[FFmpeg-cvslog] Add Dolby/DPLII downmix support to libavresample
John Stebbins
git at videolan.org
Mon Jun 18 20:25:45 CEST 2012
ffmpeg | branch: master | John Stebbins <jstebbins.hb at gmail.com> | Sat Jun 9 13:45:49 2012 -0700| [3971be0eb5d382b85f1d8772c51f21cccdc68eab] | committer: Justin Ruggles
Add Dolby/DPLII downmix support to libavresample
Signed-off-by: Justin Ruggles <justin.ruggles at gmail.com>
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=3971be0eb5d382b85f1d8772c51f21cccdc68eab
---
doc/APIchanges | 3 ++
libavresample/audio_mix.c | 3 +-
libavresample/audio_mix_matrix.c | 60 +++++++++++++++++++++++++++++++++-----
libavresample/avresample.h | 3 +-
libavresample/internal.h | 1 +
libavresample/options.c | 4 +++
libavresample/version.h | 2 +-
libavutil/audioconvert.h | 7 +++++
8 files changed, 72 insertions(+), 11 deletions(-)
diff --git a/doc/APIchanges b/doc/APIchanges
index 50cc787..4d14882 100644
--- a/doc/APIchanges
+++ b/doc/APIchanges
@@ -13,6 +13,9 @@ libavutil: 2011-04-18
API changes, most recent first:
+2012-xx-xx - xxxxxxx - lavr 0.0.3
+ Add a parameter to avresample_build_matrix() for Dolby/DPLII downmixing.
+
2012-xx-xx - xxxxxxx - lavfi 2.23.0 - avfilter.h
Add AVFilterContext.nb_inputs/outputs. Deprecate
AVFilterContext.input/output_count.
diff --git a/libavresample/audio_mix.c b/libavresample/audio_mix.c
index 7ab11b0..9319222 100644
--- a/libavresample/audio_mix.c
+++ b/libavresample/audio_mix.c
@@ -320,7 +320,8 @@ int ff_audio_mix_init(AVAudioResampleContext *avr)
avr->center_mix_level,
avr->surround_mix_level,
avr->lfe_mix_level, 1, matrix_dbl,
- avr->in_channels);
+ avr->in_channels,
+ avr->matrix_encoding);
if (ret < 0) {
av_free(matrix_dbl);
return ret;
diff --git a/libavresample/audio_mix_matrix.c b/libavresample/audio_mix_matrix.c
index 6135b02..f7121c8 100644
--- a/libavresample/audio_mix_matrix.c
+++ b/libavresample/audio_mix_matrix.c
@@ -54,6 +54,8 @@
#define SURROUND_DIRECT_LEFT 33
#define SURROUND_DIRECT_RIGHT 34
+#define SQRT3_2 1.22474487139158904909 /* sqrt(3/2) */
+
static av_always_inline int even(uint64_t layout)
{
return (!layout || (layout & (layout - 1)));
@@ -83,14 +85,21 @@ static int sane_layout(uint64_t layout)
int avresample_build_matrix(uint64_t in_layout, uint64_t out_layout,
double center_mix_level, double surround_mix_level,
double lfe_mix_level, int normalize,
- double *matrix_out, int stride)
+ double *matrix_out, int stride,
+ enum AVMatrixEncoding matrix_encoding)
{
int i, j, out_i, out_j;
double matrix[64][64] = {{0}};
- int64_t unaccounted = in_layout & ~out_layout;
+ int64_t unaccounted;
double maxcoef = 0;
int in_channels, out_channels;
+ if ((out_layout & AV_CH_LAYOUT_STEREO_DOWNMIX) == AV_CH_LAYOUT_STEREO_DOWNMIX) {
+ out_layout = AV_CH_LAYOUT_STEREO;
+ }
+
+ unaccounted = in_layout & ~out_layout;
+
in_channels = av_get_channel_layout_nb_channels( in_layout);
out_channels = av_get_channel_layout_nb_channels(out_layout);
@@ -140,8 +149,19 @@ int avresample_build_matrix(uint64_t in_layout, uint64_t out_layout,
matrix[SIDE_LEFT ][BACK_CENTER] += M_SQRT1_2;
matrix[SIDE_RIGHT][BACK_CENTER] += M_SQRT1_2;
} else if (out_layout & AV_CH_FRONT_LEFT) {
- matrix[FRONT_LEFT ][BACK_CENTER] += surround_mix_level * M_SQRT1_2;
- matrix[FRONT_RIGHT][BACK_CENTER] += surround_mix_level * M_SQRT1_2;
+ if (matrix_encoding == AV_MATRIX_ENCODING_DOLBY ||
+ matrix_encoding == AV_MATRIX_ENCODING_DPLII) {
+ if (unaccounted & (AV_CH_BACK_LEFT | AV_CH_SIDE_LEFT)) {
+ matrix[FRONT_LEFT ][BACK_CENTER] -= surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][BACK_CENTER] += surround_mix_level * M_SQRT1_2;
+ } else {
+ matrix[FRONT_LEFT ][BACK_CENTER] -= surround_mix_level;
+ matrix[FRONT_RIGHT][BACK_CENTER] += surround_mix_level;
+ }
+ } else {
+ matrix[FRONT_LEFT ][BACK_CENTER] += surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][BACK_CENTER] += surround_mix_level * M_SQRT1_2;
+ }
} else if (out_layout & AV_CH_FRONT_CENTER) {
matrix[FRONT_CENTER][BACK_CENTER] += surround_mix_level * M_SQRT1_2;
} else
@@ -163,8 +183,20 @@ int avresample_build_matrix(uint64_t in_layout, uint64_t out_layout,
matrix[SIDE_RIGHT][BACK_RIGHT] += 1.0;
}
} else if (out_layout & AV_CH_FRONT_LEFT) {
- matrix[FRONT_LEFT ][BACK_LEFT ] += surround_mix_level;
- matrix[FRONT_RIGHT][BACK_RIGHT] += surround_mix_level;
+ if (matrix_encoding == AV_MATRIX_ENCODING_DOLBY) {
+ matrix[FRONT_LEFT ][BACK_LEFT ] -= surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_LEFT ][BACK_RIGHT] -= surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][BACK_LEFT ] += surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][BACK_RIGHT] += surround_mix_level * M_SQRT1_2;
+ } else if (matrix_encoding == AV_MATRIX_ENCODING_DPLII) {
+ matrix[FRONT_LEFT ][BACK_LEFT ] -= surround_mix_level * SQRT3_2;
+ matrix[FRONT_LEFT ][BACK_RIGHT] -= surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][BACK_LEFT ] += surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][BACK_RIGHT] += surround_mix_level * SQRT3_2;
+ } else {
+ matrix[FRONT_LEFT ][BACK_LEFT ] += surround_mix_level;
+ matrix[FRONT_RIGHT][BACK_RIGHT] += surround_mix_level;
+ }
} else if (out_layout & AV_CH_FRONT_CENTER) {
matrix[FRONT_CENTER][BACK_LEFT ] += surround_mix_level * M_SQRT1_2;
matrix[FRONT_CENTER][BACK_RIGHT] += surround_mix_level * M_SQRT1_2;
@@ -187,8 +219,20 @@ int avresample_build_matrix(uint64_t in_layout, uint64_t out_layout,
matrix[BACK_CENTER][SIDE_LEFT ] += M_SQRT1_2;
matrix[BACK_CENTER][SIDE_RIGHT] += M_SQRT1_2;
} else if (out_layout & AV_CH_FRONT_LEFT) {
- matrix[FRONT_LEFT ][SIDE_LEFT ] += surround_mix_level;
- matrix[FRONT_RIGHT][SIDE_RIGHT] += surround_mix_level;
+ if (matrix_encoding == AV_MATRIX_ENCODING_DOLBY) {
+ matrix[FRONT_LEFT ][SIDE_LEFT ] -= surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_LEFT ][SIDE_RIGHT] -= surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][SIDE_LEFT ] += surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][SIDE_RIGHT] += surround_mix_level * M_SQRT1_2;
+ } else if (matrix_encoding == AV_MATRIX_ENCODING_DPLII) {
+ matrix[FRONT_LEFT ][SIDE_LEFT ] -= surround_mix_level * SQRT3_2;
+ matrix[FRONT_LEFT ][SIDE_RIGHT] -= surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][SIDE_LEFT ] += surround_mix_level * M_SQRT1_2;
+ matrix[FRONT_RIGHT][SIDE_RIGHT] += surround_mix_level * SQRT3_2;
+ } else {
+ matrix[FRONT_LEFT ][SIDE_LEFT ] += surround_mix_level;
+ matrix[FRONT_RIGHT][SIDE_RIGHT] += surround_mix_level;
+ }
} else if (out_layout & AV_CH_FRONT_CENTER) {
matrix[FRONT_CENTER][SIDE_LEFT ] += surround_mix_level * M_SQRT1_2;
matrix[FRONT_CENTER][SIDE_RIGHT] += surround_mix_level * M_SQRT1_2;
diff --git a/libavresample/avresample.h b/libavresample/avresample.h
index 65d4d2d..002bec2 100644
--- a/libavresample/avresample.h
+++ b/libavresample/avresample.h
@@ -131,12 +131,13 @@ void avresample_free(AVAudioResampleContext **avr);
* the weight of input channel i in output channel o.
* @param stride distance between adjacent input channels in the
* matrix array
+ * @param matrix_encoding matrixed stereo downmix mode (e.g. dplii)
* @return 0 on success, negative AVERROR code on failure
*/
int avresample_build_matrix(uint64_t in_layout, uint64_t out_layout,
double center_mix_level, double surround_mix_level,
double lfe_mix_level, int normalize, double *matrix,
- int stride);
+ int stride, enum AVMatrixEncoding matrix_encoding);
/**
* Get the current channel mixing matrix.
diff --git a/libavresample/internal.h b/libavresample/internal.h
index 49ea6a6..fa9499a 100644
--- a/libavresample/internal.h
+++ b/libavresample/internal.h
@@ -70,6 +70,7 @@ struct AVAudioResampleContext {
AudioConvert *ac_out; /**< output sample format conversion context */
ResampleContext *resample; /**< resampling context */
AudioMix *am; /**< channel mixing context */
+ enum AVMatrixEncoding matrix_encoding; /**< matrixed stereo encoding */
};
#endif /* AVRESAMPLE_INTERNAL_H */
diff --git a/libavresample/options.c b/libavresample/options.c
index 5430c4d..a1a0b0c 100644
--- a/libavresample/options.c
+++ b/libavresample/options.c
@@ -52,6 +52,10 @@ static const AVOption options[] = {
{ "phase_shift", "Resampling Phase Shift", OFFSET(phase_shift), AV_OPT_TYPE_INT, { 10 }, 0, 30, /* ??? */ PARAM },
{ "linear_interp", "Use Linear Interpolation", OFFSET(linear_interp), AV_OPT_TYPE_INT, { 0 }, 0, 1, PARAM },
{ "cutoff", "Cutoff Frequency Ratio", OFFSET(cutoff), AV_OPT_TYPE_DOUBLE, { 0.8 }, 0.0, 1.0, PARAM },
+ { "matrix_encoding", "Matrixed Stereo Encoding", OFFSET(matrix_encoding), AV_OPT_TYPE_INT, { AV_MATRIX_ENCODING_NONE}, AV_MATRIX_ENCODING_NONE, AV_MATRIX_ENCODING_NB-1, PARAM, "matrix_encoding" },
+ { "none", "None", 0, AV_OPT_TYPE_CONST, { AV_MATRIX_ENCODING_NONE }, INT_MIN, INT_MAX, PARAM, "matrix_encoding" },
+ { "dolby", "Dolby", 0, AV_OPT_TYPE_CONST, { AV_MATRIX_ENCODING_DOLBY }, INT_MIN, INT_MAX, PARAM, "matrix_encoding" },
+ { "dplii", "Dolby Pro Logic II", 0, AV_OPT_TYPE_CONST, { AV_MATRIX_ENCODING_DPLII }, INT_MIN, INT_MAX, PARAM, "matrix_encoding" },
{ NULL },
};
diff --git a/libavresample/version.h b/libavresample/version.h
index 6211a56..63f07f5 100644
--- a/libavresample/version.h
+++ b/libavresample/version.h
@@ -21,7 +21,7 @@
#define LIBAVRESAMPLE_VERSION_MAJOR 0
#define LIBAVRESAMPLE_VERSION_MINOR 0
-#define LIBAVRESAMPLE_VERSION_MICRO 2
+#define LIBAVRESAMPLE_VERSION_MICRO 3
#define LIBAVRESAMPLE_VERSION_INT AV_VERSION_INT(LIBAVRESAMPLE_VERSION_MAJOR, \
LIBAVRESAMPLE_VERSION_MINOR, \
diff --git a/libavutil/audioconvert.h b/libavutil/audioconvert.h
index 691c64a..7e79097 100644
--- a/libavutil/audioconvert.h
+++ b/libavutil/audioconvert.h
@@ -101,6 +101,13 @@
#define AV_CH_LAYOUT_OCTAGONAL (AV_CH_LAYOUT_5POINT0|AV_CH_BACK_LEFT|AV_CH_BACK_CENTER|AV_CH_BACK_RIGHT)
#define AV_CH_LAYOUT_STEREO_DOWNMIX (AV_CH_STEREO_LEFT|AV_CH_STEREO_RIGHT)
+enum AVMatrixEncoding {
+ AV_MATRIX_ENCODING_NONE,
+ AV_MATRIX_ENCODING_DOLBY,
+ AV_MATRIX_ENCODING_DPLII,
+ AV_MATRIX_ENCODING_NB
+};
+
/**
* @}
*/
More information about the ffmpeg-cvslog
mailing list