.1/48kHz pthreaded playback. - Rewrite of power and powermng, proper shutdown, using batterylog results (see http://gerrit.rockbox.org/r/#/c/1047/). - Rewrite of configure (Android NDK) and device specific config. - Rewrite of the Android NDK specific Makefile. Misc - All plugins/games/demos activated. - Update tinyalsa to latest from https://github.com/tinyalsa/tinyalsa. Includes - http://gerrit.rockbox.org/r/#/c/993/ - http://gerrit.rockbox.org/r/#/c/1010/ - http://gerrit.rockbox.org/r/#/c/1035/ Does not include http://gerrit.rockbox.org/r/#/c/1007/ due to new backlight interface and new option for hold switch, touchscreen, physical button interaction. Rockbox needs the iBasso DX50/DX90 loader for startup, see http://gerrit.rockbox.org/r/#/c/1099/ The loader expects Rockbox to be installed in /mnt/sdcard/.rockbox/. If /mnt/sdcard/ is accessed as USB mass storage device, Rockbox will exit gracefully and the loader will restart Rockbox on USB disconnect. Tested on iBasso DX50. Compiled (not tested) for iBasso DX90. Compiled (not tested) for PLATFORM_ANDROID. Change-Id: I5f5e22e68f5b4cf29c28e2b40b2c265f2beb7ab7
2015-02-02 20:44:29 +00:00
if (i % 4 == 0) { DEBUGF("\n"); }
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
}
Make rbcodec/dsp includes more specific. Change-Id: Idb6af40df26f5b8499a40e8b98602261ef227044
2012-04-29 21:31:30 +00:00
DEBUGF("\n");
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
logf("Makeup gain:\t%.6f", (float)comp_makeup_gain / UNITY);
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
#endif
Consolidate compressor settings into a struct. Doing that makes things cleaner for later on. Change-Id: I4e279aa57ace16a348acc0fc09059592325ec95f
2012-04-26 21:19:16 +00:00
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
return active;
}
/** GET COMPRESSION GAIN
* Returns the required gain factor in S7.24 format in order to compress the
* sample in accordance with the compression curve. Always 1 or less.
*/
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
static inline int32_t get_compression_gain(struct sample_format *format,
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
int32_t sample)
{
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
const int frac_bits_offset = format->frac_bits - 15;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
/* sample must be positive */
if (sample < 0)
sample = -(sample + 1);
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
/* shift sample into 15 frac bit range */
if (frac_bits_offset > 0)
sample >>= frac_bits_offset;
if (frac_bits_offset < 0)
sample <<= -frac_bits_offset;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
/* normal case: sample isn't clipped */
if (sample < (1 << 15))
{
/* index is 6 MSB, rem is 9 LSB */
int index = sample >> 9;
int32_t rem = (sample & 0x1FF) << 22;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** interpolate from the compression curve:
* higher gain - ((rem / (1 << 31)) * (higher gain - lower gain))
*/
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
return comp_curve[index] - (FRACMUL(rem,
(comp_curve[index] - comp_curve[index + 1])));
}
/* sample is somewhat clipped, up to 2 bits of overhead */
if (sample < (1 << 17))
{
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** straight interpolation:
* higher gain - ((clipped portion of sample * 4/3
* / (1 << 31)) * (higher gain - lower gain))
*/
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
return comp_curve[64] - (FRACMUL(((sample - (1 << 15)) / 3) << 16,
(comp_curve[64] - comp_curve[65])));
}
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
/* sample is too clipped, return invalid value */
return -1;
}
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
/** DSP interface **/
/** SET COMPRESSOR
* Enable or disable the compressor based upon the settings
*/
void dsp_set_compressor(const struct compressor_settings *settings)
{
/* enable/disable the compressor depending upon settings */
struct dsp_config *dsp = dsp_get_config(CODEC_IDX_AUDIO);
Enable setting of global output samplerate on certain targets. Replaces the NATIVE_FREQUENCY constant with a configurable frequency. The user may select 48000Hz if the hardware supports it. The default is still 44100Hz and the minimum is 44100Hz. The setting is located in the playback settings, under "Frequency". "Frequency" was duplicated in english.lang for now to avoid having to fix every .lang file for the moment and throwing everything out of sync because of the new play_frequency feature in features.txt. The next cleanup should combine it with the one included for recording and generalize the ID label. If the hardware doesn't support 48000Hz, no setting will be available. On particular hardware where very high rates are practical and desireable, the upper bound can be extended by patching. The PCM mixer can be configured to play at the full hardware frequency range. The DSP core can configure to the hardware minimum up to the maximum playback setting (some buffers must be reserved according to the maximum rate). If only 44100Hz is supported or possible on a given target for playback, using the DSP and mixer at other samperates is possible if the hardware offers them. Change-Id: I6023cf0c0baa8bc6292b6919b4dd3618a6a25622 Reviewed-on: http://gerrit.rockbox.org/479 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-05-23 17:58:51 +00:00
bool enable = compressor_update(dsp, settings);
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
dsp_proc_enable(dsp, DSP_PROC_COMPRESSOR, enable);
dsp_proc_activate(dsp, DSP_PROC_COMPRESSOR, true);
}
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
/** COMPRESSOR PROCESS
* Changes the gain of the samples according to the compressor curve
*/
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
static void compressor_process(struct dsp_proc_entry *this,
struct dsp_buffer **buf_p)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
struct dsp_buffer *buf = *buf_p;
int count = buf->remcount;
int32_t *in_buf[2] = { buf->p32[0], buf->p32[1] };
const int num_chan = buf->format.num_channels;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
while (count-- > 0)
{
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/* Use the average of the channels */
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
int32_t sample_gain = UNITY;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
int32_t x = 0;
int32_t tmpx = 0;
int32_t in_buf_max_level = 0;
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
for (int ch = 0; ch < num_chan; ch++)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
tmpx = *in_buf[ch];
x += tmpx;
labuf[ch][delay_write] = tmpx;
/* Limiter detection */
if(tmpx < 0) tmpx = -(tmpx + 1);
if(tmpx > in_buf_max_level) in_buf_max_level = tmpx;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** Divide it by the number of channels, roughly
* It will be exact if the number of channels a power of 2
* it will be imperfect otherwise. Real division costs too
* much here, and most of the time it will be 2 channels (stereo)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
*/
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
x >>= (num_chan >> 1);
/** 1p HP Filters: y[n] = a*(y[n-1] + x - x[n-1])
* Zero and Pole in the same place to reduce computation
* Run the first pre-emphasis filter
*/
int32_t tmp1 = x - hpfx1 + hp1y1;
hp1y1 = FRACMUL_SHL(hp1ca, tmp1, 7);
/* Run the second pre-emphasis filter */
tmp1 = x - hpfx1 + hp2y1;
hp2y1 = FRACMUL_SHL(hp2ca, tmp1, 7);
hpfx1 = x;
/* Apply weighted sum to the pre-emphasis network */
sample_gain = (x>>1) + hp1y1 + (hp2y1<<1); /* x/2 + hp1 + 2*hp2 */
sample_gain >>= 1;
sample_gain += sample_gain >> 1;
sample_gain = get_compression_gain(&buf->format, sample_gain);
/* Exponential Attack and Release */
if ((sample_gain <= release_gain) && (sample_gain > 0))
{
/* Attack */
if(attca != UNITY)
{
int32_t this_gain = FRACMUL_SHL(release_gain, attcb, 7);
this_gain += FRACMUL_SHL(sample_gain, attca, 7);
release_gain = this_gain;
}
else
{
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
release_gain = sample_gain;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
}
/** reset it to delay time so it cannot release before the
* delayed signal releases
*/
release_holdoff = delay_time;
}
else
/* Reverse exponential decay to current gain value */
{
/* Don't start release while output is still above thresh */
if(release_holdoff > 0)
{
release_holdoff--;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
else
{
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/* Release */
int32_t this_gain = FRACMUL_SHL(release_gain, rlscb, 7);
this_gain += FRACMUL_SHL(sample_gain,rlsca,7);
release_gain = this_gain;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
}
/** total gain factor is the product of release gain and makeup gain,
* but avoid computation if possible
*/
int32_t total_gain = FRACMUL_SHL(release_gain, comp_makeup_gain, 7);
/* Look-ahead limiter */
int32_t test_gain = FRACMUL_SHL(total_gain, in_buf_max_level, 3);
if( test_gain > UNITY)
{
release_gain -= limitca;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
/** Implement the compressor: apply total gain factor (if any) to the
* output buffer sample pair/mono sample
*/
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
if (total_gain != UNITY)
{
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
for (int ch = 0; ch < num_chan; ch++)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
*in_buf[ch] = FRACMUL_SHL(total_gain, labuf[ch][delay_read], 7);
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
}
in_buf[0]++;
in_buf[1]++;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
delay_write++;
delay_read++;
if(delay_write >= MAX_DLY) delay_write = 0;
if(delay_read >= MAX_DLY) delay_read = 0;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
(void)this;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
/* DSP message hook */
static intptr_t compressor_configure(struct dsp_proc_entry *this,
struct dsp_config *dsp,
unsigned int setting,
intptr_t value)
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
{
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
int i,j;
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
switch (setting)
{
case DSP_PROC_INIT:
if (value != 0)
break; /* Already enabled */
Cleanup and simplify latest DSP code incarnation. Some things can just be a bit simpler in handling the list of stages and some things, especially format change handling, can be simplified for each stage implementation. Format changes are sent through the configure() callback. Hide some internal details and variables from processing stages and let the core deal with it. Do some miscellaneous cleanup and keep things a bit better factored. Change-Id: I19dd8ce1d0b792ba914d426013088a49a52ecb7e
2012-12-19 22:34:57 +00:00
this->process = compressor_process;
Enable setting of global output samplerate on certain targets. Replaces the NATIVE_FREQUENCY constant with a configurable frequency. The user may select 48000Hz if the hardware supports it. The default is still 44100Hz and the minimum is 44100Hz. The setting is located in the playback settings, under "Frequency". "Frequency" was duplicated in english.lang for now to avoid having to fix every .lang file for the moment and throwing everything out of sync because of the new play_frequency feature in features.txt. The next cleanup should combine it with the one included for recording and generalize the ID label. If the hardware doesn't support 48000Hz, no setting will be available. On particular hardware where very high rates are practical and desireable, the upper bound can be extended by patching. The PCM mixer can be configured to play at the full hardware frequency range. The DSP core can configure to the hardware minimum up to the maximum playback setting (some buffers must be reserved according to the maximum rate). If only 44100Hz is supported or possible on a given target for playback, using the DSP and mixer at other samperates is possible if the hardware offers them. Change-Id: I6023cf0c0baa8bc6292b6919b4dd3618a6a25622 Reviewed-on: http://gerrit.rockbox.org/479 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-05-23 17:58:51 +00:00
/* Won't have been getting frequency updates */
compressor_update(dsp, &curr_set);
Cleanup and simplify latest DSP code incarnation. Some things can just be a bit simpler in handling the list of stages and some things, especially format change handling, can be simplified for each stage implementation. Format changes are sent through the configure() callback. Hide some internal details and variables from processing stages and let the core deal with it. Do some miscellaneous cleanup and keep things a bit better factored. Change-Id: I19dd8ce1d0b792ba914d426013088a49a52ecb7e
2012-12-19 22:34:57 +00:00
/* Fall-through */
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
case DSP_RESET:
case DSP_FLUSH:
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
release_gain = UNITY;
DSP Compressor: Sidechain, Exponential Atk/Rls This is an improvement to the current compressor which I have added to my own Sansa Fuze V2 build. I am submitting here in case others find it interesting. Features added to the existing compressor: Attack, Look-ahead, Sidechain Filtering. Exponential attack and release characteristic response. Benefits from adding missing features: Attack: Preserve perceived "brightness" of tone by letting onset transients come through at a higher level than the rest of the compressed program material. Look-ahead: With Attack comes clipping on the leading several cycles of a transient onset. With look-ahead function, this can be pre-emptively mitigated with a slower gain change (less distortion). Look-ahead limiting is implemented to prevent clipping while keeping gain change ramp to an interval near 3ms instead of instant attack. The existing compressor implementation distorts the leading edge of a transient by causing instant gain change, resulting in log() distortion. This sounds "woofy" to me. Exponential Attack/Release: eMore natural sounding. On attack, this is a true straight line of 10dB per attack interval. Release is a little different, however, sounds natural as an analog compressor. Sidechain Filtering: Mild high-pass filter reduces response to low frequency onsets. For example, a hard kick drum is less likely to make the whole of the program material appear to fade in and out. Combined with a moderate attack time, such a transient will ride through with minimal audible artifact. Overall these changes make dynamic music sound more "open", more natural. The goal of a compressor is to make dyanamic music sound louder without necessarily sounding as though it has been compressed. I believe these changes come closer to this goal. Enjoy. If not, I am enjoying it Change-Id: I664eace546c364b815b4dc9ed4a72849231a0eb2 Reviewed-on: http://gerrit.rockbox.org/626 Tested: Purling Nayuki <cyq.yzfl@gmail.com> Reviewed-by: Michael Giacomelli <giac2000@hotmail.com>
2013-10-03 12:57:00 +00:00
for(i=0; i<MAX_CH; i++)
{
for(j=0; j<MAX_DLY; j++)
{
labuf[i][j] = 0; /* All Silence */
}
}
/* Delay Line Read/Write Pointers */
int32_t fs = dsp_get_output_frequency(dsp);
delay_read = 0;
delay_write = (DLY_TIME*fs/1000);
if(delay_write >= MAX_DLY) {
delay_write = MAX_DLY - 1; /* Limit to the max allocated buffer */
}
delay_time = delay_write;
release_holdoff = delay_write;
limitca = get_att_rls_coeff(DLY_TIME, fs); /** Attack time for
* look-ahead limiter
*/
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
break;
Enable setting of global output samplerate on certain targets. Replaces the NATIVE_FREQUENCY constant with a configurable frequency. The user may select 48000Hz if the hardware supports it. The default is still 44100Hz and the minimum is 44100Hz. The setting is located in the playback settings, under "Frequency". "Frequency" was duplicated in english.lang for now to avoid having to fix every .lang file for the moment and throwing everything out of sync because of the new play_frequency feature in features.txt. The next cleanup should combine it with the one included for recording and generalize the ID label. If the hardware doesn't support 48000Hz, no setting will be available. On particular hardware where very high rates are practical and desireable, the upper bound can be extended by patching. The PCM mixer can be configured to play at the full hardware frequency range. The DSP core can configure to the hardware minimum up to the maximum playback setting (some buffers must be reserved according to the maximum rate). If only 44100Hz is supported or possible on a given target for playback, using the DSP and mixer at other samperates is possible if the hardware offers them. Change-Id: I6023cf0c0baa8bc6292b6919b4dd3618a6a25622 Reviewed-on: http://gerrit.rockbox.org/479 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-05-23 17:58:51 +00:00
case DSP_SET_OUT_FREQUENCY:
compressor_update(dsp, &curr_set);
break;
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
}
Cleanup and simplify latest DSP code incarnation. Some things can just be a bit simpler in handling the list of stages and some things, especially format change handling, can be simplified for each stage implementation. Format changes are sent through the configure() callback. Hide some internal details and variables from processing stages and let the core deal with it. Do some miscellaneous cleanup and keep things a bit better factored. Change-Id: I19dd8ce1d0b792ba914d426013088a49a52ecb7e
2012-12-19 22:34:57 +00:00
return 0;
Move to compressor out of dsp.c and into its own source to reduce DSP clutter. A bit of a rough job for the moment but all works. Change-Id: Id40852e0dec99caee02f943d0da8a1cdc16f022a
2012-02-08 19:55:37 +00:00
}
Fundamentally rewrite much of the audio DSP. Creates a standard buffer passing, local data passing and messaging system for processing stages. Stages can be moved to their own source files to reduce clutter and ease assimilation of new ones. dsp.c becomes dsp_core.c which supports an engine and framework for effects. Formats and change notifications are passed along with the buffer so that they arrive at the correct time at each stage in the chain regardless of the internal delays of a particular one. Removes restrictions on the number of samples that can be processed at a time and it pays attention to destination buffer size restrictions without having to limit input count, which also allows pcmbuf to remain fuller and safely set its own buffer limits as it sees fit. There is no longer a need to query input/output counts given a certain number of input samples; just give it the sizes of the source and destination buffers. Works in harmony with stages that are not deterministic in terms of sample input/output ratio (like both resamplers but most notably the timestretch). As a result it fixes quirks with timestretch hanging up with certain settings and it now operates properly throughout its full settings range. Change-Id: Ib206ec78f6f6c79259c5af9009fe021d68be9734 Reviewed-on: http://gerrit.rockbox.org/200 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2012-03-27 23:52:15 +00:00
/* Database entry */
DSP_PROC_DB_ENTRY(
COMPRESSOR,
compressor_configure);
Reference in a new issue Copy permalink