avcodec/dca_core: export downmix matrix coefficients

$ ./ffmpeg -downmix stereo -i $fate-suite/dts/dcadec-suite/core_51_24_48_768_1.dtshd -y native.wav Before: $ ./ffmpeg -i $fate-suite/dts/dcadec-suite/core_51_24_48_768_1.dtshd -af aresample,aformat=channel_layouts=stereo -y swr.wav [...] [Parsed_aresample_0] [SWR] Using fltp internally between filters [Parsed_aresample_0] [SWR] Matrix coefficients: [Parsed_aresample_0] [SWR] FL: FL:0.414214 FR:0.000000 FC:0.292893 LFE:0.000000 SL:0.292893 SR:0.000000 [Parsed_aresample_0] [SWR] FR: FL:0.000000 FR:0.414214 FC:0.292893 LFE:0.000000 SL:0.000000 SR:0.292893 [Parsed_aresample_0] ch:6 chl:5.1(side) fmt:fltp r:48000Hz -> ch:2 chl:stereo fmt:s16 r:48000Hz [...] $ tests/tiny_psnr.exe native.wav swr.wav s16 stddev:18907.06 PSNR: 10.80 MAXDIFF:51590 bytes: 8192/ 8192 After: $ ./ffmpeg -i $fate-suite/dts/dcadec-suite/core_51_24_48_768_1.dtshd -af aresample,aformat=channel_layouts=stereo -y swr.wav [...] [Parsed_aresample_0] [SWR] Using fltp internally between filters [Parsed_aresample_0] [SWR] Custom matrix coefficients: [Parsed_aresample_0] [SWR] FL: FL:-0.100006 FR:0.000000 FC:0.000000 LFE:0.000000 SL:-0.707092 SR:0.000000 [Parsed_aresample_0] [SWR] FR: FL:0.000000 FR:-0.100006 FC:0.000000 LFE:0.000000 SL:0.000000 SR:-0.707092 [Parsed_aresample_0] ch:6 chl:5.1(side) fmt:fltp r:48000Hz -> ch:2 chl:stereo fmt:s16 r:48000Hz [...] $ tests/tiny_psnr.exe native.wav swr.wav stddev: 0.02 PSNR:129.44 MAXDIFF: 1 bytes: 8192/ 8192 Signed-off-by: James Almer <jamrial@gmail.com>
2026-10-24 02:09:03 +03:00 · 2026-06-04 14:51:57 -03:00
parent e1e6ebb065
commit 57708daf61
4 changed files with 87 additions and 11 deletions
--- a/libavcodec/dca_core.c
+++ b/libavcodec/dca_core.c
@@ -2384,6 +2384,17 @@ int ff_dca_core_filter_frame(DCACoreDecoder *s, AVFrame *frame)
    else
        avctx->bit_rate = 0;

+    if (s->request_mask == s->ch_mask && s->prim_dmix_embedded &&
+        !s->sumdiff_front && !s->sumdiff_surround &&
+        !(s->ext_audio_mask & (DCA_CSS_XXCH | DCA_CSS_XCH | DCA_EXSS_XXCH)) &&
+        s->audio_mode >= DCA_AMODE_2F2R && (s->prim_dmix_type == DCA_DMIX_TYPE_LoRo ||
+                                            s->prim_dmix_type == DCA_DMIX_TYPE_LtRt)) {
+        ret = ff_dca_export_downmix_matrix(avctx, frame, s->prim_dmix_type, s->ch_mask,
+                                           s->prim_dmix_coeff);
+        if (ret < 0)
+            return ret;
+    }
+
    if (s->audio_mode == DCA_AMODE_STEREO_TOTAL || (s->request_mask != s->ch_mask &&
                                                    s->prim_dmix_type == DCA_DMIX_TYPE_LtRt))
        matrix_encoding = AV_MATRIX_ENCODING_DOLBY;
--- a/libavcodec/dcadec.c
+++ b/libavcodec/dcadec.c
@@ -21,10 +21,12 @@
 #include "libavutil/mem.h"
 #include "libavutil/opt.h"
 #include "libavutil/channel_layout.h"
+#include "libavutil/downmix_info.h"
 #include "libavutil/thread.h"

 #include "codec_internal.h"
 #include "dcadec.h"
+#include "dcadata.h"
 #include "dcahuff.h"
 #include "dca_syncwords.h"
 #include "profiles.h"
@@ -32,18 +34,19 @@
 #define MIN_PACKET_SIZE     16
 #define MAX_PACKET_SIZE     0x104000

+
+static const uint8_t dca2wav_norm[28] = {
+     2,  0, 1, 9, 10,  3,  8,  4,  5,  9, 10, 6, 7, 12,
+    13, 14, 3, 6,  7, 11, 12, 14, 16, 15, 17, 8, 4,  5,
+};
+
+static const uint8_t dca2wav_wide[28] = {
+     2,  0, 1, 4,  5,  3,  8,  4,  5,  9, 10, 6, 7, 12,
+    13, 14, 3, 6,  7, 11, 12, 14, 16, 15, 17, 8, 4,  5,
+};
+
 int ff_dca_set_channel_layout(AVCodecContext *avctx, int *ch_remap, int dca_mask)
 {
-    static const uint8_t dca2wav_norm[28] = {
-         2,  0, 1, 9, 10,  3,  8,  4,  5,  9, 10, 6, 7, 12,
-        13, 14, 3, 6,  7, 11, 12, 14, 16, 15, 17, 8, 4,  5,
-    };
-
-    static const uint8_t dca2wav_wide[28] = {
-         2,  0, 1, 4,  5,  3,  8,  4,  5,  9, 10, 6, 7, 12,
-        13, 14, 3, 6,  7, 11, 12, 14, 16, 15, 17, 8, 4,  5,
-    };
-
    DCAContext *s = avctx->priv_data;

    int dca_ch, wav_ch, nchannels = 0;
@@ -156,6 +159,62 @@ void ff_dca_downmix_to_stereo_float(AVFloatDSPContext *fdsp, float **samples,
    }
 }

+int ff_dca_export_downmix_matrix(AVCodecContext *avctx, AVFrame *frame,
+                                 enum DCADownMixType downmix_type,
+                                 int output_mask, const int *coeff_l)
+{
+    enum AVDownmixType dmix_type = (downmix_type == DCA_DMIX_TYPE_LoRo) ?
+                                    AV_DOWNMIX_TYPE_LORO : AV_DOWNMIX_TYPE_LTRT;
+    size_t size;
+    AVDownmixMatrix *dm = av_downmix_matrix_alloc(dmix_type, av_popcount(output_mask), &size);
+    const int *coeff_r = coeff_l + av_popcount(output_mask);
+    const double scale = 1.0 / (1 << 15);
+    AVDownmixCoeff *matrix_l, *matrix_r;
+    const uint8_t *dca2wav;
+
+    if (!dm)
+        return AVERROR(ENOMEM);
+
+    if (output_mask == DCA_SPEAKER_LAYOUT_7POINT0_WIDE ||
+        output_mask == DCA_SPEAKER_LAYOUT_7POINT1_WIDE)
+        dca2wav = dca2wav_wide;
+    else
+        dca2wav = dca2wav_norm;
+
+    matrix_l = av_downmix_matrix_coeff(dm, 0, 0);
+    matrix_r = av_downmix_matrix_coeff(dm, 1, 0);
+
+    for (int i = 0; i <= av_log2(output_mask); i++) {
+        if (!(output_mask & (1U << i)))
+            continue;
+
+        int idx = av_channel_layout_index_from_channel(&avctx->ch_layout, dca2wav[i]);
+        av_assert0(idx >= 0);
+
+        if (*coeff_l)
+            matrix_l[idx] = *coeff_l * scale;
+
+        if (*coeff_r)
+            matrix_r[idx] = *coeff_r * scale;
+
+        coeff_l++;
+        coeff_r++;
+    }
+
+    AVBufferRef *buf = av_buffer_create((uint8_t *)dm, size, NULL, NULL, 0);
+    if (!buf) {
+        av_free(dm);
+        return AVERROR(ENOMEM);
+    }
+
+    if (!av_frame_new_side_data_from_buf(frame, AV_FRAME_DATA_DOWNMIX_MATRIX, buf)) {
+        av_buffer_unref(&buf);
+        return AVERROR(ENOMEM);
+    }
+
+    return 0;
+}
+
 static int dcadec_decode_frame(AVCodecContext *avctx, AVFrame *frame,
                               int *got_frame_ptr, AVPacket *avpkt)
 {
--- a/libavcodec/dcadec.h
+++ b/libavcodec/dcadec.h
@@ -80,6 +80,9 @@ void ff_dca_downmix_to_stereo_fixed(DCADSPContext *dcadsp, int32_t **samples,
                                    int *coeff_l, int nsamples, int ch_mask);
 void ff_dca_downmix_to_stereo_float(AVFloatDSPContext *fdsp, float **samples,
                                    int *coeff_l, int nsamples, int ch_mask);
+int ff_dca_export_downmix_matrix(AVCodecContext *avctx, AVFrame *frame,
+                                 enum DCADownMixType downmix_type,
+                                 int output_mask, const int *coeff_l);

 static inline int ff_dca_check_crc(AVCodecContext *avctx, GetBitContext *s,
                                   int p1, int p2)
--- a/tests/fate/dca.mak
+++ b/tests/fate/dca.mak
@@ -39,10 +39,13 @@ $(foreach N,$(DCADEC_SUITE_LOSSLESS_24),$(eval $(call FATE_DCADEC_LOSSLESS_SUITE
 $(foreach N,$(DCADEC_SUITE_LOSSY),$(eval $(call FATE_DCADEC_LOSSY_SUITE,$(N))))

 # lossy downmix tests
-FATE_DCADEC_LOSSY += fate-dca-core_51_24_48_768_1-dmix_2
+FATE_DCADEC_LOSSY += fate-dca-core_51_24_48_768_1-dmix_2 fate-dca-core_51_24_48_768_1-dmix_2_matrix
 fate-dca-core_51_24_48_768_1-dmix_2: CMD = ffmpeg -downmix stereo -flags2 skip_manual -i $(TARGET_SAMPLES)/dts/dcadec-suite/core_51_24_48_768_1.dtshd -f f32le -af aresample -
 fate-dca-core_51_24_48_768_1-dmix_2: REF = $(SAMPLES)/dts/dcadec-suite/core_51_24_48_768_1-dmix_2.f32

+fate-dca-core_51_24_48_768_1-dmix_2_matrix: CMD = ffmpeg -flags2 skip_manual -i $(TARGET_SAMPLES)/dts/dcadec-suite/core_51_24_48_768_1.dtshd -f f32le -af aresample,aformat=channel_layouts=stereo -
+fate-dca-core_51_24_48_768_1-dmix_2_matrix: REF = $(SAMPLES)/dts/dcadec-suite/core_51_24_48_768_1-dmix_2.f32
+
 FATE_DCADEC_LOSSY += fate-dca-x96_xxch_71_24_96_3840-dmix_2
 fate-dca-x96_xxch_71_24_96_3840-dmix_2: CMD = ffmpeg -downmix stereo -flags2 skip_manual -i $(TARGET_SAMPLES)/dts/dcadec-suite/x96_xxch_71_24_96_3840.dtshd -f f32le -af aresample -
 # intentionally uses the dmix_6 reference because the sample does not contain stereo downmix coefficients