mirror of https://git.ffmpeg.org/ffmpeg.git
Libspeex VAD support
Option for Voice Activity Detection is added to speex encoder. Speex detects non-speech periods and encodes them with just enough bits to reproduce the background noise, aka ``comfort noise generation''. Signed-off-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
parent
44ae7badae
commit
c785b6db73
|
@ -75,6 +75,7 @@
|
||||||
#include "internal.h"
|
#include "internal.h"
|
||||||
#include "audio_frame_queue.h"
|
#include "audio_frame_queue.h"
|
||||||
|
|
||||||
|
/* TODO: Think about converting abr, vad, dtx (still to come) and such flags to a bit field */
|
||||||
typedef struct {
|
typedef struct {
|
||||||
AVClass *class; ///< AVClass for private options
|
AVClass *class; ///< AVClass for private options
|
||||||
SpeexBits bits; ///< libspeex bitwriter context
|
SpeexBits bits; ///< libspeex bitwriter context
|
||||||
|
@ -84,6 +85,7 @@ typedef struct {
|
||||||
float vbr_quality; ///< VBR quality 0.0 to 10.0
|
float vbr_quality; ///< VBR quality 0.0 to 10.0
|
||||||
int cbr_quality; ///< CBR quality 0 to 10
|
int cbr_quality; ///< CBR quality 0 to 10
|
||||||
int abr; ///< flag to enable ABR
|
int abr; ///< flag to enable ABR
|
||||||
|
int vad; ///< flag to enable VAD
|
||||||
int pkt_frame_count; ///< frame count for the current packet
|
int pkt_frame_count; ///< frame count for the current packet
|
||||||
AudioFrameQueue afq; ///< frame queue
|
AudioFrameQueue afq; ///< frame queue
|
||||||
} LibSpeexEncContext;
|
} LibSpeexEncContext;
|
||||||
|
@ -118,6 +120,7 @@ static av_cold void print_enc_params(AVCodecContext *avctx,
|
||||||
s->frames_per_packet);
|
s->frames_per_packet);
|
||||||
av_log(avctx, AV_LOG_DEBUG, "packet size: %d\n",
|
av_log(avctx, AV_LOG_DEBUG, "packet size: %d\n",
|
||||||
avctx->frame_size * s->frames_per_packet);
|
avctx->frame_size * s->frames_per_packet);
|
||||||
|
av_log(avctx, AV_LOG_DEBUG, "voice activity detection: %d\n", s->vad);
|
||||||
}
|
}
|
||||||
|
|
||||||
static av_cold int encode_init(AVCodecContext *avctx)
|
static av_cold int encode_init(AVCodecContext *avctx)
|
||||||
|
@ -158,6 +161,7 @@ static av_cold int encode_init(AVCodecContext *avctx)
|
||||||
if (avctx->flags & CODEC_FLAG_QSCALE) {
|
if (avctx->flags & CODEC_FLAG_QSCALE) {
|
||||||
/* VBR */
|
/* VBR */
|
||||||
s->header.vbr = 1;
|
s->header.vbr = 1;
|
||||||
|
s->vad = 1; /* VAD is always implicitly activated for VBR */
|
||||||
speex_encoder_ctl(s->enc_state, SPEEX_SET_VBR, &s->header.vbr);
|
speex_encoder_ctl(s->enc_state, SPEEX_SET_VBR, &s->header.vbr);
|
||||||
s->vbr_quality = av_clipf(avctx->global_quality / (float)FF_QP2LAMBDA,
|
s->vbr_quality = av_clipf(avctx->global_quality / (float)FF_QP2LAMBDA,
|
||||||
0.0f, 10.0f);
|
0.0f, 10.0f);
|
||||||
|
@ -189,6 +193,10 @@ static av_cold int encode_init(AVCodecContext *avctx)
|
||||||
avctx->bit_rate = s->header.bitrate + (avctx->channels == 2 ? 800 : 0);
|
avctx->bit_rate = s->header.bitrate + (avctx->channels == 2 ? 800 : 0);
|
||||||
}
|
}
|
||||||
|
|
||||||
|
/* VAD is activated with VBR or can be turned on by itself */
|
||||||
|
if (s->vad)
|
||||||
|
speex_encoder_ctl(s->enc_state, SPEEX_SET_VAD, &s->vad);
|
||||||
|
|
||||||
/* set encoding complexity */
|
/* set encoding complexity */
|
||||||
if (avctx->compression_level > FF_COMPRESSION_DEFAULT) {
|
if (avctx->compression_level > FF_COMPRESSION_DEFAULT) {
|
||||||
complexity = av_clip(avctx->compression_level, 0, 10);
|
complexity = av_clip(avctx->compression_level, 0, 10);
|
||||||
|
@ -308,6 +316,7 @@ static const AVOption options[] = {
|
||||||
{ "abr", "Use average bit rate", OFFSET(abr), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, AE },
|
{ "abr", "Use average bit rate", OFFSET(abr), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, AE },
|
||||||
{ "cbr_quality", "Set quality value (0 to 10) for CBR", OFFSET(cbr_quality), AV_OPT_TYPE_INT, { .i64 = 8 }, 0, 10, AE },
|
{ "cbr_quality", "Set quality value (0 to 10) for CBR", OFFSET(cbr_quality), AV_OPT_TYPE_INT, { .i64 = 8 }, 0, 10, AE },
|
||||||
{ "frames_per_packet", "Number of frames to encode in each packet", OFFSET(frames_per_packet), AV_OPT_TYPE_INT, { .i64 = 1 }, 1, 8, AE },
|
{ "frames_per_packet", "Number of frames to encode in each packet", OFFSET(frames_per_packet), AV_OPT_TYPE_INT, { .i64 = 1 }, 1, 8, AE },
|
||||||
|
{ "vad", "Voice Activity Detection", OFFSET(vad), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, AE },
|
||||||
{ NULL },
|
{ NULL },
|
||||||
};
|
};
|
||||||
|
|
||||||
|
|
Loading…
Reference in New Issue