Mercurial > libavcodec.hg
annotate v210dec.c @ 10874:bcfe2acbf190 libavcodec
AAC: Compress codebook tables and optimise sign bit handling
The codebooks each consist of small number of values repeated in
groups of 2 or 4. Storing the codebooks as a packed list of 2- or
4-bit indexes into a table reduces their size substantially (from 7.5k
to 1.5k), resulting in less cache pressure.
For the band types with sign bits in the bitstream, storing the number
and position of non-zero codebook values using a few bits avoids
multiple get_bits() calls and floating-point comparisons which gcc
handles miserably.
Some float/int type punning also avoids gcc brain damage.
Overall speedup 20-35% on Cortex-A8, 20% on Core i7.
| author | mru |
|---|---|
| date | Wed, 13 Jan 2010 16:46:28 +0000 |
| parents | d7ed9dcc78e3 |
| children | 8a4984c5cacc |
| rev | line source |
|---|---|
| 9628 | 1 /* |
| 2 * V210 decoder | |
| 3 * | |
| 4 * Copyright (C) 2009 Michael Niedermayer <michaelni@gmx.at> | |
| 5 * Copyright (c) 2009 Baptiste Coudurier <baptiste dot coudurier at gmail dot com> | |
| 6 * | |
| 7 * This file is part of FFmpeg. | |
| 8 * | |
| 9 * FFmpeg is free software; you can redistribute it and/or | |
| 10 * modify it under the terms of the GNU Lesser General Public | |
| 11 * License as published by the Free Software Foundation; either | |
| 12 * version 2.1 of the License, or (at your option) any later version. | |
| 13 * | |
| 14 * FFmpeg is distributed in the hope that it will be useful, | |
| 15 * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
| 17 * Lesser General Public License for more details. | |
| 18 * | |
| 19 * You should have received a copy of the GNU Lesser General Public | |
| 20 * License along with FFmpeg; if not, write to the Free Software | |
| 21 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | |
| 22 */ | |
| 23 | |
| 24 #include "avcodec.h" | |
| 25 #include "libavutil/bswap.h" | |
| 26 | |
| 27 static av_cold int decode_init(AVCodecContext *avctx) | |
| 28 { | |
| 29 if (avctx->width & 1) { | |
| 30 av_log(avctx, AV_LOG_ERROR, "v210 needs even width\n"); | |
| 31 return -1; | |
| 32 } | |
| 10131 | 33 avctx->pix_fmt = PIX_FMT_YUV422P16; |
| 9628 | 34 avctx->bits_per_raw_sample = 10; |
| 35 | |
| 10131 | 36 avctx->coded_frame = avcodec_alloc_frame(); |
| 9628 | 37 |
| 38 return 0; | |
| 39 } | |
| 40 | |
| 10131 | 41 static int decode_frame(AVCodecContext *avctx, void *data, int *data_size, |
| 42 AVPacket *avpkt) | |
| 9628 | 43 { |
| 44 int h, w; | |
| 45 AVFrame *pic = avctx->coded_frame; | |
| 46 const uint8_t *psrc = avpkt->data; | |
| 47 uint16_t *y, *u, *v; | |
| 48 int aligned_width = ((avctx->width + 47) / 48) * 48; | |
| 49 int stride = aligned_width * 8 / 3; | |
| 50 | |
| 51 if (pic->data[0]) | |
| 52 avctx->release_buffer(avctx, pic); | |
| 53 | |
| 54 if (avpkt->size < stride * avctx->height) { | |
| 55 av_log(avctx, AV_LOG_ERROR, "packet too small\n"); | |
| 56 return -1; | |
| 57 } | |
| 58 | |
| 59 pic->reference = 0; | |
| 60 if (avctx->get_buffer(avctx, pic) < 0) | |
| 61 return -1; | |
| 62 | |
| 9630 | 63 y = (uint16_t*)pic->data[0]; |
| 64 u = (uint16_t*)pic->data[1]; | |
| 65 v = (uint16_t*)pic->data[2]; | |
| 9628 | 66 pic->pict_type = FF_I_TYPE; |
| 67 pic->key_frame = 1; | |
| 68 | |
| 10131 | 69 #define READ_PIXELS(a, b, c) \ |
| 70 do { \ | |
| 71 val = le2me_32(*src++); \ | |
| 9628 | 72 *a++ = val << 6; \ |
| 73 *b++ = (val >> 4) & 0xFFC0; \ | |
| 74 *c++ = (val >> 14) & 0xFFC0; \ | |
| 10169 | 75 } while (0) |
| 9628 | 76 |
| 77 for (h = 0; h < avctx->height; h++) { | |
| 9630 | 78 const uint32_t *src = (const uint32_t*)psrc; |
| 9628 | 79 uint32_t val; |
| 80 for (w = 0; w < avctx->width - 5; w += 6) { | |
| 81 READ_PIXELS(u, y, v); | |
| 82 READ_PIXELS(y, u, y); | |
| 83 READ_PIXELS(v, y, u); | |
| 84 READ_PIXELS(y, v, y); | |
| 85 } | |
| 86 if (w < avctx->width - 1) { | |
| 87 READ_PIXELS(u, y, v); | |
| 88 | |
| 89 val = le2me_32(*src++); | |
| 90 *y++ = val << 6; | |
| 91 } | |
| 92 if (w < avctx->width - 3) { | |
| 93 *u++ = (val >> 4) & 0xFFC0; | |
| 94 *y++ = (val >> 14) & 0xFFC0; | |
| 95 | |
| 96 val = le2me_32(*src++); | |
| 97 *v++ = val << 6; | |
| 98 *y++ = (val >> 4) & 0xFFC0; | |
| 99 } | |
| 100 | |
| 101 psrc += stride; | |
| 10131 | 102 y += pic->linesize[0] / 2 - avctx->width; |
| 103 u += pic->linesize[1] / 2 - avctx->width / 2; | |
| 104 v += pic->linesize[2] / 2 - avctx->width / 2; | |
| 9628 | 105 } |
| 106 | |
| 107 *data_size = sizeof(AVFrame); | |
| 108 *(AVFrame*)data = *avctx->coded_frame; | |
| 109 | |
| 110 return avpkt->size; | |
| 111 } | |
| 112 | |
| 113 static av_cold int decode_close(AVCodecContext *avctx) | |
| 114 { | |
|
10396
2bb3882075b6
Call release_buffer on close for v210dec and v210x
reimar
parents:
10169
diff
changeset
|
115 AVFrame *pic = avctx->coded_frame; |
|
2bb3882075b6
Call release_buffer on close for v210dec and v210x
reimar
parents:
10169
diff
changeset
|
116 if (pic->data[0]) |
|
2bb3882075b6
Call release_buffer on close for v210dec and v210x
reimar
parents:
10169
diff
changeset
|
117 avctx->release_buffer(avctx, pic); |
| 9628 | 118 av_freep(&avctx->coded_frame); |
| 119 | |
| 120 return 0; | |
| 121 } | |
| 122 | |
| 123 AVCodec v210_decoder = { | |
| 124 "v210", | |
| 125 CODEC_TYPE_VIDEO, | |
| 126 CODEC_ID_V210, | |
| 127 0, | |
| 128 decode_init, | |
| 129 NULL, | |
| 130 decode_close, | |
| 131 decode_frame, | |
| 132 CODEC_CAP_DR1, | |
| 133 .long_name = NULL_IF_CONFIG_SMALL("Uncompressed 4:2:2 10-bit"), | |
| 134 }; |
