Mercurial > libavcodec.hg
annotate v210enc.c @ 12197:fbf4d5b1b664 libavcodec
Remove FF_MM_SSE2/3 flags for CPUs where this is generally not faster than
regular MMX code. Examples of this are the Core1 CPU. Instead, set a new flag,
FF_MM_SSE2/3SLOW, which can be checked for particular SSE2/3 functions that
have been checked specifically on such CPUs and are actually faster than
their MMX counterparts.
In addition, use this flag to enable particular VP8 and LPC SSE2 functions
that are faster than their MMX counterparts.
Based on a patch by Loren Merritt <lorenm AT u washington edu>.
author | rbultje |
---|---|
date | Mon, 19 Jul 2010 22:38:23 +0000 |
parents | 8a4984c5cacc |
children |
rev | line source |
---|---|
9628 | 1 /* |
2 * V210 encoder | |
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 "libavcodec/bytestream.h" | |
26 | |
27 static av_cold int encode_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 } | |
33 | |
34 if (avctx->pix_fmt != PIX_FMT_YUV422P16) { | |
35 av_log(avctx, AV_LOG_ERROR, "v210 needs YUV422P16\n"); | |
36 return -1; | |
37 } | |
38 | |
39 if (avctx->bits_per_raw_sample != 10) | |
40 av_log(avctx, AV_LOG_WARNING, "bits per raw sample: %d != 10-bit\n", | |
41 avctx->bits_per_raw_sample); | |
42 | |
43 avctx->coded_frame = avcodec_alloc_frame(); | |
44 | |
45 avctx->coded_frame->key_frame = 1; | |
46 avctx->coded_frame->pict_type = FF_I_TYPE; | |
47 | |
48 return 0; | |
49 } | |
50 | |
10131 | 51 static int encode_frame(AVCodecContext *avctx, unsigned char *buf, |
52 int buf_size, void *data) | |
9628 | 53 { |
54 const AVFrame *pic = data; | |
55 int aligned_width = ((avctx->width + 47) / 48) * 48; | |
56 int stride = aligned_width * 8 / 3; | |
57 int h, w; | |
9631 | 58 const uint16_t *y = (const uint16_t*)pic->data[0]; |
59 const uint16_t *u = (const uint16_t*)pic->data[1]; | |
60 const uint16_t *v = (const uint16_t*)pic->data[2]; | |
9628 | 61 uint8_t *p = buf; |
62 uint8_t *pdst = buf; | |
63 | |
64 if (buf_size < aligned_width * avctx->height * 8 / 3) { | |
65 av_log(avctx, AV_LOG_ERROR, "output buffer too small\n"); | |
66 return -1; | |
67 } | |
68 | |
69 #define WRITE_PIXELS(a, b, c) \ | |
70 do { \ | |
71 val = (*a++ >> 6) | \ | |
72 ((*b++ & 0xFFC0) << 4); \ | |
73 val|= (*c++ & 0xFFC0) << 14; \ | |
74 bytestream_put_le32(&p, val); \ | |
10169 | 75 } while (0) |
9628 | 76 |
77 for (h = 0; h < avctx->height; h++) { | |
78 uint32_t val; | |
79 for (w = 0; w < avctx->width - 5; w += 6) { | |
80 WRITE_PIXELS(u, y, v); | |
81 WRITE_PIXELS(y, u, y); | |
82 WRITE_PIXELS(v, y, u); | |
83 WRITE_PIXELS(y, v, y); | |
84 } | |
85 if (w < avctx->width - 1) { | |
86 WRITE_PIXELS(u, y, v); | |
87 | |
88 val = *y++ >> 6; | |
89 if (w == avctx->width - 2) | |
90 bytestream_put_le32(&p, val); | |
91 } | |
92 if (w < avctx->width - 3) { | |
93 val |=((*u++ & 0xFFC0) << 4) | | |
94 ((*y++ & 0xFFC0) << 14); | |
95 bytestream_put_le32(&p, val); | |
96 | |
97 val = (*v++ >> 6) | | |
98 (*y++ & 0xFFC0) << 4; | |
99 bytestream_put_le32(&p, val); | |
100 } | |
101 | |
102 pdst += stride; | |
103 memset(p, 0, pdst - p); | |
104 p = pdst; | |
10131 | 105 y += pic->linesize[0] / 2 - avctx->width; |
106 u += pic->linesize[1] / 2 - avctx->width / 2; | |
107 v += pic->linesize[2] / 2 - avctx->width / 2; | |
9628 | 108 } |
109 | |
110 return p - buf; | |
111 } | |
112 | |
113 static av_cold int encode_close(AVCodecContext *avctx) | |
114 { | |
115 av_freep(&avctx->coded_frame); | |
116 | |
117 return 0; | |
118 } | |
119 | |
120 AVCodec v210_encoder = { | |
121 "v210", | |
11560
8a4984c5cacc
Define AVMediaType enum, and use it instead of enum CodecType, which
stefano
parents:
10169
diff
changeset
|
122 AVMEDIA_TYPE_VIDEO, |
9628 | 123 CODEC_ID_V210, |
124 0, | |
125 encode_init, | |
126 encode_frame, | |
127 encode_close, | |
10146
38cfe222e1a4
Mark all pix_fmts and supported_framerates compound literals as const.
reimar
parents:
10131
diff
changeset
|
128 .pix_fmts = (const enum PixelFormat[]){PIX_FMT_YUV422P16, PIX_FMT_NONE}, |
9628 | 129 .long_name = NULL_IF_CONFIG_SMALL("Uncompressed 4:2:2 10-bit"), |
130 }; |