rockbox/lib/rbcodec/codecs/codecs.h

297 lines
10 KiB
C
Raw Normal View History

/***************************************************************************
* __________ __ ___.
* Open \______ \ ____ ____ | | _\_ |__ _______ ___
* Source | _// _ \_/ ___\| |/ /| __ \ / _ \ \/ /
* Jukebox | | ( <_> ) \___| < | \_\ ( <_> > < <
* Firmware |____|_ /\____/ \___ >__|_ \|___ /\____/__/\_ \
* \/ \/ \/ \/ \/
* $Id$
*
* Copyright (C) 2002 Björn Stenberg
*
* 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.
*
****************************************************************************/
#ifndef _CODECS_H_
#define _CODECS_H_
/* instruct simulator code to not redefine any symbols when compiling codecs.
(the CODEC macro is defined in codecs.make) */
#ifdef CODEC
#define NO_REDEFINES_PLEASE
#endif
#include "rbcodecconfig.h"
#include "metadata.h"
#include "audio.h"
#ifdef RB_PROFILE
#include "profile.h"
#include "thread.h"
#endif
#ifdef HAVE_RECORDING
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
#include "enc_base.h"
#endif
#include "dsp_core.h"
#include "dsp_misc.h"
#include "dsp-util.h"
#include "gcc_extensions.h"
#include "load_code.h"
#include <limits.h>
#ifdef CODEC
#if defined(DEBUG) || defined(SIMULATOR)
#undef DEBUGF
#define DEBUGF ci->debugf
#undef LDEBUGF
#define LDEBUGF ci->debugf
#else
#define DEBUGF(...)
#define LDEBUGF(...)
#endif
#ifdef ROCKBOX_HAS_LOGF
#undef LOGF
#define LOGF ci->logf
#else
#define LOGF(...)
#endif
#endif
/* magic for normal codecs */
#define CODEC_MAGIC 0x52434F44 /* RCOD */
/* magic for encoder codecs */
#define CODEC_ENC_MAGIC 0x52454E43 /* RENC */
/* increase this every time the api struct changes */
#define CODEC_API_VERSION 48
/* update this to latest version if a change to the api struct breaks
backwards compatibility (and please take the opportunity to sort in any
new function which are "waiting" at the end of the function table) */
#define CODEC_MIN_API_VERSION 48
/* reasons for calling codec main entrypoint */
enum codec_entry_call_reason {
CODEC_LOAD = 0,
CODEC_UNLOAD
};
/* codec return codes */
enum codec_status {
CODEC_OK = 0,
CODEC_ERROR = -1,
};
/* codec command action codes */
enum codec_command_action {
CODEC_ACTION_HALT = -1,
CODEC_ACTION_NULL = 0,
CODEC_ACTION_SEEK_TIME = 1,
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
#ifdef HAVE_RECORDING
CODEC_ACTION_STREAM_FINISH = 2,
#endif
CODEC_ACTION_MIN = LONG_MIN,
CODEC_ACTION_MAX = LONG_MAX,
};
/* NOTE: To support backwards compatibility, only add new functions at
the end of the structure. Every time you add a new function,
remember to increase CODEC_API_VERSION. If you make changes to the
existing APIs then also update CODEC_MIN_API_VERSION to current
version
*/
struct codec_api {
off_t filesize; /* Total file length */
off_t curpos; /* Current buffer position */
struct mp3entry *id3; /* TAG metadata pointer */
int audio_hid; /* Current audio handle */
/* The dsp instance to be used for audio output */
struct dsp_config *dsp;
/* Returns buffer to malloc array. Only codeclib should need this. */
void* (*codec_get_buffer)(size_t *size);
/* Insert PCM data into audio buffer for playback. Playback will start
automatically. */
void (*pcmbuf_insert)(const void *ch1, const void *ch2, int count);
/* Set song position in WPS (value in ms). */
void (*set_elapsed)(unsigned long value);
/* Read next <size> amount bytes from file buffer to <ptr>.
Will return number of bytes read or 0 if end of file. */
size_t (*read_filebuf)(void *ptr, size_t size);
/* Request pointer to file buffer which can be used to read
<realsize> amount of data. <reqsize> tells the buffer system
how much data it should try to allocate. If <realsize> is 0,
end of file is reached. */
void* (*request_buffer)(size_t *realsize, size_t reqsize);
/* Advance file buffer position by <amount> amount of bytes. */
void (*advance_buffer)(size_t amount);
/* Seek file buffer to position <newpos> beginning of file. */
bool (*seek_buffer)(size_t newpos);
/* Codec should call this function when it has done the seeking. */
void (*seek_complete)(void);
/* Update the current position */
void (*set_offset)(size_t value);
/* Configure different codec buffer parameters. */
void (*configure)(int setting, intptr_t value);
/* Obtain command action on what to do next */
long (*get_command)(intptr_t *param);
/* Determine whether the track should be looped, if applicable. */
bool (*loop_track)(void);
/* kernel/ system */
#if defined(CPU_ARM) && CONFIG_PLATFORM & PLATFORM_NATIVE
void (*__div0)(void);
#endif
unsigned (*sleep)(unsigned ticks);
void (*yield)(void);
#if NUM_CORES > 1
unsigned int
(*create_thread)(void (*function)(void), void* stack,
size_t stack_size, unsigned flags, const char *name
IF_PRIO(, int priority)
IF_COP(, unsigned int core));
void (*thread_thaw)(unsigned int thread_id);
void (*thread_wait)(unsigned int thread_id);
void (*semaphore_init)(struct semaphore *s, int max, int start);
int (*semaphore_wait)(struct semaphore *s, int timeout);
void (*semaphore_release)(struct semaphore *s);
#endif /* NUM_CORES */
void (*commit_dcache)(void);
void (*commit_discard_dcache)(void);
void (*commit_discard_idcache)(void);
/* strings and memory */
char* (*strcpy)(char *dst, const char *src);
size_t (*strlen)(const char *str);
int (*strcmp)(const char *, const char *);
char *(*strcat)(char *s1, const char *s2);
void* (*memset)(void *dst, int c, size_t length);
void* (*memcpy)(void *out, const void *in, size_t n);
void* (*memmove)(void *out, const void *in, size_t n);
int (*memcmp)(const void *s1, const void *s2, size_t n);
void *(*memchr)(const void *s1, int c, size_t n);
#if defined(DEBUG) || defined(SIMULATOR)
void (*debugf)(const char *fmt, ...) ATTRIBUTE_PRINTF(1, 2);
#endif
#ifdef ROCKBOX_HAS_LOGF
void (*logf)(const char *fmt, ...) ATTRIBUTE_PRINTF(1, 2);
#endif
/* Tremor requires qsort */
void (*qsort)(void *base, size_t nmemb, size_t size,
int(*compar)(const void *, const void *));
#ifdef RB_PROFILE
void (*profile_thread)(void);
void (*profstop)(void);
void (*profile_func_enter)(void *this_fn, void *call_site);
void (*profile_func_exit)(void *this_fn, void *call_site);
#endif
#ifdef HAVE_RECORDING
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
int (*enc_pcmbuf_read)(void *buf, int count);
int (*enc_pcmbuf_advance)(int count);
struct enc_chunk_data * (*enc_encbuf_get_buffer)(size_t need);
void (*enc_encbuf_finish_buffer)(void);
ssize_t (*enc_stream_read)(void *buf, size_t count);
off_t (*enc_stream_lseek)(off_t offset, int whence);
ssize_t (*enc_stream_write)(const void *buf, size_t count);
int (*round_value_to_list32)(unsigned long value,
const unsigned long list[],
int count,
bool signd);
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
#endif /* HAVE_RECORDING */
/* new stuff at the end, sort into place next time
the API gets incompatible */
};
/* codec header */
struct codec_header {
struct lc_header lc_hdr; /* must be first */
enum codec_status(*entry_point)(enum codec_entry_call_reason reason);
enum codec_status(*run_proc)(void);
struct codec_api **api;
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
void * rec_extension[]; /* extension for encoders */
};
#ifdef CODEC
#if (CONFIG_PLATFORM & PLATFORM_NATIVE)
/* plugin_* is correct, codecs use the plugin linker script */
extern unsigned char plugin_start_addr[];
extern unsigned char plugin_end_addr[];
/* decoders */
#define CODEC_HEADER \
const struct codec_header __header \
__attribute__ ((section (".header")))= { \
{ CODEC_MAGIC, TARGET_ID, CODEC_API_VERSION, \
plugin_start_addr, plugin_end_addr }, codec_start, \
codec_run, &ci };
/* encoders */
#define CODEC_ENC_HEADER \
const struct codec_header __header \
__attribute__ ((section (".header")))= { \
{ CODEC_ENC_MAGIC, TARGET_ID, CODEC_API_VERSION, \
plugin_start_addr, plugin_end_addr }, codec_start, \
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
codec_run, &ci, { enc_callback } };
#else /* def SIMULATOR */
/* decoders */
#define CODEC_HEADER \
const struct codec_header __header \
__attribute__((visibility("default"))) = { \
{ CODEC_MAGIC, TARGET_ID, CODEC_API_VERSION, NULL, NULL }, \
codec_start, codec_run, &ci };
/* encoders */
#define CODEC_ENC_HEADER \
const struct codec_header __header \
__attribute__((visibility("default"))) = { \
{ CODEC_ENC_MAGIC, TARGET_ID, CODEC_API_VERSION, NULL, NULL }, \
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
codec_start, codec_run, &ci, { enc_callback } };
#endif /* SIMULATOR */
#endif /* CODEC */
/* create full codec path from root filenames in audio_formats[]
assumes buffer size is MAX_PATH */
void codec_get_full_path(char *path, const char *codec_root_fn);
/* Returns pointer to and size of free codec RAM */
void *codec_get_buffer_callback(size_t *size);
/* defined by the codec loader (codec.c) */
int codec_load_buf(int hid, struct codec_api *api);
int codec_load_file(const char* codec, struct codec_api *api);
int codec_run_proc(void);
int codec_close(void);
#if defined(HAVE_RECORDING)
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
enc_callback_t codec_get_enc_callback(void);
#endif
/* defined by the codec */
enum codec_status codec_start(enum codec_entry_call_reason reason);
enum codec_status codec_main(enum codec_entry_call_reason reason);
enum codec_status codec_run(void);
#if defined(HAVE_RECORDING)
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
int enc_callback(enum enc_callback_reason reason, void *params);
#else
#define codec_get_enc_callback() NULL
Update software recording engine to latest codec interface. Basically, just give it a good rewrite. Software codec recording can be implemented in a more straightforward and simple manner and made more robust through the better codec control now available. Encoded audio buffer uses a packed format instead of fixed-size chunks and uses smaller data headers leading to more efficient usage. The greatest benefit is with a VBR format like wavpack which needs to request a maximum size but only actually ends up committing part of that request. No guard buffers are used for either PCM or encoded audio. PCM is read into the codec's provided buffer and mono conversion done at that time in the core if required. Any highly-specialized sample conversion is still done within the codec itself, such as 32-bit (wavpack) or interleaved mono (mp3). There is no longer a separate filename array. All metadata goes onto the main encoded audio buffer, eliminating any predermined file limit on the buffer as well as not wasting the space for unused path queue slots. The core and codec interface is less awkward and a bit more sensible. Some less useful interface features were removed. Threads are kept on narrow code paths ie. the audio thread never calls encoding functions and the codec thread never calls file functions as before. Codecs no longer call file functions directly. Writes are buffered in the core and data written to storage in larger chunks to speed up flushing of data. In fact, codecs are no longer aware of the stream being a file at all and have no access to the fd. SPDIF frequency detection no longer requires a restart of recording or plugging the source before entering the screen. It will poll for changes and update when stopped or prerecording (which does discard now-invalid prerecorded data). I've seen to it that writing a proper header on full disk works when the format makes it reasonably practical to do so. Other cases may have incorrect data sizes but sample info will be in tact. File left that way may play anyway. mp3_enc.codec acquires the ability to write 'Info' headers with LAME tags to make it gapless (bonus). Change-Id: I670685166d5eb32ef58ef317f50b8af766ceb653 Reviewed-on: http://gerrit.rockbox.org/493 Reviewed-by: Michael Sevakis <jethead71@rockbox.org> Tested-by: Michael Sevakis <jethead71@rockbox.org>
2013-06-22 20:41:16 +00:00
#endif
#endif /* _CODECS_H_ */