FFmpeg
Data Structures | Macros | Enumerations | Functions | Variables
af_atempo.c File Reference

tempo scaling audio filter – an implementation of WSOLA algorithm More...

#include <float.h>
#include "libavcodec/avfft.h"
#include "libavutil/avassert.h"
#include "libavutil/avstring.h"
#include "libavutil/channel_layout.h"
#include "libavutil/eval.h"
#include "libavutil/opt.h"
#include "libavutil/samplefmt.h"
#include "avfilter.h"
#include "audio.h"
#include "internal.h"

Go to the source code of this file.

Data Structures

struct  AudioFragment
 A fragment of audio waveform. More...
 
struct  ATempoContext
 Filter state machine. More...
 

Macros

#define YAE_ATEMPO_MIN   0.5
 
#define YAE_ATEMPO_MAX   100.0
 
#define OFFSET(x)   offsetof(ATempoContext, x)
 
#define RE_MALLOC_OR_FAIL(field, field_size)
 
#define yae_init_xdat(scalar_type, scalar_max)
 A helper macro for initializing complex data buffer with scalar data of a given type. More...
 
#define yae_blend(scalar_type)
 A helper macro for blending the overlap region of previous and current audio fragment. More...
 

Enumerations

enum  FilterState {
  YAE_LOAD_FRAGMENT, YAE_ADJUST_POSITION, YAE_RELOAD_FRAGMENT, YAE_OUTPUT_OVERLAP_ADD,
  YAE_FLUSH_OUTPUT
}
 Filter state machine states. More...
 

Functions

 AVFILTER_DEFINE_CLASS (atempo)
 
static AudioFragmentyae_curr_frag (ATempoContext *atempo)
 
static AudioFragmentyae_prev_frag (ATempoContext *atempo)
 
static void yae_clear (ATempoContext *atempo)
 Reset filter to initial state, do not deallocate existing local buffers. More...
 
static void yae_release_buffers (ATempoContext *atempo)
 Reset filter to initial state and deallocate all buffers. More...
 
static int yae_reset (ATempoContext *atempo, enum AVSampleFormat format, int sample_rate, int channels)
 Prepare filter for processing audio data of given format, sample rate and number of channels. More...
 
static int yae_update (AVFilterContext *ctx)
 
static void yae_downmix (ATempoContext *atempo, AudioFragment *frag)
 Initialize complex data buffer of a given audio fragment with down-mixed mono data of appropriate scalar type. More...
 
static int yae_load_data (ATempoContext *atempo, const uint8_t **src_ref, const uint8_t *src_end, int64_t stop_here)
 Populate the internal data buffer on as-needed basis. More...
 
static int yae_load_frag (ATempoContext *atempo, const uint8_t **src_ref, const uint8_t *src_end)
 Populate current audio fragment data buffer. More...
 
static void yae_advance_to_next_frag (ATempoContext *atempo)
 Prepare for loading next audio fragment. More...
 
static void yae_xcorr_via_rdft (FFTSample *xcorr, RDFTContext *complex_to_real, const FFTComplex *xa, const FFTComplex *xb, const int window)
 Calculate cross-correlation via rDFT. More...
 
static int yae_align (AudioFragment *frag, const AudioFragment *prev, const int window, const int delta_max, const int drift, FFTSample *correlation, RDFTContext *complex_to_real)
 Calculate alignment offset for given fragment relative to the previous fragment. More...
 
static int yae_adjust_position (ATempoContext *atempo)
 Adjust current fragment position for better alignment with previous fragment. More...
 
static int yae_overlap_add (ATempoContext *atempo, uint8_t **dst_ref, uint8_t *dst_end)
 Blend the overlap region of previous and current audio fragment and output the results to the given destination buffer. More...
 
static void yae_apply (ATempoContext *atempo, const uint8_t **src_ref, const uint8_t *src_end, uint8_t **dst_ref, uint8_t *dst_end)
 Feed as much data to the filter as it is able to consume and receive as much processed data in the destination buffer as it is able to produce or store. More...
 
static int yae_flush (ATempoContext *atempo, uint8_t **dst_ref, uint8_t *dst_end)
 Flush any buffered data from the filter. More...
 
static av_cold int init (AVFilterContext *ctx)
 
static av_cold void uninit (AVFilterContext *ctx)
 
static int query_formats (AVFilterContext *ctx)
 
static int config_props (AVFilterLink *inlink)
 
static int push_samples (ATempoContext *atempo, AVFilterLink *outlink, int n_out)
 
static int filter_frame (AVFilterLink *inlink, AVFrame *src_buffer)
 
static int request_frame (AVFilterLink *outlink)
 
static int process_command (AVFilterContext *ctx, const char *cmd, const char *arg, char *res, int res_len, int flags)
 

Variables

static const AVOption atempo_options []
 
static const AVFilterPad atempo_inputs []
 
static const AVFilterPad atempo_outputs []
 
AVFilter ff_af_atempo
 

Detailed Description

tempo scaling audio filter – an implementation of WSOLA algorithm

Based on MIT licensed yaeAudioTempoFilter.h and yaeAudioFragment.h from Apprentice Video player by Pavel Koshevoy. https://sourceforge.net/projects/apprenticevideo/

An explanation of SOLA algorithm is available at http://www.surina.net/article/time-and-pitch-scaling.html

WSOLA is very similar to SOLA, only one major difference exists between these algorithms. SOLA shifts audio fragments along the output stream, where as WSOLA shifts audio fragments along the input stream.

The advantage of WSOLA algorithm is that the overlap region size is always the same, therefore the blending function is constant and can be precomputed.

Definition in file af_atempo.c.

Macro Definition Documentation

#define YAE_ATEMPO_MIN   0.5

Definition at line 155 of file af_atempo.c.

#define YAE_ATEMPO_MAX   100.0

Definition at line 156 of file af_atempo.c.

#define OFFSET (   x)    offsetof(ATempoContext, x)

Definition at line 158 of file af_atempo.c.

#define RE_MALLOC_OR_FAIL (   field,
  field_size 
)
Value:
do { \
field = av_malloc(field_size); \
if (!field) { \
return AVERROR(ENOMEM); \
} \
} while (0)
#define av_malloc(s)
it s the only field you need to keep assuming you have a context There is some magic you don t need to care about around this field
if(ret)
static void yae_release_buffers(ATempoContext *atempo)
Reset filter to initial state and deallocate all buffers.
Definition: af_atempo.c:225
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several the filter must be ready for frames arriving randomly on any input any filter with several inputs will most likely require some kind of queuing mechanism It is perfectly acceptable to have a limited queue and to drop frames when the inputs are too unbalanced request_frame For filters that do not use the this method is called when a frame is wanted on an output For a it should directly call filter_frame on the corresponding output For a if there are queued frames already one of these frames should be pushed If the filter should request a frame on one of its repeatedly until at least one frame has been pushed Return or at least make progress towards producing a it should return
#define av_freep(p)
Filter the word “frame” indicates either a video frame or a group of audio as stored in an AVFrame structure Format for each input and each output the list of supported formats For video that means pixel format For audio that means channel sample they are references to shared objects When the negotiation mechanism computes the intersection of the formats supported at each end of a all references to both lists are replaced with a reference to the intersection And when a single format is eventually chosen for a link amongst the remaining all references to the list are updated That means that if a filter requires that its input and output have the same format amongst a supported all it has to do is use a reference to the same list of formats query_formats can leave some formats unset and return AVERROR(EAGAIN) to cause the negotiation mechanism toagain later.That can be used by filters with complex requirements to use the format negotiated on one link to set the formats supported on another.Frame references ownership and permissions

Definition at line 247 of file af_atempo.c.

Referenced by yae_reset().

#define yae_init_xdat (   scalar_type,
  scalar_max 
)

A helper macro for initializing complex data buffer with scalar data of a given type.

Definition at line 346 of file af_atempo.c.

Referenced by yae_downmix().

#define yae_blend (   scalar_type)
Value:
do { \
const scalar_type *aaa = (const scalar_type *)a; \
const scalar_type *bbb = (const scalar_type *)b; \
\
scalar_type *out = (scalar_type *)dst; \
scalar_type *out_end = (scalar_type *)dst_end; \
int64_t i; \
for (i = 0; i < overlap && out < out_end; \
i++, atempo->position[1]++, wa++, wb++) { \
float w0 = *wa; \
float w1 = *wb; \
int j; \
for (j = 0; j < atempo->channels; \
j++, aaa++, bbb++, out++) { \
float t0 = (float)*aaa; \
float t1 = (float)*bbb; \
\
*out = \
frag->position[0] + i < 0 ? \
*aaa : \
(scalar_type)(t0 * w0 + t1 * w1); \
} \
} \
dst = (uint8_t *)out; \
} while (0)
The reader does not expect b to be semantically here and if the code is changed by maybe adding a a division or other the signedness will almost certainly be mistaken To avoid this confusion a new type was SUINT is the C unsigned type but it holds a signed int to use the same example SUINT a
Definition: undefined.txt:36
uint8_t
#define t0
Definition: regdef.h:28
#define i(width, name, range_min, range_max)
Definition: cbs_h2645.c:259
#define t1
Definition: regdef.h:29
#define b
Definition: input.c:41
int
FILE * out
Definition: movenc.c:54
for(j=16;j >0;--j)

A helper macro for blending the overlap region of previous and current audio fragment.

Definition at line 727 of file af_atempo.c.

Referenced by yae_overlap_add().

Enumeration Type Documentation

Filter state machine states.

Enumerator
YAE_LOAD_FRAGMENT 
YAE_ADJUST_POSITION 
YAE_RELOAD_FRAGMENT 
YAE_OUTPUT_OVERLAP_ADD 
YAE_FLUSH_OUTPUT 

Definition at line 76 of file af_atempo.c.

Function Documentation

AVFILTER_DEFINE_CLASS ( atempo  )
static AudioFragment* yae_curr_frag ( ATempoContext atempo)
inlinestatic
static AudioFragment* yae_prev_frag ( ATempoContext atempo)
inlinestatic
static void yae_clear ( ATempoContext atempo)
static

Reset filter to initial state, do not deallocate existing local buffers.

Definition at line 184 of file af_atempo.c.

Referenced by yae_release_buffers(), and yae_reset().

static void yae_release_buffers ( ATempoContext atempo)
static

Reset filter to initial state and deallocate all buffers.

Definition at line 225 of file af_atempo.c.

Referenced by uninit(), and yae_reset().

static int yae_reset ( ATempoContext atempo,
enum AVSampleFormat  format,
int  sample_rate,
int  channels 
)
static

Prepare filter for processing audio data of given format, sample rate and number of channels.

Definition at line 261 of file af_atempo.c.

Referenced by config_props().

static int yae_update ( AVFilterContext ctx)
static

Definition at line 331 of file af_atempo.c.

Referenced by process_command().

static void yae_downmix ( ATempoContext atempo,
AudioFragment frag 
)
static

Initialize complex data buffer of a given audio fragment with down-mixed mono data of appropriate scalar type.

Definition at line 396 of file af_atempo.c.

Referenced by yae_apply(), and yae_flush().

static int yae_load_data ( ATempoContext atempo,
const uint8_t **  src_ref,
const uint8_t src_end,
int64_t  stop_here 
)
static

Populate the internal data buffer on as-needed basis.

Returns
0 if requested data was already available or was successfully loaded, AVERROR(EAGAIN) if more input data is required.

Definition at line 424 of file af_atempo.c.

Referenced by yae_load_frag().

static int yae_load_frag ( ATempoContext atempo,
const uint8_t **  src_ref,
const uint8_t src_end 
)
static

Populate current audio fragment data buffer.

Returns
0 when the fragment is ready, AVERROR(EAGAIN) if more input data is required.

Definition at line 499 of file af_atempo.c.

Referenced by yae_apply(), and yae_flush().

static void yae_advance_to_next_frag ( ATempoContext atempo)
static

Prepare for loading next audio fragment.

Definition at line 579 of file af_atempo.c.

Referenced by yae_apply(), and yae_flush().

static void yae_xcorr_via_rdft ( FFTSample xcorr,
RDFTContext complex_to_real,
const FFTComplex xa,
const FFTComplex xb,
const int  window 
)
static

Calculate cross-correlation via rDFT.

Multiply two vectors of complex numbers (result of real_to_complex rDFT) and transform back via complex_to_real rDFT.

Definition at line 601 of file af_atempo.c.

Referenced by yae_align().

static int yae_align ( AudioFragment frag,
const AudioFragment prev,
const int  window,
const int  delta_max,
const int  drift,
FFTSample correlation,
RDFTContext complex_to_real 
)
static

Calculate alignment offset for given fragment relative to the previous fragment.

Returns
alignment offset of current fragment relative to previous.

Definition at line 635 of file af_atempo.c.

Referenced by yae_adjust_position().

static int yae_adjust_position ( ATempoContext atempo)
static

Adjust current fragment position for better alignment with previous fragment.

Returns
alignment correction.

Definition at line 689 of file af_atempo.c.

Referenced by yae_apply(), and yae_flush().

static int yae_overlap_add ( ATempoContext atempo,
uint8_t **  dst_ref,
uint8_t dst_end 
)
static

Blend the overlap region of previous and current audio fragment and output the results to the given destination buffer.

Returns
0 if the overlap region was completely stored in the dst buffer, AVERROR(EAGAIN) if more destination buffer space is required.

Definition at line 764 of file af_atempo.c.

Referenced by yae_apply(), and yae_flush().

static void yae_apply ( ATempoContext atempo,
const uint8_t **  src_ref,
const uint8_t src_end,
uint8_t **  dst_ref,
uint8_t dst_end 
)
static

Feed as much data to the filter as it is able to consume and receive as much processed data in the destination buffer as it is able to produce or store.

Definition at line 819 of file af_atempo.c.

Referenced by filter_frame().

static int yae_flush ( ATempoContext atempo,
uint8_t **  dst_ref,
uint8_t dst_end 
)
static

Flush any buffered data from the filter.

Returns
0 if all data was completely stored in the dst buffer, AVERROR(EAGAIN) if more destination buffer space is required.

Definition at line 893 of file af_atempo.c.

Referenced by request_frame().

static av_cold int init ( AVFilterContext ctx)
static

Definition at line 982 of file af_atempo.c.

static av_cold void uninit ( AVFilterContext ctx)
static

Definition at line 990 of file af_atempo.c.

static int query_formats ( AVFilterContext ctx)
static

Definition at line 996 of file af_atempo.c.

static int config_props ( AVFilterLink inlink)
static

Definition at line 1040 of file af_atempo.c.

static int push_samples ( ATempoContext atempo,
AVFilterLink outlink,
int  n_out 
)
static

Definition at line 1051 of file af_atempo.c.

Referenced by check_size(), filter_frame(), and request_frame().

static int filter_frame ( AVFilterLink inlink,
AVFrame src_buffer 
)
static

Definition at line 1077 of file af_atempo.c.

static int request_frame ( AVFilterLink outlink)
static

Definition at line 1125 of file af_atempo.c.

static int process_command ( AVFilterContext ctx,
const char *  cmd,
const char *  arg,
char *  res,
int  res_len,
int  flags 
)
static

Definition at line 1171 of file af_atempo.c.

Variable Documentation

const AVOption atempo_options[]
static
Initial value:
= {
{ "tempo", "set tempo scale factor",
OFFSET(tempo), AV_OPT_TYPE_DOUBLE, { .dbl = 1.0 },
{ NULL }
}
#define NULL
Definition: coverity.c:32
#define AV_OPT_FLAG_AUDIO_PARAM
Definition: opt.h:278
#define YAE_ATEMPO_MAX
Definition: af_atempo.c:156
#define AV_OPT_FLAG_FILTERING_PARAM
a generic parameter which can be set by the user for filtering
Definition: opt.h:292
#define AV_OPT_FLAG_RUNTIME_PARAM
a generic parameter which can be set by the user at runtime
Definition: opt.h:291
#define YAE_ATEMPO_MIN
Definition: af_atempo.c:155
#define OFFSET(x)
Definition: af_atempo.c:158

Definition at line 160 of file af_atempo.c.

const AVFilterPad atempo_inputs[]
static
Initial value:
= {
{
.name = "default",
.filter_frame = filter_frame,
.config_props = config_props,
},
{ NULL }
}
#define NULL
Definition: coverity.c:32
static int config_props(AVFilterLink *inlink)
Definition: af_atempo.c:1040
static int filter_frame(AVFilterLink *inlink, AVFrame *src_buffer)
Definition: af_atempo.c:1077

Definition at line 1186 of file af_atempo.c.

const AVFilterPad atempo_outputs[]
static
Initial value:
= {
{
.name = "default",
.request_frame = request_frame,
},
{ NULL }
}
#define NULL
Definition: coverity.c:32
static int request_frame(AVFilterLink *outlink)
Definition: af_atempo.c:1125

Definition at line 1196 of file af_atempo.c.

AVFilter ff_af_atempo
Initial value:
= {
.name = "atempo",
.description = NULL_IF_CONFIG_SMALL("Adjust audio tempo."),
.init = init,
.uninit = uninit,
.query_formats = query_formats,
.process_command = process_command,
.priv_size = sizeof(ATempoContext),
.priv_class = &atempo_class,
}
static const AVFilterPad atempo_outputs[]
Definition: af_atempo.c:1196
Filter state machine.
Definition: af_atempo.c:87
static av_cold int init(AVFilterContext *ctx)
Definition: af_atempo.c:982
#define NULL_IF_CONFIG_SMALL(x)
Return NULL if CONFIG_SMALL is true, otherwise the argument without modification. ...
Definition: internal.h:186
static const AVFilterPad atempo_inputs[]
Definition: af_atempo.c:1186
static int process_command(AVFilterContext *ctx, const char *cmd, const char *arg, char *res, int res_len, int flags)
Definition: af_atempo.c:1171
static const AVFilterPad outputs[]
Definition: af_acontrast.c:203
static int query_formats(AVFilterContext *ctx)
Definition: af_atempo.c:996
these buffered frames must be flushed immediately if a new input produces new the filter must not call request_frame to get more It must just process the frame or queue it The task of requesting more frames is left to the filter s request_frame method or the application If a filter has several inputs
static av_cold void uninit(AVFilterContext *ctx)
Definition: af_atempo.c:990

Definition at line 1205 of file af_atempo.c.