Merge commit '8e134e5104e99a69cd4cea10540a7ce9c3682a2c'

* commit '8e134e5104e99a69cd4cea10540a7ce9c3682a2c':
  lavc: clarify get_buffer() documentation
  mpegaudiodec: use planar sample format for output unless packed is requested
  x86: h264 qpel: use the correct number of utilized xmm regs in cglobal

Merged-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
Michael Niedermayer 2012-11-26 14:24:14 +01:00
commit a13148f633
4 changed files with 73 additions and 55 deletions

View File

@ -2406,7 +2406,12 @@ typedef struct AVCodecContext {
* *
* Decoders cannot use the buffer after returning from * Decoders cannot use the buffer after returning from
* avcodec_decode_audio4(), so they will not call release_buffer(), as it * avcodec_decode_audio4(), so they will not call release_buffer(), as it
* is assumed to be released immediately upon return. * is assumed to be released immediately upon return. In some rare cases,
* a decoder may need to call get_buffer() more than once in a single
* call to avcodec_decode_audio4(). In that case, when get_buffer() is
* called again after it has already been called once, the previously
* acquired buffer is assumed to be released at that time and may not be
* reused by the decoder.
* *
* As a convenience, av_samples_get_buffer_size() and * As a convenience, av_samples_get_buffer_size() and
* av_samples_fill_arrays() in libavutil may be used by custom get_buffer() * av_samples_fill_arrays() in libavutil may be used by custom get_buffer()

View File

@ -95,7 +95,8 @@ typedef struct MPADecodeContext {
# define MULH3(x, y, s) ((s)*(y)*(x)) # define MULH3(x, y, s) ((s)*(y)*(x))
# define MULLx(x, y, s) ((y)*(x)) # define MULLx(x, y, s) ((y)*(x))
# define RENAME(a) a ## _float # define RENAME(a) a ## _float
# define OUT_FMT AV_SAMPLE_FMT_FLT # define OUT_FMT AV_SAMPLE_FMT_FLT
# define OUT_FMT_P AV_SAMPLE_FMT_FLTP
#else #else
# define SHR(a,b) ((a)>>(b)) # define SHR(a,b) ((a)>>(b))
/* WARNING: only correct for positive numbers */ /* WARNING: only correct for positive numbers */
@ -105,7 +106,8 @@ typedef struct MPADecodeContext {
# define MULH3(x, y, s) MULH((s)*(x), y) # define MULH3(x, y, s) MULH((s)*(x), y)
# define MULLx(x, y, s) MULL(x,y,s) # define MULLx(x, y, s) MULL(x,y,s)
# define RENAME(a) a ## _fixed # define RENAME(a) a ## _fixed
# define OUT_FMT AV_SAMPLE_FMT_S16 # define OUT_FMT AV_SAMPLE_FMT_S16
# define OUT_FMT_P AV_SAMPLE_FMT_S16P
#endif #endif
/****************/ /****************/
@ -441,7 +443,11 @@ static av_cold int decode_init(AVCodecContext * avctx)
ff_mpadsp_init(&s->mpadsp); ff_mpadsp_init(&s->mpadsp);
ff_dsputil_init(&s->dsp, avctx); ff_dsputil_init(&s->dsp, avctx);
avctx->sample_fmt= OUT_FMT; if (avctx->request_sample_fmt == OUT_FMT &&
avctx->codec_id != AV_CODEC_ID_MP3ON4)
avctx->sample_fmt = OUT_FMT;
else
avctx->sample_fmt = OUT_FMT_P;
s->err_recognition = avctx->err_recognition; s->err_recognition = avctx->err_recognition;
if (avctx->codec_id == AV_CODEC_ID_MP3ADU) if (avctx->codec_id == AV_CODEC_ID_MP3ADU)
@ -1564,7 +1570,7 @@ static int mp_decode_layer3(MPADecodeContext *s)
return nb_granules * 18; return nb_granules * 18;
} }
static int mp_decode_frame(MPADecodeContext *s, OUT_INT *samples, static int mp_decode_frame(MPADecodeContext *s, OUT_INT **samples,
const uint8_t *buf, int buf_size) const uint8_t *buf, int buf_size)
{ {
int i, nb_frames, ch, ret; int i, nb_frames, ch, ret;
@ -1627,20 +1633,26 @@ static int mp_decode_frame(MPADecodeContext *s, OUT_INT *samples,
av_log(s->avctx, AV_LOG_ERROR, "get_buffer() failed\n"); av_log(s->avctx, AV_LOG_ERROR, "get_buffer() failed\n");
return ret; return ret;
} }
samples = (OUT_INT *)s->frame.data[0]; samples = (OUT_INT **)s->frame.extended_data;
} }
/* apply the synthesis filter */ /* apply the synthesis filter */
for (ch = 0; ch < s->nb_channels; ch++) { for (ch = 0; ch < s->nb_channels; ch++) {
samples_ptr = samples + ch; int sample_stride;
if (s->avctx->sample_fmt == OUT_FMT_P) {
samples_ptr = samples[ch];
sample_stride = 1;
} else {
samples_ptr = samples[0] + ch;
sample_stride = s->nb_channels;
}
for (i = 0; i < nb_frames; i++) { for (i = 0; i < nb_frames; i++) {
RENAME(ff_mpa_synth_filter)( RENAME(ff_mpa_synth_filter)(&s->mpadsp, s->synth_buf[ch],
&s->mpadsp, &(s->synth_buf_offset[ch]),
s->synth_buf[ch], &(s->synth_buf_offset[ch]), RENAME(ff_mpa_synth_window),
RENAME(ff_mpa_synth_window), &s->dither_state, &s->dither_state, samples_ptr,
samples_ptr, s->nb_channels, sample_stride, s->sb_samples[ch][i]);
s->sb_samples[ch][i]); samples_ptr += 32 * sample_stride;
samples_ptr += 32 * s->nb_channels;
} }
} }
@ -1789,7 +1801,6 @@ typedef struct MP3On4DecodeContext {
int syncword; ///< syncword patch int syncword; ///< syncword patch
const uint8_t *coff; ///< channel offsets in output buffer const uint8_t *coff; ///< channel offsets in output buffer
MPADecodeContext *mp3decctx[5]; ///< MPADecodeContext for every decoder instance MPADecodeContext *mp3decctx[5]; ///< MPADecodeContext for every decoder instance
OUT_INT *decoded_buf; ///< output buffer for decoded samples
} MP3On4DecodeContext; } MP3On4DecodeContext;
#include "mpeg4audio.h" #include "mpeg4audio.h"
@ -1831,8 +1842,6 @@ static av_cold int decode_close_mp3on4(AVCodecContext * avctx)
for (i = 0; i < s->frames; i++) for (i = 0; i < s->frames; i++)
av_free(s->mp3decctx[i]); av_free(s->mp3decctx[i]);
av_freep(&s->decoded_buf);
return 0; return 0;
} }
@ -1893,14 +1902,6 @@ static int decode_init_mp3on4(AVCodecContext * avctx)
s->mp3decctx[i]->mpadsp = s->mp3decctx[0]->mpadsp; s->mp3decctx[i]->mpadsp = s->mp3decctx[0]->mpadsp;
} }
/* Allocate buffer for multi-channel output if needed */
if (s->frames > 1) {
s->decoded_buf = av_malloc(MPA_FRAME_SIZE * MPA_MAX_CHANNELS *
sizeof(*s->decoded_buf));
if (!s->decoded_buf)
goto alloc_fail;
}
return 0; return 0;
alloc_fail: alloc_fail:
decode_close_mp3on4(avctx); decode_close_mp3on4(avctx);
@ -1927,9 +1928,9 @@ static int decode_frame_mp3on4(AVCodecContext *avctx, void *data,
MPADecodeContext *m; MPADecodeContext *m;
int fsize, len = buf_size, out_size = 0; int fsize, len = buf_size, out_size = 0;
uint32_t header; uint32_t header;
OUT_INT *out_samples; OUT_INT **out_samples;
OUT_INT *outptr, *bp; OUT_INT *outptr[2];
int fr, j, n, ch, ret; int fr, ch, ret;
/* get output buffer */ /* get output buffer */
s->frame->nb_samples = s->frames * MPA_FRAME_SIZE; s->frame->nb_samples = s->frames * MPA_FRAME_SIZE;
@ -1937,15 +1938,12 @@ static int decode_frame_mp3on4(AVCodecContext *avctx, void *data,
av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n"); av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n");
return ret; return ret;
} }
out_samples = (OUT_INT *)s->frame->data[0]; out_samples = (OUT_INT **)s->frame->extended_data;
// Discard too short frames // Discard too short frames
if (buf_size < HEADER_SIZE) if (buf_size < HEADER_SIZE)
return AVERROR_INVALIDDATA; return AVERROR_INVALIDDATA;
// If only one decoder interleave is not needed
outptr = s->frames == 1 ? out_samples : s->decoded_buf;
avctx->bit_rate = 0; avctx->bit_rate = 0;
ch = 0; ch = 0;
@ -1973,6 +1971,10 @@ static int decode_frame_mp3on4(AVCodecContext *avctx, void *data,
} }
ch += m->nb_channels; ch += m->nb_channels;
outptr[0] = out_samples[s->coff[fr]];
if (m->nb_channels > 1)
outptr[1] = out_samples[s->coff[fr] + 1];
if ((ret = mp_decode_frame(m, outptr, buf, fsize)) < 0) if ((ret = mp_decode_frame(m, outptr, buf, fsize)) < 0)
return ret; return ret;
@ -1980,23 +1982,6 @@ static int decode_frame_mp3on4(AVCodecContext *avctx, void *data,
buf += fsize; buf += fsize;
len -= fsize; len -= fsize;
if (s->frames > 1) {
n = m->avctx->frame_size*m->nb_channels;
/* interleave output data */
bp = out_samples + s->coff[fr];
if (m->nb_channels == 1) {
for (j = 0; j < n; j++) {
*bp = s->decoded_buf[j];
bp += avctx->channels;
}
} else {
for (j = 0; j < n; j++) {
bp[0] = s->decoded_buf[j++];
bp[1] = s->decoded_buf[j];
bp += avctx->channels;
}
}
}
avctx->bit_rate += m->bit_rate; avctx->bit_rate += m->bit_rate;
} }
@ -2023,6 +2008,9 @@ AVCodec ff_mp1_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush, .flush = flush,
.long_name = NULL_IF_CONFIG_SMALL("MP1 (MPEG audio layer 1)"), .long_name = NULL_IF_CONFIG_SMALL("MP1 (MPEG audio layer 1)"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16P,
AV_SAMPLE_FMT_S16,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif
#if CONFIG_MP2_DECODER #if CONFIG_MP2_DECODER
@ -2036,6 +2024,9 @@ AVCodec ff_mp2_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush, .flush = flush,
.long_name = NULL_IF_CONFIG_SMALL("MP2 (MPEG audio layer 2)"), .long_name = NULL_IF_CONFIG_SMALL("MP2 (MPEG audio layer 2)"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16P,
AV_SAMPLE_FMT_S16,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif
#if CONFIG_MP3_DECODER #if CONFIG_MP3_DECODER
@ -2049,6 +2040,9 @@ AVCodec ff_mp3_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush, .flush = flush,
.long_name = NULL_IF_CONFIG_SMALL("MP3 (MPEG audio layer 3)"), .long_name = NULL_IF_CONFIG_SMALL("MP3 (MPEG audio layer 3)"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16P,
AV_SAMPLE_FMT_S16,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif
#if CONFIG_MP3ADU_DECODER #if CONFIG_MP3ADU_DECODER
@ -2062,6 +2056,9 @@ AVCodec ff_mp3adu_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush, .flush = flush,
.long_name = NULL_IF_CONFIG_SMALL("ADU (Application Data Unit) MP3 (MPEG audio layer 3)"), .long_name = NULL_IF_CONFIG_SMALL("ADU (Application Data Unit) MP3 (MPEG audio layer 3)"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16P,
AV_SAMPLE_FMT_S16,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif
#if CONFIG_MP3ON4_DECODER #if CONFIG_MP3ON4_DECODER
@ -2076,6 +2073,8 @@ AVCodec ff_mp3on4_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush_mp3on4, .flush = flush_mp3on4,
.long_name = NULL_IF_CONFIG_SMALL("MP3onMP4"), .long_name = NULL_IF_CONFIG_SMALL("MP3onMP4"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_S16P,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif
#endif #endif

View File

@ -33,6 +33,9 @@ AVCodec ff_mp1float_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush, .flush = flush,
.long_name = NULL_IF_CONFIG_SMALL("MP1 (MPEG audio layer 1)"), .long_name = NULL_IF_CONFIG_SMALL("MP1 (MPEG audio layer 1)"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLTP,
AV_SAMPLE_FMT_FLT,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif
#if CONFIG_MP2FLOAT_DECODER #if CONFIG_MP2FLOAT_DECODER
@ -46,6 +49,9 @@ AVCodec ff_mp2float_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush, .flush = flush,
.long_name = NULL_IF_CONFIG_SMALL("MP2 (MPEG audio layer 2)"), .long_name = NULL_IF_CONFIG_SMALL("MP2 (MPEG audio layer 2)"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLTP,
AV_SAMPLE_FMT_FLT,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif
#if CONFIG_MP3FLOAT_DECODER #if CONFIG_MP3FLOAT_DECODER
@ -59,6 +65,9 @@ AVCodec ff_mp3float_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush, .flush = flush,
.long_name = NULL_IF_CONFIG_SMALL("MP3 (MPEG audio layer 3)"), .long_name = NULL_IF_CONFIG_SMALL("MP3 (MPEG audio layer 3)"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLTP,
AV_SAMPLE_FMT_FLT,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif
#if CONFIG_MP3ADUFLOAT_DECODER #if CONFIG_MP3ADUFLOAT_DECODER
@ -72,6 +81,9 @@ AVCodec ff_mp3adufloat_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush, .flush = flush,
.long_name = NULL_IF_CONFIG_SMALL("ADU (Application Data Unit) MP3 (MPEG audio layer 3)"), .long_name = NULL_IF_CONFIG_SMALL("ADU (Application Data Unit) MP3 (MPEG audio layer 3)"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLTP,
AV_SAMPLE_FMT_FLT,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif
#if CONFIG_MP3ON4FLOAT_DECODER #if CONFIG_MP3ON4FLOAT_DECODER
@ -86,5 +98,7 @@ AVCodec ff_mp3on4float_decoder = {
.capabilities = CODEC_CAP_DR1, .capabilities = CODEC_CAP_DR1,
.flush = flush_mp3on4, .flush = flush_mp3on4,
.long_name = NULL_IF_CONFIG_SMALL("MP3onMP4"), .long_name = NULL_IF_CONFIG_SMALL("MP3onMP4"),
.sample_fmts = (const enum AVSampleFormat[]) { AV_SAMPLE_FMT_FLTP,
AV_SAMPLE_FMT_NONE },
}; };
#endif #endif

View File

@ -157,7 +157,7 @@ QPEL8_H_LOWPASS_OP put
QPEL8_H_LOWPASS_OP avg QPEL8_H_LOWPASS_OP avg
%macro QPEL8_H_LOWPASS_OP_XMM 1 %macro QPEL8_H_LOWPASS_OP_XMM 1
cglobal %1_h264_qpel8_h_lowpass, 4,5,7 ; dst, src, dstStride, srcStride cglobal %1_h264_qpel8_h_lowpass, 4,5,8 ; dst, src, dstStride, srcStride
movsxdifnidn r2, r2d movsxdifnidn r2, r2d
movsxdifnidn r3, r3d movsxdifnidn r3, r3d
mov r4d, 8 mov r4d, 8
@ -312,7 +312,7 @@ QPEL8_H_LOWPASS_L2_OP avg
%macro QPEL8_H_LOWPASS_L2_OP_XMM 1 %macro QPEL8_H_LOWPASS_L2_OP_XMM 1
cglobal %1_h264_qpel8_h_lowpass_l2, 5,6,7 ; dst, src, src2, dstStride, src2Stride cglobal %1_h264_qpel8_h_lowpass_l2, 5,6,8 ; dst, src, src2, dstStride, src2Stride
movsxdifnidn r3, r3d movsxdifnidn r3, r3d
movsxdifnidn r4, r4d movsxdifnidn r4, r4d
mov r5d, 8 mov r5d, 8
@ -415,13 +415,13 @@ QPEL4_V_LOWPASS_OP avg
%macro QPEL8OR16_V_LOWPASS_OP 1 %macro QPEL8OR16_V_LOWPASS_OP 1
%if cpuflag(sse2) %if cpuflag(sse2)
cglobal %1_h264_qpel8or16_v_lowpass, 5,5,7 ; dst, src, dstStride, srcStride, h cglobal %1_h264_qpel8or16_v_lowpass, 5,5,8 ; dst, src, dstStride, srcStride, h
movsxdifnidn r2, r2d movsxdifnidn r2, r2d
movsxdifnidn r3, r3d movsxdifnidn r3, r3d
sub r1, r3 sub r1, r3
sub r1, r3 sub r1, r3
%else %else
cglobal %1_h264_qpel8or16_v_lowpass_op, 5,5,7 ; dst, src, dstStride, srcStride, h cglobal %1_h264_qpel8or16_v_lowpass_op, 5,5,8 ; dst, src, dstStride, srcStride, h
movsxdifnidn r2, r2d movsxdifnidn r2, r2d
movsxdifnidn r3, r3d movsxdifnidn r3, r3d
%endif %endif
@ -543,7 +543,7 @@ QPEL4_HV1_LOWPASS_OP put
QPEL4_HV1_LOWPASS_OP avg QPEL4_HV1_LOWPASS_OP avg
%macro QPEL8OR16_HV1_LOWPASS_OP 1 %macro QPEL8OR16_HV1_LOWPASS_OP 1
cglobal %1_h264_qpel8or16_hv1_lowpass_op, 4,4,7 ; src, tmp, srcStride, size cglobal %1_h264_qpel8or16_hv1_lowpass_op, 4,4,8 ; src, tmp, srcStride, size
movsxdifnidn r2, r2d movsxdifnidn r2, r2d
pxor m7, m7 pxor m7, m7
movh m0, [r0] movh m0, [r0]
@ -635,7 +635,7 @@ QPEL8OR16_HV2_LOWPASS_OP put
QPEL8OR16_HV2_LOWPASS_OP avg QPEL8OR16_HV2_LOWPASS_OP avg
%macro QPEL8OR16_HV2_LOWPASS_OP_XMM 1 %macro QPEL8OR16_HV2_LOWPASS_OP_XMM 1
cglobal %1_h264_qpel8or16_hv2_lowpass, 5,5,7 ; dst, tmp, dstStride, tmpStride, size cglobal %1_h264_qpel8or16_hv2_lowpass, 5,5,8 ; dst, tmp, dstStride, tmpStride, size
movsxdifnidn r2, r2d movsxdifnidn r2, r2d
movsxdifnidn r3, r3d movsxdifnidn r3, r3d
cmp r4d, 16 cmp r4d, 16