July 2020 - ffmpeg-cvslog - lists.ffmpeg.org

libsvtav1: Fix the documentation to match the actual options
by Mark Thompson 31 Jul '20

31 Jul '20

ffmpeg | branch: master | Mark Thompson <sw(a)jkqxz.net> | Thu Jul 30 22:32:00 2020 +0100| [11a2d05b3fa977ea9437725531a4875b735c977e] | committer: Mark Thompson libsvtav1: Fix the documentation to match the actual options > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=11a2d05b3fa977ea9… --- doc/encoders.texi | 48 ++++++++++++++++++++++++++++++++++++------------ libavcodec/libsvtav1.c | 6 +++--- 2 files changed, 39 insertions(+), 15 deletions(-) diff --git a/doc/encoders.texi b/doc/encoders.texi index af7d2ba983..2f5457fb26 100644 --- a/doc/encoders.texi +++ b/doc/encoders.texi @@ -1687,7 +1687,7 @@ Enable smooth interintra mode. Default is true. @end table -@section libsvt-av1 +@section libsvtav1 SVT-AV1 encoder wrapper. @@ -1697,31 +1697,55 @@ You need to explicitly configure the build with @code{--enable-libsvtav1}. @subsection Options @table @option +@item profile +Set the encoding profile. + +@item level +Set the operating point level. + +@item tier +Set the operating point tier. + +@item rc +Set the rate control mode to use. + +Possible modes: +@table @option +@item cqp +Constant quantizer: use fixed values of qindex (dependent on the frame type) +throughout the stream. This mode is the default. + +@item vbr +Variable bitrate: use a target bitrate for the whole stream. + +@item cvbr +Constrained variable bitrate: use a target bitrate for each GOP. +@end table + @item qmax -Sets the maximum quantizer to use when using bitrate mode. +Set the maximum quantizer to use when using a bitrate mode. @item qmin -Sets the minimum quantizer to use when using bitrate mode. +Set the minimum quantizer to use when using a bitrate mode. @item qp -Uses quantizer mode to encode at the given quantizer (0-63). +Set the quantizer used in cqp rate control mode (0-63). -@item hielevel +@item sc_detection Enable scene change detection. @item la_depth Set number of frames to look ahead (0-120). @item preset -Set encoding preset. +Set the quality-speed tradeoff, in the range 0 to 8. Higher values are +faster but lower quality. Defaults to 8 (highest speed). -@item tier - -@item tile-rows -Selects how many rows of tiles to encode with (0-6). +@item tile_rows +Set log2 of the number of rows of tiles to use (0-6). -@item tile-columns -Selects how many columns of tiles to encode with (0-4). +@item tile_columns +Set log2 of the number of columns of tiles to use (0-4). @end table diff --git a/libavcodec/libsvtav1.c b/libavcodec/libsvtav1.c index 13d3adc0f8..40942c6f6a 100644 --- a/libavcodec/libsvtav1.c +++ b/libavcodec/libsvtav1.c @@ -489,7 +489,7 @@ static const AVOption options[] = { { "preset", "Encoding preset [0, 8]", OFFSET(enc_mode), AV_OPT_TYPE_INT, { .i64 = MAX_ENC_PRESET }, 0, MAX_ENC_PRESET, VE }, - { "tier", "Set tier (general_tier_flag)", OFFSET(tier), + { "tier", "Set operating point tier", OFFSET(tier), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1, VE, "tier" }, { "main", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, 0, 0, VE, "tier" }, { "high", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = 1 }, 0, 0, VE, "tier" }, @@ -526,11 +526,11 @@ static const AVOption options[] = { { "rc", "Bit rate control mode", OFFSET(rc_mode), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 3, VE , "rc"}, - { "cqp", "Const Quantization Parameter", 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, INT_MIN, INT_MAX, VE, "rc" }, + { "cqp", "Constant quantizer", 0, AV_OPT_TYPE_CONST, { .i64 = 0 }, INT_MIN, INT_MAX, VE, "rc" }, { "vbr", "Variable Bit Rate, use a target bitrate for the entire stream", 0, AV_OPT_TYPE_CONST, { .i64 = 1 }, INT_MIN, INT_MAX, VE, "rc" }, { "cvbr", "Constrained Variable Bit Rate, use a target bitrate for each GOP", 0, AV_OPT_TYPE_CONST,{ .i64 = 2 }, INT_MIN, INT_MAX, VE, "rc" }, - { "qp", "QP value for intra frames", OFFSET(qp), + { "qp", "Quantizer to use with cqp rate control mode", OFFSET(qp), AV_OPT_TYPE_INT, { .i64 = 50 }, 0, 63, VE }, { "sc_detection", "Scene change detection", OFFSET(scd),

1 0

libsvtav1: Use _ rather than - in options
by Mark Thompson 31 Jul '20

31 Jul '20

ffmpeg | branch: master | Mark Thompson <sw(a)jkqxz.net> | Thu Jul 30 22:31:59 2020 +0100| [ad7aa1e8e7e60d3db1264c16c8cd891782ab8887] | committer: Mark Thompson libsvtav1: Use _ rather than - in options > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=ad7aa1e8e7e60d3db… --- libavcodec/libsvtav1.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/libavcodec/libsvtav1.c b/libavcodec/libsvtav1.c index 0c951286c7..13d3adc0f8 100644 --- a/libavcodec/libsvtav1.c +++ b/libavcodec/libsvtav1.c @@ -536,8 +536,8 @@ static const AVOption options[] = { { "sc_detection", "Scene change detection", OFFSET(scd), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, VE }, - { "tile-columns", "Log2 of number of tile columns to use", OFFSET(tile_columns), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 4, VE}, - { "tile-rows", "Log2 of number of tile rows to use", OFFSET(tile_rows), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 6, VE}, + { "tile_columns", "Log2 of number of tile columns to use", OFFSET(tile_columns), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 4, VE}, + { "tile_rows", "Log2 of number of tile rows to use", OFFSET(tile_rows), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 6, VE}, {NULL}, };

1 0

libsvtav1: Rename without a _
by Mark Thompson 31 Jul '20

31 Jul '20

ffmpeg | branch: master | Mark Thompson <sw(a)jkqxz.net> | Thu Jul 30 22:31:58 2020 +0100| [30a4bdbc1f4853c1deb2cd551d8c661f7b067503] | committer: Mark Thompson libsvtav1: Rename without a _ The external library is called libsvtav1, so use this name everywhere. > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=30a4bdbc1f4853c1d… --- configure | 2 +- libavcodec/Makefile | 2 +- libavcodec/allcodecs.c | 2 +- libavcodec/{libsvt_av1.c => libsvtav1.c} | 8 ++++---- 4 files changed, 7 insertions(+), 7 deletions(-) diff --git a/configure b/configure index 169f23e17f..b137669dd3 100755 --- a/configure +++ b/configure @@ -3245,7 +3245,7 @@ libshine_encoder_select="audio_frame_queue" libspeex_decoder_deps="libspeex" libspeex_encoder_deps="libspeex" libspeex_encoder_select="audio_frame_queue" -libsvt_av1_encoder_deps="libsvtav1" +libsvtav1_encoder_deps="libsvtav1" libtheora_encoder_deps="libtheora" libtwolame_encoder_deps="libtwolame" libvo_amrwbenc_encoder_deps="libvo_amrwbenc" diff --git a/libavcodec/Makefile b/libavcodec/Makefile index 9d4d52d048..98b388e260 100644 --- a/libavcodec/Makefile +++ b/libavcodec/Makefile @@ -1026,7 +1026,7 @@ OBJS-$(CONFIG_LIBRAV1E_ENCODER) += librav1e.o OBJS-$(CONFIG_LIBSHINE_ENCODER) += libshine.o OBJS-$(CONFIG_LIBSPEEX_DECODER) += libspeexdec.o OBJS-$(CONFIG_LIBSPEEX_ENCODER) += libspeexenc.o -OBJS-$(CONFIG_LIBSVT_AV1_ENCODER) += libsvt_av1.o +OBJS-$(CONFIG_LIBSVTAV1_ENCODER) += libsvtav1.o OBJS-$(CONFIG_LIBTHEORA_ENCODER) += libtheoraenc.o OBJS-$(CONFIG_LIBTWOLAME_ENCODER) += libtwolame.o OBJS-$(CONFIG_LIBVO_AMRWBENC_ENCODER) += libvo-amrwbenc.o diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c index 80142899fe..2e291f749f 100644 --- a/libavcodec/allcodecs.c +++ b/libavcodec/allcodecs.c @@ -727,7 +727,7 @@ extern AVCodec ff_librsvg_decoder; extern AVCodec ff_libshine_encoder; extern AVCodec ff_libspeex_encoder; extern AVCodec ff_libspeex_decoder; -extern AVCodec ff_libsvt_av1_encoder; +extern AVCodec ff_libsvtav1_encoder; extern AVCodec ff_libtheora_encoder; extern AVCodec ff_libtwolame_encoder; extern AVCodec ff_libvo_amrwbenc_encoder; diff --git a/libavcodec/libsvt_av1.c b/libavcodec/libsvtav1.c similarity index 99% rename from libavcodec/libsvt_av1.c rename to libavcodec/libsvtav1.c index c7ae5f9691..0c951286c7 100644 --- a/libavcodec/libsvt_av1.c +++ b/libavcodec/libsvtav1.c @@ -543,7 +543,7 @@ static const AVOption options[] = { }; static const AVClass class = { - .class_name = "libsvt_av1", + .class_name = "libsvtav1", .item_name = av_default_item_name, .option = options, .version = LIBAVUTIL_VERSION_INT, @@ -557,8 +557,8 @@ static const AVCodecDefault eb_enc_defaults[] = { { NULL }, }; -AVCodec ff_libsvt_av1_encoder = { - .name = "libsvt_av1", +AVCodec ff_libsvtav1_encoder = { + .name = "libsvtav1", .long_name = NULL_IF_CONFIG_SMALL("SVT-AV1(Scalable Video Technology for AV1) encoder"), .priv_data_size = sizeof(SvtContext), .type = AVMEDIA_TYPE_VIDEO, @@ -573,5 +573,5 @@ AVCodec ff_libsvt_av1_encoder = { .priv_class = &class, .defaults = eb_enc_defaults, .caps_internal = FF_CODEC_CAP_INIT_CLEANUP, - .wrapper_name = "libsvt_av1", + .wrapper_name = "libsvtav1", };

1 0

tests/imgutils: test the output of av_image_fill_* functions
by James Almer 30 Jul '20

30 Jul '20

ffmpeg | branch: master | James Almer <jamrial(a)gmail.com> | Sat Jul 25 00:10:38 2020 -0300| [134a48a88023a02f941afdd7d21c34897ade4d08] | committer: James Almer tests/imgutils: test the output of av_image_fill_* functions Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=134a48a88023a02f9… --- libavutil/tests/imgutils.c | 38 ++++++++++ tests/ref/fate/imgutils | 183 +++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 221 insertions(+) diff --git a/libavutil/tests/imgutils.c b/libavutil/tests/imgutils.c index 571045c857..748bd6c9d2 100644 --- a/libavutil/tests/imgutils.c +++ b/libavutil/tests/imgutils.c @@ -22,6 +22,7 @@ int main(void) { + const AVPixFmtDescriptor *desc = NULL; int64_t x, y; for (y = -1; y<UINT_MAX; y+= y/2 + 1) { @@ -31,6 +32,43 @@ int main(void) } printf("\n"); } + printf("\n"); + + while (desc = av_pix_fmt_desc_next(desc)) { + uint8_t *data[4]; + size_t sizes[4]; + ptrdiff_t linesizes1[4], offsets[3] = { 0 }; + int i, total_size, w = 64, h = 48, linesizes[4]; + enum AVPixelFormat pix_fmt = av_pix_fmt_desc_get_id(desc); + + if (av_image_fill_linesizes(linesizes, pix_fmt, w) < 0) + continue; + for (i = 0; i < 4; i++) + linesizes1[i] = linesizes[i]; + if (av_image_fill_plane_sizes(sizes, pix_fmt, h, linesizes1) < 0) + continue; + total_size = av_image_fill_pointers(data, pix_fmt, h, (void *)1, linesizes); + if (total_size < 0) + continue; + printf("%-16s", desc->name); + for (i = 0; i < 4 && data[i]; i++); + printf("planes: %d", i); + // Test the output of av_image_fill_linesizes() + printf(", linesizes:"); + for (i = 0; i < 4; i++) + printf(" %3d", linesizes[i]); + // Test the output of av_image_fill_plane_sizes() + printf(", plane_sizes:"); + for (i = 0; i < 4; i++) + printf(" %5"SIZE_SPECIFIER, sizes[i]); + // Test the output of av_image_fill_pointers() + for (i = 0; i < 3 && data[i + 1]; i++) + offsets[i] = data[i + 1] - data[i]; + printf(", plane_offsets:"); + for (i = 0; i < 3; i++) + printf(" %5"PTRDIFF_SPECIFIER, offsets[i]); + printf(", total_size: %d\n", total_size); + } return 0; } diff --git a/tests/ref/fate/imgutils b/tests/ref/fate/imgutils index aba482221f..c968ea0e85 100644 --- a/tests/ref/fate/imgutils +++ b/tests/ref/fate/imgutils @@ -53,3 +53,186 @@ 0000000000000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000000000000 0000000000000000000000000000000000000000000000000000000 + +yuv420p planes: 3, linesizes: 64 32 32 0, plane_sizes: 3072 768 768 0, plane_offsets: 3072 768 0, total_size: 4608 +yuyv422 planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +rgb24 planes: 1, linesizes: 192 0 0 0, plane_sizes: 9216 0 0 0, plane_offsets: 0 0 0, total_size: 9216 +bgr24 planes: 1, linesizes: 192 0 0 0, plane_sizes: 9216 0 0 0, plane_offsets: 0 0 0, total_size: 9216 +yuv422p planes: 3, linesizes: 64 32 32 0, plane_sizes: 3072 1536 1536 0, plane_offsets: 3072 1536 0, total_size: 6144 +yuv444p planes: 3, linesizes: 64 64 64 0, plane_sizes: 3072 3072 3072 0, plane_offsets: 3072 3072 0, total_size: 9216 +yuv410p planes: 3, linesizes: 64 16 16 0, plane_sizes: 3072 192 192 0, plane_offsets: 3072 192 0, total_size: 3456 +yuv411p planes: 3, linesizes: 64 16 16 0, plane_sizes: 3072 768 768 0, plane_offsets: 3072 768 0, total_size: 4608 +gray planes: 2, linesizes: 64 0 0 0, plane_sizes: 3072 1024 0 0, plane_offsets: 3072 0 0, total_size: 4096 +monow planes: 1, linesizes: 8 0 0 0, plane_sizes: 384 0 0 0, plane_offsets: 0 0 0, total_size: 384 +monob planes: 1, linesizes: 8 0 0 0, plane_sizes: 384 0 0 0, plane_offsets: 0 0 0, total_size: 384 +pal8 planes: 2, linesizes: 64 0 0 0, plane_sizes: 3072 1024 0 0, plane_offsets: 3072 0 0, total_size: 4096 +yuvj420p planes: 3, linesizes: 64 32 32 0, plane_sizes: 3072 768 768 0, plane_offsets: 3072 768 0, total_size: 4608 +yuvj422p planes: 3, linesizes: 64 32 32 0, plane_sizes: 3072 1536 1536 0, plane_offsets: 3072 1536 0, total_size: 6144 +yuvj444p planes: 3, linesizes: 64 64 64 0, plane_sizes: 3072 3072 3072 0, plane_offsets: 3072 3072 0, total_size: 9216 +uyvy422 planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +uyyvyy411 planes: 1, linesizes: 96 0 0 0, plane_sizes: 4608 0 0 0, plane_offsets: 0 0 0, total_size: 4608 +bgr8 planes: 2, linesizes: 64 0 0 0, plane_sizes: 3072 1024 0 0, plane_offsets: 3072 0 0, total_size: 4096 +bgr4 planes: 1, linesizes: 32 0 0 0, plane_sizes: 1536 0 0 0, plane_offsets: 0 0 0, total_size: 1536 +bgr4_byte planes: 2, linesizes: 64 0 0 0, plane_sizes: 3072 1024 0 0, plane_offsets: 3072 0 0, total_size: 4096 +rgb8 planes: 2, linesizes: 64 0 0 0, plane_sizes: 3072 1024 0 0, plane_offsets: 3072 0 0, total_size: 4096 +rgb4 planes: 1, linesizes: 32 0 0 0, plane_sizes: 1536 0 0 0, plane_offsets: 0 0 0, total_size: 1536 +rgb4_byte planes: 2, linesizes: 64 0 0 0, plane_sizes: 3072 1024 0 0, plane_offsets: 3072 0 0, total_size: 4096 +nv12 planes: 2, linesizes: 64 64 0 0, plane_sizes: 3072 1536 0 0, plane_offsets: 3072 0 0, total_size: 4608 +nv21 planes: 2, linesizes: 64 64 0 0, plane_sizes: 3072 1536 0 0, plane_offsets: 3072 0 0, total_size: 4608 +argb planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +rgba planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +abgr planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +bgra planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +gray16be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +gray16le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +yuv440p planes: 3, linesizes: 64 64 64 0, plane_sizes: 3072 1536 1536 0, plane_offsets: 3072 1536 0, total_size: 6144 +yuvj440p planes: 3, linesizes: 64 64 64 0, plane_sizes: 3072 1536 1536 0, plane_offsets: 3072 1536 0, total_size: 6144 +yuva420p planes: 4, linesizes: 64 32 32 64, plane_sizes: 3072 768 768 3072, plane_offsets: 3072 768 768, total_size: 7680 +rgb48be planes: 1, linesizes: 384 0 0 0, plane_sizes: 18432 0 0 0, plane_offsets: 0 0 0, total_size: 18432 +rgb48le planes: 1, linesizes: 384 0 0 0, plane_sizes: 18432 0 0 0, plane_offsets: 0 0 0, total_size: 18432 +rgb565be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +rgb565le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +rgb555be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +rgb555le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bgr565be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bgr565le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bgr555be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bgr555le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +yuv420p16le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv420p16be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv422p16le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv422p16be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv444p16le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuv444p16be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +rgb444le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +rgb444be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bgr444le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bgr444be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +ya8 planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bgr48be planes: 1, linesizes: 384 0 0 0, plane_sizes: 18432 0 0 0, plane_offsets: 0 0 0, total_size: 18432 +bgr48le planes: 1, linesizes: 384 0 0 0, plane_sizes: 18432 0 0 0, plane_offsets: 0 0 0, total_size: 18432 +yuv420p9be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv420p9le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv420p10be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv420p10le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv422p10be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv422p10le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv444p9be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuv444p9le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuv444p10be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuv444p10le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuv422p9be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv422p9le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +gbrp planes: 3, linesizes: 64 64 64 0, plane_sizes: 3072 3072 3072 0, plane_offsets: 3072 3072 0, total_size: 9216 +gbrp9be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +gbrp9le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +gbrp10be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +gbrp10le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +gbrp16be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +gbrp16le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuva422p planes: 4, linesizes: 64 32 32 64, plane_sizes: 3072 1536 1536 3072, plane_offsets: 3072 1536 1536, total_size: 9216 +yuva444p planes: 4, linesizes: 64 64 64 64, plane_sizes: 3072 3072 3072 3072, plane_offsets: 3072 3072 3072, total_size: 12288 +yuva420p9be planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 1536 1536 6144, plane_offsets: 6144 1536 1536, total_size: 15360 +yuva420p9le planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 1536 1536 6144, plane_offsets: 6144 1536 1536, total_size: 15360 +yuva422p9be planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 3072 3072 6144, plane_offsets: 6144 3072 3072, total_size: 18432 +yuva422p9le planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 3072 3072 6144, plane_offsets: 6144 3072 3072, total_size: 18432 +yuva444p9be planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +yuva444p9le planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +yuva420p10be planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 1536 1536 6144, plane_offsets: 6144 1536 1536, total_size: 15360 +yuva420p10le planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 1536 1536 6144, plane_offsets: 6144 1536 1536, total_size: 15360 +yuva422p10be planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 3072 3072 6144, plane_offsets: 6144 3072 3072, total_size: 18432 +yuva422p10le planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 3072 3072 6144, plane_offsets: 6144 3072 3072, total_size: 18432 +yuva444p10be planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +yuva444p10le planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +yuva420p16be planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 1536 1536 6144, plane_offsets: 6144 1536 1536, total_size: 15360 +yuva420p16le planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 1536 1536 6144, plane_offsets: 6144 1536 1536, total_size: 15360 +yuva422p16be planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 3072 3072 6144, plane_offsets: 6144 3072 3072, total_size: 18432 +yuva422p16le planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 3072 3072 6144, plane_offsets: 6144 3072 3072, total_size: 18432 +yuva444p16be planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +yuva444p16le planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +xyz12le planes: 1, linesizes: 384 0 0 0, plane_sizes: 18432 0 0 0, plane_offsets: 0 0 0, total_size: 18432 +xyz12be planes: 1, linesizes: 384 0 0 0, plane_sizes: 18432 0 0 0, plane_offsets: 0 0 0, total_size: 18432 +nv16 planes: 2, linesizes: 64 64 0 0, plane_sizes: 3072 3072 0 0, plane_offsets: 3072 0 0, total_size: 6144 +nv20le planes: 2, linesizes: 128 128 0 0, plane_sizes: 6144 6144 0 0, plane_offsets: 6144 0 0, total_size: 12288 +nv20be planes: 2, linesizes: 128 128 0 0, plane_sizes: 6144 6144 0 0, plane_offsets: 6144 0 0, total_size: 12288 +rgba64be planes: 1, linesizes: 512 0 0 0, plane_sizes: 24576 0 0 0, plane_offsets: 0 0 0, total_size: 24576 +rgba64le planes: 1, linesizes: 512 0 0 0, plane_sizes: 24576 0 0 0, plane_offsets: 0 0 0, total_size: 24576 +bgra64be planes: 1, linesizes: 512 0 0 0, plane_sizes: 24576 0 0 0, plane_offsets: 0 0 0, total_size: 24576 +bgra64le planes: 1, linesizes: 512 0 0 0, plane_sizes: 24576 0 0 0, plane_offsets: 0 0 0, total_size: 24576 +yvyu422 planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +ya16be planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +ya16le planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +gbrap planes: 4, linesizes: 64 64 64 64, plane_sizes: 3072 3072 3072 3072, plane_offsets: 3072 3072 3072, total_size: 12288 +gbrap16be planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +gbrap16le planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +0rgb planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +rgb0 planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +0bgr planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +bgr0 planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +yuv420p12be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv420p12le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv420p14be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv420p14le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 1536 1536 0, plane_offsets: 6144 1536 0, total_size: 9216 +yuv422p12be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv422p12le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv422p14be planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv422p14le planes: 3, linesizes: 128 64 64 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv444p12be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuv444p12le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuv444p14be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuv444p14le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +gbrp12be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +gbrp12le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +gbrp14be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +gbrp14le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 6144 6144 0, plane_offsets: 6144 6144 0, total_size: 18432 +yuvj411p planes: 3, linesizes: 64 16 16 0, plane_sizes: 3072 768 768 0, plane_offsets: 3072 768 0, total_size: 4608 +bayer_bggr8 planes: 1, linesizes: 64 0 0 0, plane_sizes: 3072 0 0 0, plane_offsets: 0 0 0, total_size: 3072 +bayer_rggb8 planes: 1, linesizes: 64 0 0 0, plane_sizes: 3072 0 0 0, plane_offsets: 0 0 0, total_size: 3072 +bayer_gbrg8 planes: 1, linesizes: 64 0 0 0, plane_sizes: 3072 0 0 0, plane_offsets: 0 0 0, total_size: 3072 +bayer_grbg8 planes: 1, linesizes: 64 0 0 0, plane_sizes: 3072 0 0 0, plane_offsets: 0 0 0, total_size: 3072 +bayer_bggr16le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bayer_bggr16be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bayer_rggb16le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bayer_rggb16be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bayer_gbrg16le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bayer_gbrg16be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bayer_grbg16le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +bayer_grbg16be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +yuv440p10le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv440p10be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv440p12le planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +yuv440p12be planes: 3, linesizes: 128 128 128 0, plane_sizes: 6144 3072 3072 0, plane_offsets: 6144 3072 0, total_size: 12288 +ayuv64le planes: 1, linesizes: 512 0 0 0, plane_sizes: 24576 0 0 0, plane_offsets: 0 0 0, total_size: 24576 +ayuv64be planes: 1, linesizes: 512 0 0 0, plane_sizes: 24576 0 0 0, plane_offsets: 0 0 0, total_size: 24576 +p010le planes: 2, linesizes: 128 128 0 0, plane_sizes: 6144 3072 0 0, plane_offsets: 6144 0 0, total_size: 9216 +p010be planes: 2, linesizes: 128 128 0 0, plane_sizes: 6144 3072 0 0, plane_offsets: 6144 0 0, total_size: 9216 +gbrap12be planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +gbrap12le planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +gbrap10be planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +gbrap10le planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +gray12be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +gray12le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +gray10be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +gray10le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +p016le planes: 2, linesizes: 128 128 0 0, plane_sizes: 6144 3072 0 0, plane_offsets: 6144 0 0, total_size: 9216 +p016be planes: 2, linesizes: 128 128 0 0, plane_sizes: 6144 3072 0 0, plane_offsets: 6144 0 0, total_size: 9216 +gray9be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +gray9le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +gbrpf32be planes: 3, linesizes: 256 256 256 0, plane_sizes: 12288 12288 12288 0, plane_offsets: 12288 12288 0, total_size: 36864 +gbrpf32le planes: 3, linesizes: 256 256 256 0, plane_sizes: 12288 12288 12288 0, plane_offsets: 12288 12288 0, total_size: 36864 +gbrapf32be planes: 4, linesizes: 256 256 256 256, plane_sizes: 12288 12288 12288 12288, plane_offsets: 12288 12288 12288, total_size: 49152 +gbrapf32le planes: 4, linesizes: 256 256 256 256, plane_sizes: 12288 12288 12288 12288, plane_offsets: 12288 12288 12288, total_size: 49152 +gray14be planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +gray14le planes: 1, linesizes: 128 0 0 0, plane_sizes: 6144 0 0 0, plane_offsets: 0 0 0, total_size: 6144 +grayf32be planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +grayf32le planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +yuva422p12be planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 3072 3072 6144, plane_offsets: 6144 3072 3072, total_size: 18432 +yuva422p12le planes: 4, linesizes: 128 64 64 128, plane_sizes: 6144 3072 3072 6144, plane_offsets: 6144 3072 3072, total_size: 18432 +yuva444p12be planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +yuva444p12le planes: 4, linesizes: 128 128 128 128, plane_sizes: 6144 6144 6144 6144, plane_offsets: 6144 6144 6144, total_size: 24576 +nv24 planes: 2, linesizes: 64 128 0 0, plane_sizes: 3072 6144 0 0, plane_offsets: 3072 0 0, total_size: 9216 +nv42 planes: 2, linesizes: 64 128 0 0, plane_sizes: 3072 6144 0 0, plane_offsets: 3072 0 0, total_size: 9216 +y210be planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +y210le planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +x2rgb10le planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288 +x2rgb10be planes: 1, linesizes: 256 0 0 0, plane_sizes: 12288 0 0 0, plane_offsets: 0 0 0, total_size: 12288

1 0

avformat/apm: fix APM_FILE_HEADER_SIZE value
by Zane van Iperen 30 Jul '20

30 Jul '20

ffmpeg | branch: master | Zane van Iperen <zane(a)zanevaniperen.com> | Wed Jul 29 22:06:50 2020 +1000| [be3c10d5517824318254a99007c7b1c07559c264] | committer: Zane van Iperen avformat/apm: fix APM_FILE_HEADER_SIZE value Value was incorrectly changed in 3bf1be210150b435c51c7c8eb8fd05a1fca08814. Signed-off-by: Zane van Iperen <zane(a)zanevaniperen.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=be3c10d5517824318… --- libavformat/apm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/libavformat/apm.c b/libavformat/apm.c index 4c5844c83e..38a0f6382a 100644 --- a/libavformat/apm.c +++ b/libavformat/apm.c @@ -26,7 +26,7 @@ #include "libavutil/internal.h" #include "libavutil/intreadwrite.h" -#define APM_FILE_HEADER_SIZE 18 +#define APM_FILE_HEADER_SIZE 20 #define APM_FILE_EXTRADATA_SIZE 80 #define APM_EXTRADATA_SIZE 28 @@ -287,7 +287,7 @@ static int apm_write_trailer(AVFormatContext *s) int64_t file_size, data_size; file_size = avio_tell(s->pb); - data_size = file_size - (APM_FILE_HEADER_SIZE + 2 + APM_FILE_EXTRADATA_SIZE); + data_size = file_size - (APM_FILE_HEADER_SIZE + APM_FILE_EXTRADATA_SIZE); if (file_size >= UINT32_MAX) { av_log(s, AV_LOG_ERROR,

1 0

avcodec/options_table: make AVCodecContext->level search for child constants
by James Almer 29 Jul '20

29 Jul '20

ffmpeg | branch: master | James Almer <jamrial(a)gmail.com> | Wed Jul 29 19:08:17 2020 -0300| [8a3f561c341fb2c2b7667c38a6801e2633257577] | committer: James Almer avcodec/options_table: make AVCodecContext->level search for child constants This change makes it possible for child encoders to define custom level option names which can be used for setting the AVCodecContext->level. Based on 337fe4bcc2 Reviewed-by: jkqxz Signed-off-by: James Almer <jamrial(a)gmail.com> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=8a3f561c341fb2c2b… --- libavcodec/options_table.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/libavcodec/options_table.h b/libavcodec/options_table.h index 8ba137f51e..1d0db1b5a4 100644 --- a/libavcodec/options_table.h +++ b/libavcodec/options_table.h @@ -264,8 +264,8 @@ static const AVOption avcodec_options[] = { {"profile", NULL, OFFSET(profile), AV_OPT_TYPE_INT, {.i64 = FF_PROFILE_UNKNOWN }, INT_MIN, INT_MAX, V|A|E|CC, "avctx.profile"}, {"unknown", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = FF_PROFILE_UNKNOWN }, INT_MIN, INT_MAX, V|A|E, "avctx.profile"}, {"main10", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = FF_PROFILE_HEVC_MAIN_10 }, INT_MIN, INT_MAX, V|E, "avctx.profile"}, -{"level", NULL, OFFSET(level), AV_OPT_TYPE_INT, {.i64 = FF_LEVEL_UNKNOWN }, INT_MIN, INT_MAX, V|A|E, "level"}, -{"unknown", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = FF_LEVEL_UNKNOWN }, INT_MIN, INT_MAX, V|A|E, "level"}, +{"level", NULL, OFFSET(level), AV_OPT_TYPE_INT, {.i64 = FF_LEVEL_UNKNOWN }, INT_MIN, INT_MAX, V|A|E|CC, "avctx.level"}, +{"unknown", NULL, 0, AV_OPT_TYPE_CONST, {.i64 = FF_LEVEL_UNKNOWN }, INT_MIN, INT_MAX, V|A|E, "avctx.level"}, {"lowres", "decode at 1= 1/2, 2=1/4, 3=1/8 resolutions", OFFSET(lowres), AV_OPT_TYPE_INT, {.i64 = 0 }, 0, INT_MAX, V|A|D}, #if FF_API_PRIVATE_OPT {"skip_threshold", "frame skip threshold", OFFSET(frame_skip_threshold), AV_OPT_TYPE_INT, {.i64 = DEFAULT }, INT_MIN, INT_MAX, V|E},

1 0

libavcodec/pgxdec Fix pix format selection error
by Gautam Ramakrishnan 29 Jul '20

29 Jul '20

ffmpeg | branch: master | Gautam Ramakrishnan <gautamramk(a)gmail.com> | Wed Jul 29 21:37:11 2020 +0530| [8821df6fcfc0a3b75c7079617f5917c031d0a63c] | committer: Michael Niedermayer libavcodec/pgxdec Fix pix format selection error This patch makes the pgx decoder select the correct byte order instead of selecting big endian format for 16 bit images. Signed-off-by: Michael Niedermayer <michael(a)niedermayer.cc> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=8821df6fcfc0a3b75… --- libavcodec/pgxdec.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libavcodec/pgxdec.c b/libavcodec/pgxdec.c index 93b9f4e7a0..a72d6fe6fe 100644 --- a/libavcodec/pgxdec.c +++ b/libavcodec/pgxdec.c @@ -137,7 +137,7 @@ static int pgx_decode_frame(AVCodecContext *avctx, void *data, avctx->pix_fmt = AV_PIX_FMT_GRAY8; bpp = 8; } else if (depth <= 16) { - avctx->pix_fmt = AV_PIX_FMT_GRAY16BE; + avctx->pix_fmt = AV_PIX_FMT_GRAY16; bpp = 16; } else { av_log(avctx, AV_LOG_ERROR, "Maximum depth of 16 bits supported.\n");

1 0

libavcodec/j2kenc: Encoding up to 16 bits
by Gautam Ramakrishnan 29 Jul '20

29 Jul '20

ffmpeg | branch: master | Gautam Ramakrishnan <gautamramk(a)gmail.com> | Wed Jul 29 21:37:10 2020 +0530| [9d302efdf2bc1b48896e34574f5a675c6e441a81] | committer: Michael Niedermayer libavcodec/j2kenc: Encoding up to 16 bits This patch allows the JPEG2000 encoder to encode images to up to 16 bits. Signed-off-by: Michael Niedermayer <michael(a)niedermayer.cc> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=9d302efdf2bc1b488… --- libavcodec/j2kenc.c | 99 ++++++++++++++++++++++++++++++----------------------- 1 file changed, 57 insertions(+), 42 deletions(-) diff --git a/libavcodec/j2kenc.c b/libavcodec/j2kenc.c index 38643c9a28..5ebc1f3a99 100644 --- a/libavcodec/j2kenc.c +++ b/libavcodec/j2kenc.c @@ -298,7 +298,7 @@ static int put_siz(Jpeg2000EncoderContext *s) bytestream_put_be16(&s->buf, s->ncomponents); // CSiz for (i = 0; i < s->ncomponents; i++){ // Ssiz_i XRsiz_i, YRsiz_i - bytestream_put_byte(&s->buf, 7); + bytestream_put_byte(&s->buf, s->cbps[i] - 1); bytestream_put_byte(&s->buf, i?1<<s->chroma_shift[0]:1); bytestream_put_byte(&s->buf, i?1<<s->chroma_shift[1]:1); } @@ -447,43 +447,49 @@ static int init_tiles(Jpeg2000EncoderContext *s) return 0; } -static void copy_frame(Jpeg2000EncoderContext *s) -{ - int tileno, compno, i, y, x; - uint8_t *line; - for (tileno = 0; tileno < s->numXtiles * s->numYtiles; tileno++){ - Jpeg2000Tile *tile = s->tile + tileno; - if (s->planar){ - for (compno = 0; compno < s->ncomponents; compno++){ - Jpeg2000Component *comp = tile->comp + compno; - int *dst = comp->i_data; - line = s->picture->data[compno] - + comp->coord[1][0] * s->picture->linesize[compno] - + comp->coord[0][0]; - for (y = comp->coord[1][0]; y < comp->coord[1][1]; y++){ - uint8_t *ptr = line; - for (x = comp->coord[0][0]; x < comp->coord[0][1]; x++) - *dst++ = *ptr++ - (1 << 7); - line += s->picture->linesize[compno]; - } - } - } else{ - line = s->picture->data[0] + tile->comp[0].coord[1][0] * s->picture->linesize[0] - + tile->comp[0].coord[0][0] * s->ncomponents; - - i = 0; - for (y = tile->comp[0].coord[1][0]; y < tile->comp[0].coord[1][1]; y++){ - uint8_t *ptr = line; - for (x = tile->comp[0].coord[0][0]; x < tile->comp[0].coord[0][1]; x++, i++){ - for (compno = 0; compno < s->ncomponents; compno++){ - tile->comp[compno].i_data[i] = *ptr++ - (1 << 7); - } - } - line += s->picture->linesize[0]; - } - } +#define COPY_FRAME(D, PIXEL) \ + static void copy_frame_ ##D(Jpeg2000EncoderContext *s) \ + { \ + int tileno, compno, i, y, x; \ + PIXEL *line; \ + for (tileno = 0; tileno < s->numXtiles * s->numYtiles; tileno++){ \ + Jpeg2000Tile *tile = s->tile + tileno; \ + if (s->planar){ \ + for (compno = 0; compno < s->ncomponents; compno++){ \ + Jpeg2000Component *comp = tile->comp + compno; \ + int *dst = comp->i_data; \ + int cbps = s->cbps[compno]; \ + line = (PIXEL*)s->picture->data[compno] \ + + comp->coord[1][0] * (s->picture->linesize[compno] / sizeof(PIXEL)) \ + + comp->coord[0][0]; \ + for (y = comp->coord[1][0]; y < comp->coord[1][1]; y++){ \ + PIXEL *ptr = line; \ + for (x = comp->coord[0][0]; x < comp->coord[0][1]; x++) \ + *dst++ = *ptr++ - (1 << (cbps - 1)); \ + line += s->picture->linesize[compno] / sizeof(PIXEL); \ + } \ + } \ + } else{ \ + line = (PIXEL*)s->picture->data[0] + tile->comp[0].coord[1][0] * (s->picture->linesize[0] / sizeof(PIXEL)) \ + + tile->comp[0].coord[0][0] * s->ncomponents; \ + \ + i = 0; \ + for (y = tile->comp[0].coord[1][0]; y < tile->comp[0].coord[1][1]; y++){ \ + PIXEL *ptr = line; \ + for (x = tile->comp[0].coord[0][0]; x < tile->comp[0].coord[0][1]; x++, i++){ \ + for (compno = 0; compno < s->ncomponents; compno++){ \ + int cbps = s->cbps[compno]; \ + tile->comp[compno].i_data[i] = *ptr++ - (1 << (cbps - 1)); \ + } \ + } \ + line += s->picture->linesize[0] / sizeof(PIXEL); \ + } \ + } \ + } \ } -} + +COPY_FRAME(8, uint8_t) +COPY_FRAME(16, uint16_t) static void init_quantization(Jpeg2000EncoderContext *s) { @@ -1015,7 +1021,11 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt, s->lambda = s->picture->quality * LAMBDA_SCALE; - copy_frame(s); + if (avctx->pix_fmt == AV_PIX_FMT_BGR48 || avctx->pix_fmt == AV_PIX_FMT_GRAY16) + copy_frame_16(s); + else + copy_frame_8(s); + reinit(s); if (s->format == CODEC_JP2) { @@ -1180,12 +1190,16 @@ FF_ENABLE_DEPRECATION_WARNINGS s->width = avctx->width; s->height = avctx->height; - for (i = 0; i < 3; i++) - s->cbps[i] = 8; + for (i = 0; i < 3; i++) { + if (avctx->pix_fmt == AV_PIX_FMT_GRAY16 || avctx->pix_fmt == AV_PIX_FMT_RGB48) + s->cbps[i] = 16; + else + s->cbps[i] = 8; + } - if (avctx->pix_fmt == AV_PIX_FMT_RGB24){ + if (avctx->pix_fmt == AV_PIX_FMT_RGB24 || avctx->pix_fmt == AV_PIX_FMT_RGB48){ s->ncomponents = 3; - } else if (avctx->pix_fmt == AV_PIX_FMT_GRAY8 || avctx->pix_fmt == AV_PIX_FMT_PAL8){ + } else if (avctx->pix_fmt == AV_PIX_FMT_GRAY8 || avctx->pix_fmt == AV_PIX_FMT_PAL8 || avctx->pix_fmt == AV_PIX_FMT_GRAY16){ s->ncomponents = 1; } else{ // planar YUV s->planar = 1; @@ -1255,6 +1269,7 @@ AVCodec ff_jpeg2000_encoder = { AV_PIX_FMT_YUV420P, AV_PIX_FMT_YUV422P, AV_PIX_FMT_YUV410P, AV_PIX_FMT_YUV411P, AV_PIX_FMT_PAL8, + AV_PIX_FMT_RGB48, AV_PIX_FMT_GRAY16, AV_PIX_FMT_NONE }, .priv_class = &j2k_class,

1 0

avcodec/mips: Fix segfault in imdct36_mips_float.
by Shiyou Yin 29 Jul '20

29 Jul '20

ffmpeg | branch: master | Shiyou Yin <yinshiyou-hf(a)loongson.cn> | Wed Jul 29 18:11:01 2020 +0800| [bd4f37f2ebf5870083c2343cbed6846aee63e6a5] | committer: Michael Niedermayer avcodec/mips: Fix segfault in imdct36_mips_float. 'li.s' is a synthesized instruction, it does not work properly when compiled with clang on mips, and A segfault occurred. Signed-off-by: Michael Niedermayer <michael(a)niedermayer.cc> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=bd4f37f2ebf587008… --- libavcodec/mips/aacpsdsp_mips.c | 13 +- libavcodec/mips/aacpsy_mips.h | 14 +- libavcodec/mips/fft_mips.c | 12 +- libavcodec/mips/mpegaudiodsp_mips_float.c | 492 +++++++++++++++--------------- 4 files changed, 264 insertions(+), 267 deletions(-) diff --git a/libavcodec/mips/aacpsdsp_mips.c b/libavcodec/mips/aacpsdsp_mips.c index ef47e31a9e..f63541330d 100644 --- a/libavcodec/mips/aacpsdsp_mips.c +++ b/libavcodec/mips/aacpsdsp_mips.c @@ -293,16 +293,17 @@ static void ps_decorrelate_mips(float (*out)[2], float (*delay)[2], float phi_fract0 = phi_fract[0]; float phi_fract1 = phi_fract[1]; float temp0, temp1, temp2, temp3, temp4, temp5, temp6, temp7, temp8, temp9; + float f1, f2, f3; float *p_delay_end = (p_delay + (len << 1)); /* merged 2 loops */ + f1 = 0.65143905753106; + f2 = 0.56471812200776; + f3 = 0.48954165955695; __asm__ volatile( ".set push \n\t" ".set noreorder \n\t" - "li.s %[ag0], 0.65143905753106 \n\t" - "li.s %[ag1], 0.56471812200776 \n\t" - "li.s %[ag2], 0.48954165955695 \n\t" "mul.s %[ag0], %[ag0], %[g_decay_slope] \n\t" "mul.s %[ag1], %[ag1], %[g_decay_slope] \n\t" "mul.s %[ag2], %[ag2], %[g_decay_slope] \n\t" @@ -378,10 +379,10 @@ static void ps_decorrelate_mips(float (*out)[2], float (*delay)[2], [temp3]"=&f"(temp3), [temp4]"=&f"(temp4), [temp5]"=&f"(temp5), [temp6]"=&f"(temp6), [temp7]"=&f"(temp7), [temp8]"=&f"(temp8), [temp9]"=&f"(temp9), [p_delay]"+r"(p_delay), [p_ap_delay]"+r"(p_ap_delay), - [p_Q_fract]"+r"(p_Q_fract), [p_t_gain]"+r"(p_t_gain), [p_out]"+r"(p_out), - [ag0]"=&f"(ag0), [ag1]"=&f"(ag1), [ag2]"=&f"(ag2) + [p_Q_fract]"+r"(p_Q_fract), [p_t_gain]"+r"(p_t_gain), [p_out]"+r"(p_out) : [phi_fract0]"f"(phi_fract0), [phi_fract1]"f"(phi_fract1), - [p_delay_end]"r"(p_delay_end), [g_decay_slope]"f"(g_decay_slope) + [p_delay_end]"r"(p_delay_end), [g_decay_slope]"f"(g_decay_slope), + [ag0]"f"(f1), [ag1]"f"(f2), [ag2]"f"(f3) : "memory" ); } diff --git a/libavcodec/mips/aacpsy_mips.h b/libavcodec/mips/aacpsy_mips.h index a1fe5ccea9..7d27d32f18 100644 --- a/libavcodec/mips/aacpsy_mips.h +++ b/libavcodec/mips/aacpsy_mips.h @@ -135,11 +135,11 @@ static void psy_hp_filter_mips(const float *firbuf, float *hpfsmpl, const float float coeff3 = psy_fir_coeffs[7]; float coeff4 = psy_fir_coeffs[9]; + float f1 = 32768.0; __asm__ volatile ( ".set push \n\t" ".set noreorder \n\t" - "li.s $f12, 32768 \n\t" "1: \n\t" "lwc1 $f0, 40(%[fb]) \n\t" "lwc1 $f1, 4(%[fb]) \n\t" @@ -203,14 +203,14 @@ static void psy_hp_filter_mips(const float *firbuf, float *hpfsmpl, const float "madd.s %[sum2], %[sum2], $f9, %[coeff4] \n\t" "madd.s %[sum4], %[sum4], $f6, %[coeff4] \n\t" "madd.s %[sum3], %[sum3], $f3, %[coeff4] \n\t" - "mul.s %[sum1], %[sum1], $f12 \n\t" - "mul.s %[sum2], %[sum2], $f12 \n\t" + "mul.s %[sum1], %[sum1], %[f1] \n\t" + "mul.s %[sum2], %[sum2], %[f1] \n\t" "madd.s %[sum4], %[sum4], $f11, %[coeff4] \n\t" "madd.s %[sum3], %[sum3], $f8, %[coeff4] \n\t" "swc1 %[sum1], 0(%[hp]) \n\t" "swc1 %[sum2], 4(%[hp]) \n\t" - "mul.s %[sum4], %[sum4], $f12 \n\t" - "mul.s %[sum3], %[sum3], $f12 \n\t" + "mul.s %[sum4], %[sum4], %[f1] \n\t" + "mul.s %[sum3], %[sum3], %[f1] \n\t" "swc1 %[sum4], 12(%[hp]) \n\t" "swc1 %[sum3], 8(%[hp]) \n\t" "bne %[fb], %[fb_end], 1b \n\t" @@ -223,9 +223,9 @@ static void psy_hp_filter_mips(const float *firbuf, float *hpfsmpl, const float [fb]"+r"(fb), [hp]"+r"(hp) : [coeff0]"f"(coeff0), [coeff1]"f"(coeff1), [coeff2]"f"(coeff2), [coeff3]"f"(coeff3), - [coeff4]"f"(coeff4), [fb_end]"r"(fb_end) + [coeff4]"f"(coeff4), [fb_end]"r"(fb_end), [f1]"f"(f1) : "$f0", "$f1", "$f2", "$f3", "$f4", "$f5", "$f6", - "$f7", "$f8", "$f9", "$f10", "$f11", "$f12", + "$f7", "$f8", "$f9", "$f10", "$f11", "memory" ); } diff --git a/libavcodec/mips/fft_mips.c b/libavcodec/mips/fft_mips.c index 03dcbad4d8..69abdc8a08 100644 --- a/libavcodec/mips/fft_mips.c +++ b/libavcodec/mips/fft_mips.c @@ -71,6 +71,7 @@ static void ff_fft_calc_mips(FFTContext *s, FFTComplex *z) float temp, temp1, temp3, temp4; FFTComplex * tmpz_n2, * tmpz_n34, * tmpz_n4; FFTComplex * tmpz_n2_i, * tmpz_n34_i, * tmpz_n4_i, * tmpz_i; + float f1 = 0.7071067812; num_transforms = (21845 >> (17 - s->nbits)) | 1; @@ -148,7 +149,6 @@ static void ff_fft_calc_mips(FFTContext *s, FFTComplex *z) "swc1 %[pom2], 4(%[tmpz]) \n\t" // tmpz[0].im = tmpz[0].im + tmp6; "lwc1 %[pom1], 16(%[tmpz]) \n\t" "lwc1 %[pom3], 20(%[tmpz]) \n\t" - "li.s %[pom], 0.7071067812 \n\t" // float pom = 0.7071067812f; "add.s %[temp1],%[tmp1], %[tmp2] \n\t" "sub.s %[temp], %[pom1], %[tmp8] \n\t" "add.s %[pom2], %[pom3], %[tmp7] \n\t" @@ -159,10 +159,10 @@ static void ff_fft_calc_mips(FFTContext *s, FFTComplex *z) "add.s %[pom1], %[pom1], %[tmp8] \n\t" "sub.s %[pom3], %[pom3], %[tmp7] \n\t" "add.s %[tmp3], %[tmp3], %[tmp4] \n\t" - "mul.s %[tmp5], %[pom], %[temp1] \n\t" // tmp5 = pom * (tmp1 + tmp2); - "mul.s %[tmp7], %[pom], %[temp3] \n\t" // tmp7 = pom * (tmp3 - tmp4); - "mul.s %[tmp6], %[pom], %[temp4] \n\t" // tmp6 = pom * (tmp2 - tmp1); - "mul.s %[tmp8], %[pom], %[tmp3] \n\t" // tmp8 = pom * (tmp3 + tmp4); + "mul.s %[tmp5], %[f1], %[temp1] \n\t" // tmp5 = pom * (tmp1 + tmp2); + "mul.s %[tmp7], %[f1], %[temp3] \n\t" // tmp7 = pom * (tmp3 - tmp4); + "mul.s %[tmp6], %[f1], %[temp4] \n\t" // tmp6 = pom * (tmp2 - tmp1); + "mul.s %[tmp8], %[f1], %[tmp3] \n\t" // tmp8 = pom * (tmp3 + tmp4); "swc1 %[pom1], 16(%[tmpz]) \n\t" // tmpz[2].re = tmpz[2].re + tmp8; "swc1 %[pom3], 20(%[tmpz]) \n\t" // tmpz[2].im = tmpz[2].im - tmp7; "add.s %[tmp1], %[tmp5], %[tmp7] \n\t" // tmp1 = tmp5 + tmp7; @@ -193,7 +193,7 @@ static void ff_fft_calc_mips(FFTContext *s, FFTComplex *z) [tmp3]"=&f"(tmp3), [tmp2]"=&f"(tmp2), [tmp4]"=&f"(tmp4), [tmp5]"=&f"(tmp5), [tmp7]"=&f"(tmp7), [tmp6]"=&f"(tmp6), [tmp8]"=&f"(tmp8), [pom3]"=&f"(pom3),[temp]"=&f"(temp), [temp1]"=&f"(temp1), [temp3]"=&f"(temp3), [temp4]"=&f"(temp4) - : [tmpz]"r"(tmpz) + : [tmpz]"r"(tmpz), [f1]"f"(f1) : "memory" ); } diff --git a/libavcodec/mips/mpegaudiodsp_mips_float.c b/libavcodec/mips/mpegaudiodsp_mips_float.c index 481b69c10e..ae130c752e 100644 --- a/libavcodec/mips/mpegaudiodsp_mips_float.c +++ b/libavcodec/mips/mpegaudiodsp_mips_float.c @@ -287,9 +287,16 @@ static void ff_dct32_mips_float(float *out, const float *tab) val8 , val9 , val10, val11, val12, val13, val14, val15, val16, val17, val18, val19, val20, val21, val22, val23, val24, val25, val26, val27, val28, val29, val30, val31; - float fTmp1, fTmp2, fTmp3, fTmp4, fTmp5, fTmp6, fTmp7, fTmp8, - fTmp9, fTmp10, fTmp11; + float fTmp1, fTmp2, fTmp3, fTmp4, fTmp5, fTmp6, fTmp8, fTmp9; + float f1, f2, f3, f4, f5, f6, f7; + f1 = 0.50241928618815570551; + f2 = 0.50060299823519630134; + f3 = 10.19000812354805681150; + f4 = 5.10114861868916385802; + f5 = 0.67480834145500574602; + f6 = 0.74453627100229844977; + f7 = 0.50979557910415916894; /** * instructions are scheduled to minimize pipeline stall. */ @@ -298,149 +305,142 @@ static void ff_dct32_mips_float(float *out, const float *tab) "lwc1 %[fTmp2], 31*4(%[tab]) \n\t" "lwc1 %[fTmp3], 15*4(%[tab]) \n\t" "lwc1 %[fTmp4], 16*4(%[tab]) \n\t" - "li.s %[fTmp7], 0.50241928618815570551 \n\t" "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t" "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t" "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t" "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t" - "li.s %[fTmp10], 0.50060299823519630134 \n\t" - "li.s %[fTmp11], 10.19000812354805681150 \n\t" - "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t" + "mul.s %[fTmp8], %[fTmp8], %[f2] \n\t" "add.s %[val0], %[fTmp5], %[fTmp6] \n\t" "sub.s %[val15], %[fTmp5], %[fTmp6] \n\t" "lwc1 %[fTmp1], 7*4(%[tab]) \n\t" "lwc1 %[fTmp2], 24*4(%[tab]) \n\t" - "madd.s %[val16], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "nmsub.s %[val31], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "mul.s %[val15], %[val15], %[fTmp7] \n\t" + "madd.s %[val16], %[fTmp8], %[fTmp9], %[f3] \n\t" + "nmsub.s %[val31], %[fTmp8], %[fTmp9], %[f3] \n\t" + "mul.s %[val15], %[val15], %[f1] \n\t" "lwc1 %[fTmp3], 8*4(%[tab]) \n\t" "lwc1 %[fTmp4], 23*4(%[tab]) \n\t" "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t" - "mul.s %[val31], %[val31], %[fTmp7] \n\t" + "mul.s %[val31], %[val31], %[f1] \n\t" "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t" "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t" "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t" - "li.s %[fTmp7], 5.10114861868916385802 \n\t" - "li.s %[fTmp10], 0.67480834145500574602 \n\t" - "li.s %[fTmp11], 0.74453627100229844977 \n\t" "add.s %[val7], %[fTmp5], %[fTmp6] \n\t" "sub.s %[val8], %[fTmp5], %[fTmp6] \n\t" - "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t" - "li.s %[fTmp1], 0.50979557910415916894 \n\t" + "mul.s %[fTmp8], %[fTmp8], %[f5] \n\t" "sub.s %[fTmp2], %[val0], %[val7] \n\t" - "mul.s %[val8], %[val8], %[fTmp7] \n\t" - "madd.s %[val23], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "nmsub.s %[val24], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" + "mul.s %[val8], %[val8], %[f4] \n\t" + "madd.s %[val23], %[fTmp8], %[fTmp9], %[f6] \n\t" + "nmsub.s %[val24], %[fTmp8], %[fTmp9], %[f6] \n\t" "add.s %[val0], %[val0], %[val7] \n\t" - "mul.s %[val7], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val7], %[f7], %[fTmp2] \n\t" "sub.s %[fTmp2], %[val15], %[val8] \n\t" "add.s %[val8], %[val15], %[val8] \n\t" - "mul.s %[val24], %[val24], %[fTmp7] \n\t" + "mul.s %[val24], %[val24], %[f4] \n\t" "sub.s %[fTmp3], %[val16], %[val23] \n\t" "add.s %[val16], %[val16], %[val23] \n\t" - "mul.s %[val15], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val15], %[f7], %[fTmp2] \n\t" "sub.s %[fTmp4], %[val31], %[val24] \n\t" - "mul.s %[val23], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val23], %[f7], %[fTmp3] \n\t" "add.s %[val24], %[val31], %[val24] \n\t" - "mul.s %[val31], %[fTmp1], %[fTmp4] \n\t" + "mul.s %[val31], %[f7], %[fTmp4] \n\t" : [fTmp1] "=&f" (fTmp1), [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), [fTmp5] "=&f" (fTmp5), [fTmp6] "=&f" (fTmp6), - [fTmp7] "=&f" (fTmp7), [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9), - [fTmp10] "=&f" (fTmp10), [fTmp11] "=&f" (fTmp11), - [val0] "=f" (val0), [val7] "=f" (val7), - [val8] "=f" (val8), [val15] "=f" (val15), - [val16] "=f" (val16), [val23] "=f" (val23), - [val24] "=f" (val24), [val31] "=f" (val31) - : [tab] "r" (tab) + [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9), + [val0] "=&f" (val0), [val7] "=&f" (val7), + [val8] "=&f" (val8), [val15] "=&f" (val15), + [val16] "=&f" (val16), [val23] "=&f" (val23), + [val24] "=&f" (val24), [val31] "=&f" (val31) + : [tab] "r" (tab), [f1]"f"(f1), [f2]"f"(f2), [f3]"f"(f3), + [f4]"f"(f4), [f5]"f"(f5), [f6]"f"(f6), [f7]"f"(f7) : "memory" ); + f1 = 0.64682178335999012954; + f2 = 0.53104259108978417447; + f3 = 1.48416461631416627724; + f4 = 0.78815462345125022473; + f5 = 0.55310389603444452782; + f6 = 1.16943993343288495515; + f7 = 2.56291544774150617881; __asm__ volatile ( "lwc1 %[fTmp1], 3*4(%[tab]) \n\t" "lwc1 %[fTmp2], 28*4(%[tab]) \n\t" "lwc1 %[fTmp3], 12*4(%[tab]) \n\t" "lwc1 %[fTmp4], 19*4(%[tab]) \n\t" - "li.s %[fTmp7], 0.64682178335999012954 \n\t" "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t" "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t" "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t" "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t" - "li.s %[fTmp10], 0.53104259108978417447 \n\t" - "li.s %[fTmp11], 1.48416461631416627724 \n\t" - "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t" + "mul.s %[fTmp8], %[fTmp8], %[f2] \n\t" "add.s %[val3], %[fTmp5], %[fTmp6] \n\t" "sub.s %[val12], %[fTmp5], %[fTmp6] \n\t" "lwc1 %[fTmp1], 4*4(%[tab]) \n\t" "lwc1 %[fTmp2], 27*4(%[tab]) \n\t" - "madd.s %[val19], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "nmsub.s %[val28], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "mul.s %[val12], %[val12], %[fTmp7] \n\t" + "madd.s %[val19], %[fTmp8], %[fTmp9], %[f3] \n\t" + "nmsub.s %[val28], %[fTmp8], %[fTmp9], %[f3] \n\t" + "mul.s %[val12], %[val12], %[f1] \n\t" "lwc1 %[fTmp3], 11*4(%[tab]) \n\t" "lwc1 %[fTmp4], 20*4(%[tab]) \n\t" "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t" - "mul.s %[val28], %[val28], %[fTmp7] \n\t" + "mul.s %[val28], %[val28], %[f1] \n\t" "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t" - "li.s %[fTmp7], 0.78815462345125022473 \n\t" "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t" "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t" - "li.s %[fTmp10], 0.55310389603444452782 \n\t" - "li.s %[fTmp11], 1.16943993343288495515 \n\t" - "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t" + "mul.s %[fTmp8], %[fTmp8], %[f5] \n\t" "add.s %[val4], %[fTmp5], %[fTmp6] \n\t" "sub.s %[val11], %[fTmp5], %[fTmp6] \n\t" - "li.s %[fTmp1], 2.56291544774150617881 \n\t" - "madd.s %[val20], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "nmsub.s %[val27], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "mul.s %[val11], %[val11], %[fTmp7] \n\t" + "madd.s %[val20], %[fTmp8], %[fTmp9], %[f6] \n\t" + "nmsub.s %[val27], %[fTmp8], %[fTmp9], %[f6] \n\t" + "mul.s %[val11], %[val11], %[f4] \n\t" "sub.s %[fTmp2], %[val3], %[val4] \n\t" "add.s %[val3], %[val3], %[val4] \n\t" "sub.s %[fTmp4], %[val19], %[val20] \n\t" - "mul.s %[val27], %[val27], %[fTmp7] \n\t" + "mul.s %[val27], %[val27], %[f4] \n\t" "sub.s %[fTmp3], %[val12], %[val11] \n\t" - "mul.s %[val4], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val4], %[f7], %[fTmp2] \n\t" "add.s %[val11], %[val12], %[val11] \n\t" "add.s %[val19], %[val19], %[val20] \n\t" - "mul.s %[val20], %[fTmp1], %[fTmp4] \n\t" - "mul.s %[val12], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val20], %[f7], %[fTmp4] \n\t" + "mul.s %[val12], %[f7], %[fTmp3] \n\t" "sub.s %[fTmp2], %[val28], %[val27] \n\t" "add.s %[val27], %[val28], %[val27] \n\t" - "mul.s %[val28], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val28], %[f7], %[fTmp2] \n\t" : [fTmp1] "=&f" (fTmp1), [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), [fTmp5] "=&f" (fTmp5), [fTmp6] "=&f" (fTmp6), - [fTmp7] "=&f" (fTmp7), [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9), - [fTmp10] "=&f" (fTmp10), [fTmp11] "=&f" (fTmp11), - [val3] "=f" (val3), [val4] "=f" (val4), - [val11] "=f" (val11), [val12] "=f" (val12), - [val19] "=f" (val19), [val20] "=f" (val20), - [val27] "=f" (val27), [val28] "=f" (val28) - : [tab] "r" (tab) + [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9), + [val3] "=&f" (val3), [val4] "=&f" (val4), + [val11] "=&f" (val11), [val12] "=&f" (val12), + [val19] "=&f" (val19), [val20] "=&f" (val20), + [val27] "=&f" (val27), [val28] "=&f" (val28) + : [tab] "r" (tab), [f1]"f"(f1), [f2]"f"(f2), [f3]"f"(f3), + [f4]"f"(f4), [f5]"f"(f5), [f6]"f"(f6), [f7]"f"(f7) : "memory" ); + f1 = 0.54119610014619698439; __asm__ volatile ( - "li.s %[fTmp1], 0.54119610014619698439 \n\t" "sub.s %[fTmp2], %[val0], %[val3] \n\t" "add.s %[val0], %[val0], %[val3] \n\t" "sub.s %[fTmp3], %[val7], %[val4] \n\t" "add.s %[val4], %[val7], %[val4] \n\t" "sub.s %[fTmp4], %[val8], %[val11] \n\t" - "mul.s %[val3], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val3], %[f1], %[fTmp2] \n\t" "add.s %[val8], %[val8], %[val11] \n\t" - "mul.s %[val7], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val7], %[f1], %[fTmp3] \n\t" "sub.s %[fTmp2], %[val15], %[val12] \n\t" - "mul.s %[val11], %[fTmp1], %[fTmp4] \n\t" + "mul.s %[val11], %[f1], %[fTmp4] \n\t" "add.s %[val12], %[val15], %[val12] \n\t" - "mul.s %[val15], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val15], %[f1], %[fTmp2] \n\t" - : [val0] "+f" (val0), [val3] "+f" (val3), - [val4] "+f" (val4), [val7] "+f" (val7), - [val8] "+f" (val8), [val11] "+f" (val11), - [val12] "+f" (val12), [val15] "+f" (val15), - [fTmp1] "=f" (fTmp1), [fTmp2] "=&f" (fTmp2), + : [val0] "+&f" (val0), [val3] "+&f" (val3), + [val4] "+&f" (val4), [val7] "+&f" (val7), + [val8] "+&f" (val8), [val11] "+&f" (val11), + [val12] "+&f" (val12), [val15] "+&f" (val15), + [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4) - : + : [f1] "f" (f1) ); __asm__ volatile ( @@ -449,169 +449,169 @@ static void ff_dct32_mips_float(float *out, const float *tab) "sub.s %[fTmp3], %[val23], %[val20] \n\t" "add.s %[val20], %[val23], %[val20] \n\t" "sub.s %[fTmp4], %[val24], %[val27] \n\t" - "mul.s %[val19], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val19], %[f1], %[fTmp2] \n\t" "add.s %[val24], %[val24], %[val27] \n\t" - "mul.s %[val23], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val23], %[f1], %[fTmp3] \n\t" "sub.s %[fTmp2], %[val31], %[val28] \n\t" - "mul.s %[val27], %[fTmp1], %[fTmp4] \n\t" + "mul.s %[val27], %[f1], %[fTmp4] \n\t" "add.s %[val28], %[val31], %[val28] \n\t" - "mul.s %[val31], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val31], %[f1], %[fTmp2] \n\t" : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), - [val16] "+f" (val16), [val19] "+f" (val19), [val20] "+f" (val20), - [val23] "+f" (val23), [val24] "+f" (val24), [val27] "+f" (val27), - [val28] "+f" (val28), [val31] "+f" (val31) - : [fTmp1] "f" (fTmp1) + [val16] "+&f" (val16), [val19] "+&f" (val19), [val20] "+&f" (val20), + [val23] "+&f" (val23), [val24] "+&f" (val24), [val27] "+&f" (val27), + [val28] "+&f" (val28), [val31] "+&f" (val31) + : [f1] "f" (f1) ); + f1 = 0.52249861493968888062; + f2 = 0.50547095989754365998; + f3 = 3.40760841846871878570; + f4 = 1.72244709823833392782; + f5 = 0.62250412303566481615; + f6 = 0.83934964541552703873; + f7 = 0.60134488693504528054; __asm__ volatile ( "lwc1 %[fTmp1], 1*4(%[tab]) \n\t" "lwc1 %[fTmp2], 30*4(%[tab]) \n\t" "lwc1 %[fTmp3], 14*4(%[tab]) \n\t" "lwc1 %[fTmp4], 17*4(%[tab]) \n\t" - "li.s %[fTmp7], 0.52249861493968888062 \n\t" "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t" "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t" "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t" "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t" - "li.s %[fTmp10], 0.50547095989754365998 \n\t" - "li.s %[fTmp11], 3.40760841846871878570 \n\t" - "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t" + "mul.s %[fTmp8], %[fTmp8], %[f2] \n\t" "add.s %[val1], %[fTmp5], %[fTmp6] \n\t" "sub.s %[val14], %[fTmp5], %[fTmp6] \n\t" "lwc1 %[fTmp1], 6*4(%[tab]) \n\t" "lwc1 %[fTmp2], 25*4(%[tab]) \n\t" - "madd.s %[val17], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "nmsub.s %[val30], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "mul.s %[val14], %[val14], %[fTmp7] \n\t" + "madd.s %[val17], %[fTmp8], %[fTmp9], %[f3] \n\t" + "nmsub.s %[val30], %[fTmp8], %[fTmp9], %[f3] \n\t" + "mul.s %[val14], %[val14], %[f1] \n\t" "lwc1 %[fTmp3], 9*4(%[tab]) \n\t" "lwc1 %[fTmp4], 22*4(%[tab]) \n\t" "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t" - "mul.s %[val30], %[val30], %[fTmp7] \n\t" + "mul.s %[val30], %[val30], %[f1] \n\t" "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t" "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t" "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t" - "li.s %[fTmp7], 1.72244709823833392782 \n\t" - "li.s %[fTmp10], 0.62250412303566481615 \n\t" - "li.s %[fTmp11], 0.83934964541552703873 \n\t" "add.s %[val6], %[fTmp5], %[fTmp6] \n\t" "sub.s %[val9], %[fTmp5], %[fTmp6] \n\t" - "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t" - "li.s %[fTmp1], 0.60134488693504528054 \n\t" + "mul.s %[fTmp8], %[fTmp8], %[f5] \n\t" "sub.s %[fTmp2], %[val1], %[val6] \n\t" "add.s %[val1], %[val1], %[val6] \n\t" - "mul.s %[val9], %[val9], %[fTmp7] \n\t" - "madd.s %[val22], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "nmsub.s %[val25], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "mul.s %[val6], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val9], %[val9], %[f4] \n\t" + "madd.s %[val22], %[fTmp8], %[fTmp9], %[f6] \n\t" + "nmsub.s %[val25], %[fTmp8], %[fTmp9], %[f6] \n\t" + "mul.s %[val6], %[f7], %[fTmp2] \n\t" "sub.s %[fTmp2], %[val14], %[val9] \n\t" "add.s %[val9], %[val14], %[val9] \n\t" - "mul.s %[val25], %[val25], %[fTmp7] \n\t" + "mul.s %[val25], %[val25], %[f4] \n\t" "sub.s %[fTmp3], %[val17], %[val22] \n\t" "add.s %[val17], %[val17], %[val22] \n\t" - "mul.s %[val14], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val14], %[f7], %[fTmp2] \n\t" "sub.s %[fTmp2], %[val30], %[val25] \n\t" - "mul.s %[val22], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val22], %[f7], %[fTmp3] \n\t" "add.s %[val25], %[val30], %[val25] \n\t" - "mul.s %[val30], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val30], %[f7], %[fTmp2] \n\t" : [fTmp1] "=&f" (fTmp1), [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), [fTmp5] "=&f" (fTmp5), [fTmp6] "=&f" (fTmp6), - [fTmp7] "=&f" (fTmp7), [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9), - [fTmp10] "=&f" (fTmp10), [fTmp11] "=&f" (fTmp11), - [val1] "=f" (val1), [val6] "=f" (val6), - [val9] "=f" (val9), [val14] "=f" (val14), - [val17] "=f" (val17), [val22] "=f" (val22), - [val25] "=f" (val25), [val30] "=f" (val30) - : [tab] "r" (tab) + [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9), + [val1] "=&f" (val1), [val6] "=&f" (val6), + [val9] "=&f" (val9), [val14] "=&f" (val14), + [val17] "=&f" (val17), [val22] "=&f" (val22), + [val25] "=&f" (val25), [val30] "=&f" (val30) + : [tab] "r" (tab), [f1]"f"(f1), [f2]"f"(f2), [f3]"f"(f3), + [f4]"f"(f4), [f5]"f"(f5), [f6]"f"(f6), [f7]"f"(f7) : "memory" ); + f1 = 0.56694403481635770368; + f2 = 0.51544730992262454697; + f3 = 2.05778100995341155085; + f4 = 1.06067768599034747134; + f5 = 0.58293496820613387367; + f6 = 0.97256823786196069369; + f7 = 0.89997622313641570463; __asm__ volatile ( "lwc1 %[fTmp1], 2*4(%[tab]) \n\t" "lwc1 %[fTmp2], 29*4(%[tab]) \n\t" "lwc1 %[fTmp3], 13*4(%[tab]) \n\t" "lwc1 %[fTmp4], 18*4(%[tab]) \n\t" - "li.s %[fTmp7], 0.56694403481635770368 \n\t" "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t" "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t" "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t" "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t" - "li.s %[fTmp10], 0.51544730992262454697 \n\t" - "li.s %[fTmp11], 2.05778100995341155085 \n\t" - "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t" + "mul.s %[fTmp8], %[fTmp8], %[f2] \n\t" "add.s %[val2], %[fTmp5], %[fTmp6] \n\t" "sub.s %[val13], %[fTmp5], %[fTmp6] \n\t" "lwc1 %[fTmp1], 5*4(%[tab]) \n\t" "lwc1 %[fTmp2], 26*4(%[tab]) \n\t" - "madd.s %[val18], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "nmsub.s %[val29], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "mul.s %[val13], %[val13], %[fTmp7] \n\t" + "madd.s %[val18], %[fTmp8], %[fTmp9], %[f3] \n\t" + "nmsub.s %[val29], %[fTmp8], %[fTmp9], %[f3] \n\t" + "mul.s %[val13], %[val13], %[f1] \n\t" "lwc1 %[fTmp3], 10*4(%[tab]) \n\t" "lwc1 %[fTmp4], 21*4(%[tab]) \n\t" - "mul.s %[val29], %[val29], %[fTmp7] \n\t" + "mul.s %[val29], %[val29], %[f1] \n\t" "add.s %[fTmp5], %[fTmp1], %[fTmp2] \n\t" "sub.s %[fTmp8], %[fTmp1], %[fTmp2] \n\t" "add.s %[fTmp6], %[fTmp3], %[fTmp4] \n\t" "sub.s %[fTmp9], %[fTmp3], %[fTmp4] \n\t" - "li.s %[fTmp7], 1.06067768599034747134 \n\t" - "li.s %[fTmp10], 0.58293496820613387367 \n\t" - "li.s %[fTmp11], 0.97256823786196069369 \n\t" "add.s %[val5], %[fTmp5], %[fTmp6] \n\t" "sub.s %[val10], %[fTmp5], %[fTmp6] \n\t" - "mul.s %[fTmp8], %[fTmp8], %[fTmp10] \n\t" - "li.s %[fTmp1], 0.89997622313641570463 \n\t" + "mul.s %[fTmp8], %[fTmp8], %[f5] \n\t" "sub.s %[fTmp2], %[val2], %[val5] \n\t" - "mul.s %[val10], %[val10], %[fTmp7] \n\t" - "madd.s %[val21], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" - "nmsub.s %[val26], %[fTmp8], %[fTmp9], %[fTmp11] \n\t" + "mul.s %[val10], %[val10], %[f4] \n\t" + "madd.s %[val21], %[fTmp8], %[fTmp9], %[f6] \n\t" + "nmsub.s %[val26], %[fTmp8], %[fTmp9], %[f6] \n\t" "add.s %[val2], %[val2], %[val5] \n\t" - "mul.s %[val5], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val5], %[f7], %[fTmp2] \n\t" "sub.s %[fTmp3], %[val13], %[val10] \n\t" "add.s %[val10], %[val13], %[val10] \n\t" - "mul.s %[val26], %[val26], %[fTmp7] \n\t" + "mul.s %[val26], %[val26], %[f4] \n\t" "sub.s %[fTmp4], %[val18], %[val21] \n\t" "add.s %[val18], %[val18], %[val21] \n\t" - "mul.s %[val13], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val13], %[f7], %[fTmp3] \n\t" "sub.s %[fTmp2], %[val29], %[val26] \n\t" "add.s %[val26], %[val29], %[val26] \n\t" - "mul.s %[val21], %[fTmp1], %[fTmp4] \n\t" - "mul.s %[val29], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val21], %[f7], %[fTmp4] \n\t" + "mul.s %[val29], %[f7], %[fTmp2] \n\t" : [fTmp1] "=&f" (fTmp1), [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), [fTmp5] "=&f" (fTmp5), [fTmp6] "=&f" (fTmp6), - [fTmp7] "=&f" (fTmp7), [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9), - [fTmp10] "=&f" (fTmp10), [fTmp11] "=&f" (fTmp11), - [val2] "=f" (val2), [val5] "=f" (val5), - [val10] "=f" (val10), [val13] "=f" (val13), - [val18] "=f" (val18), [val21] "=f" (val21), - [val26] "=f" (val26), [val29] "=f" (val29) - : [tab] "r" (tab) + [fTmp8] "=&f" (fTmp8), [fTmp9] "=&f" (fTmp9), + [val2] "=&f" (val2), [val5] "=&f" (val5), + [val10] "=&f" (val10), [val13] "=&f" (val13), + [val18] "=&f" (val18), [val21] "=&f" (val21), + [val26] "=&f" (val26), [val29] "=&f" (val29) + : [tab] "r" (tab), [f1]"f"(f1), [f2]"f"(f2), [f3]"f"(f3), + [f4]"f"(f4), [f5]"f"(f5), [f6]"f"(f6), [f7]"f"(f7) : "memory" ); + f1 = 1.30656296487637652785; __asm__ volatile ( - "li.s %[fTmp1], 1.30656296487637652785 \n\t" "sub.s %[fTmp2], %[val1], %[val2] \n\t" "add.s %[val1], %[val1], %[val2] \n\t" "sub.s %[fTmp3], %[val6], %[val5] \n\t" "add.s %[val5], %[val6], %[val5] \n\t" "sub.s %[fTmp4], %[val9], %[val10] \n\t" - "mul.s %[val2], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val2], %[f1], %[fTmp2] \n\t" "add.s %[val9], %[val9], %[val10] \n\t" - "mul.s %[val6], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val6], %[f1], %[fTmp3] \n\t" "sub.s %[fTmp2], %[val14], %[val13] \n\t" - "mul.s %[val10], %[fTmp1], %[fTmp4] \n\t" + "mul.s %[val10], %[f1], %[fTmp4] \n\t" "add.s %[val13], %[val14], %[val13] \n\t" - "mul.s %[val14], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val14], %[f1], %[fTmp2] \n\t" - : [fTmp1] "=f" (fTmp1), [fTmp2] "=&f" (fTmp2), + : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), - [val1] "+f" (val1), [val2] "+f" (val2), - [val5] "+f" (val5), [val6] "+f" (val6), - [val9] "+f" (val9), [val10] "+f" (val10), - [val13] "+f" (val13), [val14] "+f" (val14) - : + [val1] "+&f" (val1), [val2] "+&f" (val2), + [val5] "+&f" (val5), [val6] "+&f" (val6), + [val9] "+&f" (val9), [val10] "+&f" (val10), + [val13] "+&f" (val13), [val14] "+&f" (val14) + : [f1]"f"(f1) ); __asm__ volatile ( @@ -620,39 +620,39 @@ static void ff_dct32_mips_float(float *out, const float *tab) "sub.s %[fTmp3], %[val22], %[val21] \n\t" "add.s %[val21], %[val22], %[val21] \n\t" "sub.s %[fTmp4], %[val25], %[val26] \n\t" - "mul.s %[val18], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val18], %[f1], %[fTmp2] \n\t" "add.s %[val25], %[val25], %[val26] \n\t" - "mul.s %[val22], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val22], %[f1], %[fTmp3] \n\t" "sub.s %[fTmp2], %[val30], %[val29] \n\t" - "mul.s %[val26], %[fTmp1], %[fTmp4] \n\t" + "mul.s %[val26], %[f1], %[fTmp4] \n\t" "add.s %[val29], %[val30], %[val29] \n\t" - "mul.s %[val30], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val30], %[f1], %[fTmp2] \n\t" : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), - [val17] "+f" (val17), [val18] "+f" (val18), [val21] "+f" (val21), - [val22] "+f" (val22), [val25] "+f" (val25), [val26] "+f" (val26), - [val29] "+f" (val29), [val30] "+f" (val30) - : [fTmp1] "f" (fTmp1) + [val17] "+&f" (val17), [val18] "+&f" (val18), [val21] "+&f" (val21), + [val22] "+&f" (val22), [val25] "+&f" (val25), [val26] "+&f" (val26), + [val29] "+&f" (val29), [val30] "+&f" (val30) + : [f1] "f" (f1) ); + f1 = 0.70710678118654752439; __asm__ volatile ( - "li.s %[fTmp1], 0.70710678118654752439 \n\t" "sub.s %[fTmp2], %[val0], %[val1] \n\t" "add.s %[val0], %[val0], %[val1] \n\t" "sub.s %[fTmp3], %[val3], %[val2] \n\t" "add.s %[val2], %[val3], %[val2] \n\t" "sub.s %[fTmp4], %[val4], %[val5] \n\t" - "mul.s %[val1], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val1], %[f1], %[fTmp2] \n\t" "swc1 %[val0], 0(%[out]) \n\t" - "mul.s %[val3], %[fTmp3], %[fTmp1] \n\t" + "mul.s %[val3], %[fTmp3], %[f1] \n\t" "add.s %[val4], %[val4], %[val5] \n\t" - "mul.s %[val5], %[fTmp1], %[fTmp4] \n\t" + "mul.s %[val5], %[f1], %[fTmp4] \n\t" "swc1 %[val1], 16*4(%[out]) \n\t" "sub.s %[fTmp2], %[val7], %[val6] \n\t" "add.s %[val2], %[val2], %[val3] \n\t" "swc1 %[val3], 24*4(%[out]) \n\t" "add.s %[val6], %[val7], %[val6] \n\t" - "mul.s %[val7], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val7], %[f1], %[fTmp2] \n\t" "swc1 %[val2], 8*4(%[out]) \n\t" "add.s %[val6], %[val6], %[val7] \n\t" "swc1 %[val7], 28*4(%[out]) \n\t" @@ -663,13 +663,13 @@ static void ff_dct32_mips_float(float *out, const float *tab) "swc1 %[val5], 20*4(%[out]) \n\t" "swc1 %[val6], 12*4(%[out]) \n\t" - : [fTmp1] "=f" (fTmp1), [fTmp2] "=&f" (fTmp2), + : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), - [val0] "+f" (val0), [val1] "+f" (val1), - [val2] "+f" (val2), [val3] "+f" (val3), - [val4] "+f" (val4), [val5] "+f" (val5), - [val6] "+f" (val6), [val7] "+f" (val7) - : [out] "r" (out) + [val0] "+&f" (val0), [val1] "+&f" (val1), + [val2] "+&f" (val2), [val3] "+&f" (val3), + [val4] "+&f" (val4), [val5] "+&f" (val5), + [val6] "+&f" (val6), [val7] "+&f" (val7) + : [out] "r" (out), [f1]"f"(f1) ); __asm__ volatile ( @@ -678,14 +678,14 @@ static void ff_dct32_mips_float(float *out, const float *tab) "sub.s %[fTmp3], %[val11], %[val10] \n\t" "add.s %[val10], %[val11], %[val10] \n\t" "sub.s %[fTmp4], %[val12], %[val13] \n\t" - "mul.s %[val9], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val9], %[f1], %[fTmp2] \n\t" "add.s %[val12], %[val12], %[val13] \n\t" - "mul.s %[val11], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val11], %[f1], %[fTmp3] \n\t" "sub.s %[fTmp2], %[val15], %[val14] \n\t" - "mul.s %[val13], %[fTmp1], %[fTmp4] \n\t" + "mul.s %[val13], %[f1], %[fTmp4] \n\t" "add.s %[val14], %[val15], %[val14] \n\t" "add.s %[val10], %[val10], %[val11] \n\t" - "mul.s %[val15], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val15], %[f1], %[fTmp2] \n\t" "add.s %[val14], %[val14], %[val15] \n\t" "add.s %[val12], %[val12], %[val14] \n\t" "add.s %[val14], %[val14], %[val13] \n\t" @@ -707,10 +707,10 @@ static void ff_dct32_mips_float(float *out, const float *tab) "swc1 %[val15], 30*4(%[out]) \n\t" : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), - [val8] "+f" (val8), [val9] "+f" (val9), [val10] "+f" (val10), - [val11] "+f" (val11), [val12] "+f" (val12), [val13] "+f" (val13), - [val14] "+f" (val14), [val15] "+f" (val15) - : [fTmp1] "f" (fTmp1), [out] "r" (out) + [val8] "+&f" (val8), [val9] "+&f" (val9), [val10] "+&f" (val10), + [val11] "+&f" (val11), [val12] "+&f" (val12), [val13] "+&f" (val13), + [val14] "+&f" (val14), [val15] "+&f" (val15) + : [f1] "f" (f1), [out] "r" (out) ); __asm__ volatile ( @@ -719,24 +719,24 @@ static void ff_dct32_mips_float(float *out, const float *tab) "sub.s %[fTmp3], %[val19], %[val18] \n\t" "add.s %[val18], %[val19], %[val18] \n\t" "sub.s %[fTmp4], %[val20], %[val21] \n\t" - "mul.s %[val17], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val17], %[f1], %[fTmp2] \n\t" "add.s %[val20], %[val20], %[val21] \n\t" - "mul.s %[val19], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val19], %[f1], %[fTmp3] \n\t" "sub.s %[fTmp2], %[val23], %[val22] \n\t" - "mul.s %[val21], %[fTmp1], %[fTmp4] \n\t" + "mul.s %[val21], %[f1], %[fTmp4] \n\t" "add.s %[val22], %[val23], %[val22] \n\t" "add.s %[val18], %[val18], %[val19] \n\t" - "mul.s %[val23], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val23], %[f1], %[fTmp2] \n\t" "add.s %[val22], %[val22], %[val23] \n\t" "add.s %[val20], %[val20], %[val22] \n\t" "add.s %[val22], %[val22], %[val21] \n\t" "add.s %[val21], %[val21], %[val23] \n\t" : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), - [val16] "+f" (val16), [val17] "+f" (val17), [val18] "+f" (val18), - [val19] "+f" (val19), [val20] "+f" (val20), [val21] "+f" (val21), - [val22] "+f" (val22), [val23] "+f" (val23) - : [fTmp1] "f" (fTmp1) + [val16] "+&f" (val16), [val17] "+&f" (val17), [val18] "+&f" (val18), + [val19] "+&f" (val19), [val20] "+&f" (val20), [val21] "+&f" (val21), + [val22] "+&f" (val22), [val23] "+&f" (val23) + : [f1] "f" (f1) ); __asm__ volatile ( @@ -745,14 +745,14 @@ static void ff_dct32_mips_float(float *out, const float *tab) "sub.s %[fTmp3], %[val27], %[val26] \n\t" "add.s %[val26], %[val27], %[val26] \n\t" "sub.s %[fTmp4], %[val28], %[val29] \n\t" - "mul.s %[val25], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val25], %[f1], %[fTmp2] \n\t" "add.s %[val28], %[val28], %[val29] \n\t" - "mul.s %[val27], %[fTmp1], %[fTmp3] \n\t" + "mul.s %[val27], %[f1], %[fTmp3] \n\t" "sub.s %[fTmp2], %[val31], %[val30] \n\t" - "mul.s %[val29], %[fTmp1], %[fTmp4] \n\t" + "mul.s %[val29], %[f1], %[fTmp4] \n\t" "add.s %[val30], %[val31], %[val30] \n\t" "add.s %[val26], %[val26], %[val27] \n\t" - "mul.s %[val31], %[fTmp1], %[fTmp2] \n\t" + "mul.s %[val31], %[f1], %[fTmp2] \n\t" "add.s %[val30], %[val30], %[val31] \n\t" "add.s %[val28], %[val28], %[val30] \n\t" "add.s %[val30], %[val30], %[val29] \n\t" @@ -766,10 +766,10 @@ static void ff_dct32_mips_float(float *out, const float *tab) "add.s %[val27], %[val27], %[val31] \n\t" : [fTmp2] "=&f" (fTmp2), [fTmp3] "=&f" (fTmp3), [fTmp4] "=&f" (fTmp4), - [val24] "+f" (val24), [val25] "+f" (val25), [val26] "+f" (val26), - [val27] "+f" (val27), [val28] "+f" (val28), [val29] "+f" (val29), - [val30] "+f" (val30), [val31] "+f" (val31) - : [fTmp1] "f" (fTmp1) + [val24] "+&f" (val24), [val25] "+&f" (val25), [val26] "+&f" (val26), + [val27] "+&f" (val27), [val28] "+&f" (val28), [val29] "+&f" (val29), + [val30] "+&f" (val30), [val31] "+&f" (val31) + : [f1] "f" (f1) ); out[ 1] = val16 + val24; @@ -797,7 +797,7 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) /* temporary variables */ float in1, in2, in3, in4, in5, in6; float out1, out2, out3, out4, out5; - float c1, c2, c3, c4, c5, c6, c7, c8, c9; + float f1, f2, f3, f4, f5, f6, f7, f8, f9; /** * all loops are unrolled totally, and instructions are scheduled to @@ -881,33 +881,36 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) ); /* loop 3 */ + f1 = 0.5; + f2 = 0.93969262078590838405; + f3 = -0.76604444311897803520; + f4 = -0.17364817766693034885; + f5 = -0.86602540378443864676; + f6 = 0.98480775301220805936; + f7 = -0.34202014332566873304; + f8 = 0.86602540378443864676; + f9 = -0.64278760968653932632; __asm__ volatile ( - "li.s %[c1], 0.5 \t\n" "lwc1 %[in1], 8*4(%[in]) \t\n" "lwc1 %[in2], 16*4(%[in]) \t\n" "lwc1 %[in3], 4*4(%[in]) \t\n" "lwc1 %[in4], 0(%[in]) \t\n" "lwc1 %[in5], 12*4(%[in]) \t\n" - "li.s %[c2], 0.93969262078590838405 \t\n" "add.s %[t2], %[in1], %[in2] \t\n" "add.s %[t0], %[in1], %[in3] \t\n" - "li.s %[c3], -0.76604444311897803520 \t\n" - "madd.s %[t3], %[in4], %[in5], %[c1] \t\n" + "madd.s %[t3], %[in4], %[in5], %[f1] \t\n" "sub.s %[t1], %[in4], %[in5] \t\n" "sub.s %[t2], %[t2], %[in3] \t\n" - "mul.s %[t0], %[t0], %[c2] \t\n" - "li.s %[c4], -0.17364817766693034885 \t\n" - "li.s %[c5], -0.86602540378443864676 \t\n" - "li.s %[c6], 0.98480775301220805936 \t\n" - "nmsub.s %[out1], %[t1], %[t2], %[c1] \t\n" + "mul.s %[t0], %[t0], %[f2] \t\n" + "nmsub.s %[out1], %[t1], %[t2], %[f1] \t\n" "add.s %[out2], %[t1], %[t2] \t\n" "add.s %[t2], %[in2], %[in3] \t\n" "sub.s %[t1], %[in1], %[in2] \t\n" "sub.s %[out3], %[t3], %[t0] \t\n" "swc1 %[out1], 6*4(%[tmp]) \t\n" "swc1 %[out2], 16*4(%[tmp]) \t\n" - "mul.s %[t2], %[t2], %[c3] \t\n" - "mul.s %[t1], %[t1], %[c4] \t\n" + "mul.s %[t2], %[t2], %[f3] \t\n" + "mul.s %[t1], %[t1], %[f4] \t\n" "add.s %[out1], %[t3], %[t0] \t\n" "lwc1 %[in1], 10*4(%[in]) \t\n" "lwc1 %[in2], 14*4(%[in]) \t\n" @@ -923,19 +926,16 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "add.s %[t2], %[in1], %[in3] \t\n" "sub.s %[t3], %[in1], %[in2] \t\n" "swc1 %[out2], 14*4(%[tmp]) \t\n" - "li.s %[c7], -0.34202014332566873304 \t\n" "sub.s %[out1], %[out1], %[in3] \t\n" - "mul.s %[t2], %[t2], %[c6] \t\n" - "mul.s %[t3], %[t3], %[c7] \t\n" - "li.s %[c8], 0.86602540378443864676 \t\n" - "mul.s %[t0], %[in4], %[c8] \t\n" - "mul.s %[out1], %[out1], %[c5] \t\n" + "mul.s %[t2], %[t2], %[f6] \t\n" + "mul.s %[t3], %[t3], %[f7] \t\n" + "mul.s %[t0], %[in4], %[f8] \t\n" + "mul.s %[out1], %[out1], %[f5] \t\n" "add.s %[t1], %[in2], %[in3] \t\n" - "li.s %[c9], -0.64278760968653932632 \t\n" "add.s %[out2], %[t2], %[t3] \t\n" "lwc1 %[in1], 9*4(%[in]) \t\n" "swc1 %[out1], 4*4(%[tmp]) \t\n" - "mul.s %[t1], %[t1], %[c9] \t\n" + "mul.s %[t1], %[t1], %[f9] \t\n" "lwc1 %[in2], 17*4(%[in]) \t\n" "add.s %[out2], %[out2], %[t0] \t\n" "lwc1 %[in3], 5*4(%[in]) \t\n" @@ -948,21 +948,21 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "sub.s %[out3], %[out3], %[t0] \t\n" "sub.s %[out1], %[out1], %[t0] \t\n" "add.s %[t0], %[in1], %[in3] \t\n" - "madd.s %[t3], %[in4], %[in5], %[c1] \t\n" + "madd.s %[t3], %[in4], %[in5], %[f1] \t\n" "sub.s %[t2], %[t2], %[in3] \t\n" "swc1 %[out3], 12*4(%[tmp]) \t\n" "swc1 %[out1], 8*4(%[tmp]) \t\n" "sub.s %[t1], %[in4], %[in5] \t\n" - "mul.s %[t0], %[t0], %[c2] \t\n" - "nmsub.s %[out1], %[t1], %[t2], %[c1] \t\n" + "mul.s %[t0], %[t0], %[f2] \t\n" + "nmsub.s %[out1], %[t1], %[t2], %[f1] \t\n" "add.s %[out2], %[t1], %[t2] \t\n" "add.s %[t2], %[in2], %[in3] \t\n" "sub.s %[t1], %[in1], %[in2] \t\n" "sub.s %[out3], %[t3], %[t0] \t\n" "swc1 %[out1], 7*4(%[tmp]) \t\n" "swc1 %[out2], 17*4(%[tmp]) \t\n" - "mul.s %[t2], %[t2], %[c3] \t\n" - "mul.s %[t1], %[t1], %[c4] \t\n" + "mul.s %[t2], %[t2], %[f3] \t\n" + "mul.s %[t1], %[t1], %[f4] \t\n" "add.s %[out1], %[t3], %[t0] \t\n" "lwc1 %[in1], 11*4(%[in]) \t\n" "lwc1 %[in2], 15*4(%[in]) \t\n" @@ -978,14 +978,14 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "add.s %[t2], %[in1], %[in3] \t\n" "sub.s %[t3], %[in1], %[in2] \t\n" "swc1 %[out2], 15*4(%[tmp]) \t\n" - "mul.s %[t0], %[in4], %[c8] \t\n" + "mul.s %[t0], %[in4], %[f8] \t\n" "sub.s %[out3], %[out3], %[in3] \t\n" - "mul.s %[t2], %[t2], %[c6] \t\n" - "mul.s %[t3], %[t3], %[c7] \t\n" + "mul.s %[t2], %[t2], %[f6] \t\n" + "mul.s %[t3], %[t3], %[f7] \t\n" "add.s %[t1], %[in2], %[in3] \t\n" - "mul.s %[out3], %[out3], %[c5] \t\n" + "mul.s %[out3], %[out3], %[f5] \t\n" "add.s %[out1], %[t2], %[t3] \t\n" - "mul.s %[t1], %[t1], %[c9] \t\n" + "mul.s %[t1], %[t1], %[f9] \t\n" "swc1 %[out3], 5*4(%[tmp]) \t\n" "add.s %[out1], %[out1], %[t0] \t\n" "add.s %[out2], %[t2], %[t1] \t\n" @@ -1000,26 +1000,29 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) [t2] "=&f" (t2), [t3] "=&f" (t3), [in1] "=&f" (in1), [in2] "=&f" (in2), [in3] "=&f" (in3), [in4] "=&f" (in4), - [in5] "=&f" (in5), - [out1] "=&f" (out1), [out2] "=&f" (out2), - [out3] "=&f" (out3), - [c1] "=&f" (c1), [c2] "=&f" (c2), - [c3] "=&f" (c3), [c4] "=&f" (c4), - [c5] "=&f" (c5), [c6] "=&f" (c6), - [c7] "=&f" (c7), [c8] "=&f" (c8), - [c9] "=&f" (c9) - : [in] "r" (in), [tmp] "r" (tmp) + [in5] "=&f" (in5), [out1] "=&f" (out1), + [out2] "=&f" (out2), [out3] "=&f" (out3) + : [in] "r" (in), [tmp] "r" (tmp), [f1]"f"(f1), [f2]"f"(f2), + [f3]"f"(f3), [f4]"f"(f4), [f5]"f"(f5), [f6]"f"(f6), + [f7]"f"(f7), [f8]"f"(f8), [f9]"f"(f9) : "memory" ); /* loop 4 */ + f1 = 0.50190991877167369479; + f2 = 5.73685662283492756461; + f3 = 0.51763809020504152469; + f4 = 1.93185165257813657349; + f5 = 0.55168895948124587824; + f6 = 1.18310079157624925896; + f7 = 0.61038729438072803416; + f8 = 0.87172339781054900991; + f9 = 0.70710678118654752439; __asm__ volatile ( "lwc1 %[in1], 2*4(%[tmp]) \t\n" "lwc1 %[in2], 0(%[tmp]) \t\n" "lwc1 %[in3], 3*4(%[tmp]) \t\n" "lwc1 %[in4], 1*4(%[tmp]) \t\n" - "li.s %[c1], 0.50190991877167369479 \t\n" - "li.s %[c2], 5.73685662283492756461 \t\n" "add.s %[s0], %[in1], %[in2] \t\n" "sub.s %[s2], %[in1], %[in2] \t\n" "add.s %[s1], %[in3], %[in4] \t\n" @@ -1027,15 +1030,13 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "lwc1 %[in1], 9*4(%[win]) \t\n" "lwc1 %[in2], 4*9*4(%[buf]) \t\n" "lwc1 %[in3], 8*4(%[win]) \t\n" - "mul.s %[s1], %[s1], %[c1] \t\n" - "mul.s %[s3], %[s3], %[c2] \t\n" + "mul.s %[s1], %[s1], %[f1] \t\n" + "mul.s %[s3], %[s3], %[f2] \t\n" "lwc1 %[in4], 4*8*4(%[buf]) \t\n" "lwc1 %[in5], 29*4(%[win]) \t\n" "lwc1 %[in6], 28*4(%[win]) \t\n" "add.s %[t0], %[s0], %[s1] \t\n" "sub.s %[t1], %[s0], %[s1] \t\n" - "li.s %[c1], 0.51763809020504152469 \t\n" - "li.s %[c2], 1.93185165257813657349 \t\n" "mul.s %[out3], %[in5], %[t0] \t\n" "madd.s %[out1], %[in2], %[in1], %[t1] \t\n" "madd.s %[out2], %[in4], %[in3], %[t1] \t\n" @@ -1071,14 +1072,13 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "lwc1 %[in1], 10*4(%[win]) \t\n" "lwc1 %[in2], 4*10*4(%[buf]) \t\n" "lwc1 %[in3], 7*4(%[win]) \t\n" - "mul.s %[s1], %[s1], %[c1] \t\n" - "mul.s %[s3], %[s3], %[c2] \t\n" + "mul.s %[s1], %[s1], %[f3] \t\n" + "mul.s %[s3], %[s3], %[f4] \t\n" "add.s %[t0], %[s0], %[s1] \t\n" "sub.s %[t1], %[s0], %[s1] \t\n" "lwc1 %[in4], 4*7*4(%[buf]) \t\n" "lwc1 %[in5], 30*4(%[win]) \t\n" "lwc1 %[in6], 27*4(%[win]) \t\n" - "li.s %[c1], 0.55168895948124587824 \t\n" "madd.s %[out1], %[in2], %[in1], %[t1] \t\n" "madd.s %[out2], %[in4], %[in3], %[t1] \t\n" "mul.s %[out3], %[t0], %[in5] \t\n" @@ -1105,7 +1105,6 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "swc1 %[out2], 32*4(%[out]) \t\n" "swc1 %[out3], 4*16*4(%[buf]) \t\n" "swc1 %[out4], 4*1*4(%[buf]) \t\n" - "li.s %[c2], 1.18310079157624925896 \t\n" "add.s %[s0], %[in1], %[in2] \t\n" "sub.s %[s2], %[in1], %[in2] \t\n" "lwc1 %[in3], 11*4(%[tmp]) \t\n" @@ -1115,8 +1114,8 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "lwc1 %[in1], 11*4(%[win]) \t\n" "lwc1 %[in2], 4*11*4(%[buf]) \t\n" "lwc1 %[in3], 6*4(%[win]) \t\n" - "mul.s %[s1], %[s1], %[c1] \t\n" - "mul.s %[s3], %[s3], %[c2] \t\n" + "mul.s %[s1], %[s1], %[f5] \t\n" + "mul.s %[s3], %[s3], %[f6] \t\n" "lwc1 %[in4], 4*6*4(%[buf]) \t\n" "lwc1 %[in5], 31*4(%[win]) \t\n" "lwc1 %[in6], 26*4(%[win]) \t\n" @@ -1152,15 +1151,13 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "add.s %[s0], %[in1], %[in2] \t\n" "sub.s %[s2], %[in1], %[in2] \t\n" "lwc1 %[in4], 13*4(%[tmp]) \t\n" - "li.s %[c1], 0.61038729438072803416 \t\n" - "li.s %[c2], 0.87172339781054900991 \t\n" "add.s %[s1], %[in3], %[in4] \t\n" "sub.s %[s3], %[in3], %[in4] \t\n" "lwc1 %[in1], 12*4(%[win]) \t\n" "lwc1 %[in2], 4*12*4(%[buf]) \t\n" "lwc1 %[in3], 5*4(%[win]) \t\n" - "mul.s %[s1], %[s1], %[c1] \t\n" - "mul.s %[s3], %[s3], %[c2] \t\n" + "mul.s %[s1], %[s1], %[f7] \t\n" + "mul.s %[s3], %[s3], %[f8] \t\n" "lwc1 %[in4], 4*5*4(%[buf]) \t\n" "lwc1 %[in5], 32*4(%[win]) \t\n" "lwc1 %[in6], 25*4(%[win]) \t\n" @@ -1168,7 +1165,6 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "sub.s %[t1], %[s0], %[s1] \t\n" "lwc1 %[s0], 16*4(%[tmp]) \t\n" "lwc1 %[s1], 17*4(%[tmp]) \t\n" - "li.s %[c1], 0.70710678118654752439 \t\n" "mul.s %[out3], %[t0], %[in5] \t\n" "madd.s %[out1], %[in2], %[in1], %[t1] \t\n" "madd.s %[out2], %[in4], %[in3], %[t1] \t\n" @@ -1186,7 +1182,7 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "lwc1 %[in5], 34*4(%[win]) \t\n" "lwc1 %[in6], 23*4(%[win]) \t\n" "madd.s %[out1], %[in2], %[in1], %[t1] \t\n" - "mul.s %[s1], %[s1], %[c1] \t\n" + "mul.s %[s1], %[s1], %[f9] \t\n" "madd.s %[out2], %[in4], %[in3], %[t1] \t\n" "mul.s %[out3], %[in5], %[t0] \t\n" "mul.s %[out4], %[in6], %[t0] \t\n" @@ -1211,18 +1207,18 @@ static void imdct36_mips_float(float *out, float *buf, float *in, float *win) "swc1 %[out3], 4*13*4(%[buf]) \t\n" "swc1 %[out4], 4*4*4(%[buf]) \t\n" - : [c1] "=&f" (c1), [c2] "=&f" (c2), - [in1] "=&f" (in1), [in2] "=&f" (in2), + : [in1] "=&f" (in1), [in2] "=&f" (in2), [in3] "=&f" (in3), [in4] "=&f" (in4), [in5] "=&f" (in5), [in6] "=&f" (in6), [out1] "=&f" (out1), [out2] "=&f" (out2), [out3] "=&f" (out3), [out4] "=&f" (out4), [t0] "=&f" (t0), [t1] "=&f" (t1), - [t2] "=&f" (t2), [t3] "=&f" (t3), [s0] "=&f" (s0), [s1] "=&f" (s1), [s2] "=&f" (s2), [s3] "=&f" (s3) : [tmp] "r" (tmp), [win] "r" (win), - [buf] "r" (buf), [out] "r" (out) + [buf] "r" (buf), [out] "r" (out), + [f1]"f"(f1), [f2]"f"(f2), [f3]"f"(f3), [f4]"f"(f4), + [f5]"f"(f5), [f6]"f"(f6), [f7]"f"(f7), [f8]"f"(f8), [f9]"f"(f9) : "memory" ); }

1 0

avcodec/mips/cabac: Fix a bug in get_cabac_inline_mips.
by Shiyou Yin 29 Jul '20

29 Jul '20

ffmpeg | branch: master | Shiyou Yin <yinshiyou-hf(a)loongson.cn> | Wed Jul 29 18:11:00 2020 +0800| [1563b4b4c615e2264c2d302b97b1d507e153255c] | committer: Michael Niedermayer avcodec/mips/cabac: Fix a bug in get_cabac_inline_mips. Failed fate case: fate-h264-conformance-caba2_sony_e Clang is more strict in the use of register constraint. Signed-off-by: Michael Niedermayer <michael(a)niedermayer.cc> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=1563b4b4c615e2264… --- libavcodec/mips/cabac.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libavcodec/mips/cabac.h b/libavcodec/mips/cabac.h index c595915eda..3d09e93523 100644 --- a/libavcodec/mips/cabac.h +++ b/libavcodec/mips/cabac.h @@ -109,7 +109,7 @@ static av_always_inline int get_cabac_inline_mips(CABACContext *c, [lps_off]"i"(H264_LPS_RANGE_OFFSET), [mlps_off]"i"(H264_MLPS_STATE_OFFSET + 128), [norm_off]"i"(H264_NORM_SHIFT_OFFSET), - [cabac_mask]"i"(CABAC_MASK) + [cabac_mask]"r"(CABAC_MASK) : "memory" );

1 0