mpv/video/img_format.c

246 lines
8.2 KiB
C
Raw Normal View History

/*
* This file is part of MPlayer.
*
* MPlayer is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* MPlayer is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License along
* with MPlayer; if not, write to the Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
*/
#include <assert.h>
video: decouple internal pixel formats from FourCCs mplayer's video chain traditionally used FourCCs for pixel formats. For example, it used IMGFMT_YV12 for 4:2:0 YUV, which was defined to the string 'YV12' interpreted as unsigned int. Additionally, it used to encode information into the numeric values of some formats. The RGB formats had their bit depth and endian encoded into the least significant byte. Extended planar formats (420P10 etc.) had chroma shift, endian, and component bit depth encoded. (This has been removed in recent commits.) Replace the FourCC mess with a simple enum. Remove all the redundant formats like YV12/I420/IYUV. Replace some image format names by something more intuitive, most importantly IMGFMT_YV12 -> IMGFMT_420P. Add img_fourcc.h, which contains the old IDs for code that actually uses FourCCs. Change the way demuxers, that output raw video, identify the video format: they set either MP_FOURCC_RAWVIDEO or MP_FOURCC_IMGFMT to request the rawvideo decoder, and sh_video->imgfmt specifies the pixel format. Like the previous hack, this is supposed to avoid the need for a complete codecs.cfg entry per format, or other lookup tables. (Note that the RGB raw video FourCCs mostly rely on ffmpeg's mappings for NUT raw video, but this is still considered better than adding a raw video decoder - even if trivial, it would be full of annoying lookup tables.) The TV code has not been tested. Some corrective changes regarding endian and other image format flags creep in.
2012-12-23 19:03:30 +00:00
#include <string.h>
#include <libavutil/pixfmt.h>
#include <libavutil/pixdesc.h>
#include "compat/libav.h"
#include "video/img_format.h"
#include "video/mp_image.h"
#include "video/fmt-conversion.h"
video: decouple internal pixel formats from FourCCs mplayer's video chain traditionally used FourCCs for pixel formats. For example, it used IMGFMT_YV12 for 4:2:0 YUV, which was defined to the string 'YV12' interpreted as unsigned int. Additionally, it used to encode information into the numeric values of some formats. The RGB formats had their bit depth and endian encoded into the least significant byte. Extended planar formats (420P10 etc.) had chroma shift, endian, and component bit depth encoded. (This has been removed in recent commits.) Replace the FourCC mess with a simple enum. Remove all the redundant formats like YV12/I420/IYUV. Replace some image format names by something more intuitive, most importantly IMGFMT_YV12 -> IMGFMT_420P. Add img_fourcc.h, which contains the old IDs for code that actually uses FourCCs. Change the way demuxers, that output raw video, identify the video format: they set either MP_FOURCC_RAWVIDEO or MP_FOURCC_IMGFMT to request the rawvideo decoder, and sh_video->imgfmt specifies the pixel format. Like the previous hack, this is supposed to avoid the need for a complete codecs.cfg entry per format, or other lookup tables. (Note that the RGB raw video FourCCs mostly rely on ffmpeg's mappings for NUT raw video, but this is still considered better than adding a raw video decoder - even if trivial, it would be full of annoying lookup tables.) The TV code has not been tested. Some corrective changes regarding endian and other image format flags creep in.
2012-12-23 19:03:30 +00:00
#define FMT(string, id) \
{string, id},
#define FMT_ENDIAN(string, id) \
{string, id}, \
{string "le", MP_CONCAT(id, _LE)}, \
{string "be", MP_CONCAT(id, _BE)}, \
struct mp_imgfmt_entry {
const char *name;
int fmt;
};
static const struct mp_imgfmt_entry mp_imgfmt_list[] = {
// these formats are pretty common, and the "le"/"be" suffixes enforced
// by FFmpeg are annoying
FMT("yuv420p10", IMGFMT_420P10)
FMT("yuv420p16", IMGFMT_420P16)
// these names are weirdly different from FFmpeg's
video: decouple internal pixel formats from FourCCs mplayer's video chain traditionally used FourCCs for pixel formats. For example, it used IMGFMT_YV12 for 4:2:0 YUV, which was defined to the string 'YV12' interpreted as unsigned int. Additionally, it used to encode information into the numeric values of some formats. The RGB formats had their bit depth and endian encoded into the least significant byte. Extended planar formats (420P10 etc.) had chroma shift, endian, and component bit depth encoded. (This has been removed in recent commits.) Replace the FourCC mess with a simple enum. Remove all the redundant formats like YV12/I420/IYUV. Replace some image format names by something more intuitive, most importantly IMGFMT_YV12 -> IMGFMT_420P. Add img_fourcc.h, which contains the old IDs for code that actually uses FourCCs. Change the way demuxers, that output raw video, identify the video format: they set either MP_FOURCC_RAWVIDEO or MP_FOURCC_IMGFMT to request the rawvideo decoder, and sh_video->imgfmt specifies the pixel format. Like the previous hack, this is supposed to avoid the need for a complete codecs.cfg entry per format, or other lookup tables. (Note that the RGB raw video FourCCs mostly rely on ffmpeg's mappings for NUT raw video, but this is still considered better than adding a raw video decoder - even if trivial, it would be full of annoying lookup tables.) The TV code has not been tested. Some corrective changes regarding endian and other image format flags creep in.
2012-12-23 19:03:30 +00:00
FMT_ENDIAN("rgb12", IMGFMT_RGB12)
FMT_ENDIAN("rgb15", IMGFMT_RGB15)
FMT_ENDIAN("rgb16", IMGFMT_RGB16)
FMT_ENDIAN("bgr12", IMGFMT_BGR12)
FMT_ENDIAN("bgr15", IMGFMT_BGR15)
FMT_ENDIAN("bgr16", IMGFMT_BGR16)
// the MPlayer derived names have components in reverse order
FMT("rgb8", IMGFMT_RGB8)
FMT("bgr8", IMGFMT_BGR8)
FMT("rgb4_byte", IMGFMT_RGB4_BYTE)
FMT("bgr4_byte", IMGFMT_BGR4_BYTE)
FMT("rgb4", IMGFMT_RGB4)
FMT("bgr4", IMGFMT_BGR4)
// FFmpeg names have an annoying "_vld" suffix
video: add vda decode support (with hwaccel) and direct rendering Decoding H264 using Video Decode Acceleration used the custom 'vda_h264_dec' decoder in FFmpeg. The Good: This new implementation has some advantages over the previous one: - It works with Libav: vda_h264_dec never got into Libav since they prefer client applications to use the hwaccel API. - It is way more efficient: in my tests this implementation yields a reduction of CPU usage of roughly ~50% compared to using `vda_h264_dec` and ~65-75% compared to h264 software decoding. This is mainly because `vo_corevideo` was adapted to perform direct rendering of the `CVPixelBufferRefs` created by the Video Decode Acceleration API Framework. The Bad: - `vo_corevideo` is required to use VDA decoding acceleration. - only works with versions of ffmpeg/libav new enough (needs reference refcounting). That is FFmpeg 2.0+ and Libav's git master currently. The Ugly: VDA was hardcoded to use UYVY (2vuy) for the uploaded video texture. One one end this makes the code simple since Apple's OpenGL implementation actually supports this out of the box. It would be nice to support other output image formats and choose the best format depending on the input, or at least making it configurable. My tests indicate that CPU usage actually increases with a 420p IMGFMT output which is not what I would have expected. NOTE: There is a small memory leak with old versions of FFmpeg and with Libav since the CVPixelBufferRef is not automatically released when the AVFrame is deallocated. This can cause leaks inside libavcodec for decoded frames that are discarded before mpv wraps them inside a refcounted mp_image (this only happens on seeks). For frames that enter mpv's refcounting facilities, this is not a problem since we rewrap the CVPixelBufferRef in our mp_image that properly forwards CVPixelBufferRetain/CvPixelBufferRelease calls to the underying CVPixelBufferRef. So, for FFmpeg use something more recent than `b3d63995` for Libav the patch was posted to the dev ML in July and in review since, apparently, the proposed fix is rather hacky.
2013-08-14 13:47:18 +00:00
FMT("vda", IMGFMT_VDA)
video: add vaapi decode and output support This is based on the MPlayer VA API patches. To be exact it's based on a very stripped down version of commit f1ad459a263f8537f6c from git://gitorious.org/vaapi/mplayer.git. This doesn't contain useless things like benchmarking hacks and the demo code for GLX interop. Also, unlike in the original patch, decoding and video output are split into separate source files (the separation between decoding and display also makes pixel format hacks unnecessary). On the other hand, some features not present in the original patch were added, like screenshot support. VA API is rather bad for actual video output. Dealing with older libva versions or the completely broken vdpau backend doesn't help. OSD is low quality and should be rather slow. In some cases, only either OSD or subtitles can be shown at the same time (because OSD is drawn first, OSD is prefered). Also, libva can't decide whether it accepts straight or premultiplied alpha for OSD sub-pictures: the vdpau backend seems to assume premultiplied, while a native vaapi driver uses straight. So I picked straight alpha. It doesn't matter much, because the blending code for straight alpha I added to img_convert.c is probably buggy, and ASS subtitles might be blended incorrectly. Really good video output with VA API would probably use OpenGL and the GL interop features, but at this point you might just use vo_opengl. (Patches for making HW decoding with vo_opengl have a chance of being accepted.) Despite these issues, decoding seems to work ok. I still got tearing on the Intel system I tested (Intel(R) Core(TM) i3-2350M). It was also tested with the vdpau vaapi wrapper on a nvidia system; however this was rather broken. (Fortunately, there is no reason to use mpv's VAAPI support over native VDPAU.)
2013-08-09 12:01:30 +00:00
FMT("vaapi", IMGFMT_VAAPI)
// names below this are not preferred over the FFmpeg names
FMT("none", 0)
// endian-specific aliases (not in FFmpeg)
FMT("rgb32", IMGFMT_RGB32)
FMT("bgr32", IMGFMT_BGR32)
// old names we keep around
FMT("y8", IMGFMT_Y8)
FMT("420p", IMGFMT_420P)
FMT("yv12", IMGFMT_420P)
FMT_ENDIAN("420p16", IMGFMT_420P16)
FMT_ENDIAN("420p10", IMGFMT_420P10)
FMT("444p", IMGFMT_444P)
FMT("444p9", IMGFMT_444P9)
FMT("444p10", IMGFMT_444P10)
FMT("422p", IMGFMT_422P)
FMT("422p9", IMGFMT_422P9)
FMT("422p10", IMGFMT_422P10)
{0}
};
char **mp_imgfmt_name_list(void)
{
int count = IMGFMT_END - IMGFMT_START;
char **list = talloc_zero_array(NULL, char *, count + 1);
int num = 0;
for (int n = IMGFMT_START; n < IMGFMT_END; n++) {
const char *name = mp_imgfmt_to_name(n);
if (strcmp(name, "none") != 0 && strcmp(name, "unknown") != 0)
list[num++] = (char *)name;
}
return list;
}
int mp_imgfmt_from_name(bstr name, bool allow_hwaccel)
{
int img_fmt = 0;
for (const struct mp_imgfmt_entry *p = mp_imgfmt_list; p->name; ++p) {
if (bstr_equals0(name, p->name)) {
img_fmt = p->fmt;
break;
}
}
if (!img_fmt) {
char *t = bstrdup0(NULL, name);
img_fmt = pixfmt2imgfmt(av_get_pix_fmt(t));
talloc_free(t);
}
if (!allow_hwaccel && IMGFMT_IS_HWACCEL(img_fmt))
return 0;
return img_fmt;
}
const char *mp_imgfmt_to_name(int fmt)
{
const struct mp_imgfmt_entry *p = mp_imgfmt_list;
for (; p->fmt; p++) {
if (p->name && p->fmt == fmt)
return p->name;
}
const AVPixFmtDescriptor *pixdesc = av_pix_fmt_desc_get(imgfmt2pixfmt(fmt));
if (pixdesc && pixdesc->name)
return pixdesc->name;
return "unknown";
}
struct mp_imgfmt_desc mp_imgfmt_get_desc(int mpfmt)
{
enum AVPixelFormat fmt = imgfmt2pixfmt(mpfmt);
const AVPixFmtDescriptor *pd = av_pix_fmt_desc_get(fmt);
if (!pd || fmt == AV_PIX_FMT_NONE)
return (struct mp_imgfmt_desc) {0};
struct mp_imgfmt_desc desc = {
.id = mpfmt,
.avformat = fmt,
2013-03-03 10:14:44 +00:00
.name = mp_imgfmt_to_name(mpfmt),
.chroma_xs = pd->log2_chroma_w,
.chroma_ys = pd->log2_chroma_h,
};
int planedepth[4] = {0};
int el_size = (pd->flags & AV_PIX_FMT_FLAG_BITSTREAM) ? 1 : 8;
for (int c = 0; c < pd->nb_components; c++) {
AVComponentDescriptor d = pd->comp[c];
// multiple components per plane -> Y is definitive, ignore chroma
if (!desc.bpp[d.plane])
desc.bpp[d.plane] = (d.step_minus1 + 1) * el_size;
planedepth[d.plane] += d.depth_minus1 + 1;
}
for (int p = 0; p < 4; p++) {
if (desc.bpp[p])
desc.num_planes++;
}
video: decouple internal pixel formats from FourCCs mplayer's video chain traditionally used FourCCs for pixel formats. For example, it used IMGFMT_YV12 for 4:2:0 YUV, which was defined to the string 'YV12' interpreted as unsigned int. Additionally, it used to encode information into the numeric values of some formats. The RGB formats had their bit depth and endian encoded into the least significant byte. Extended planar formats (420P10 etc.) had chroma shift, endian, and component bit depth encoded. (This has been removed in recent commits.) Replace the FourCC mess with a simple enum. Remove all the redundant formats like YV12/I420/IYUV. Replace some image format names by something more intuitive, most importantly IMGFMT_YV12 -> IMGFMT_420P. Add img_fourcc.h, which contains the old IDs for code that actually uses FourCCs. Change the way demuxers, that output raw video, identify the video format: they set either MP_FOURCC_RAWVIDEO or MP_FOURCC_IMGFMT to request the rawvideo decoder, and sh_video->imgfmt specifies the pixel format. Like the previous hack, this is supposed to avoid the need for a complete codecs.cfg entry per format, or other lookup tables. (Note that the RGB raw video FourCCs mostly rely on ffmpeg's mappings for NUT raw video, but this is still considered better than adding a raw video decoder - even if trivial, it would be full of annoying lookup tables.) The TV code has not been tested. Some corrective changes regarding endian and other image format flags creep in.
2012-12-23 19:03:30 +00:00
// Packed RGB formats are the only formats that have less than 8 bits per
// component, and still require endian dependent access.
if (pd->comp[0].depth_minus1 + 1 <= 8 &&
!(mpfmt >= IMGFMT_RGB12_LE && mpfmt <= IMGFMT_BGR16_BE))
video: decouple internal pixel formats from FourCCs mplayer's video chain traditionally used FourCCs for pixel formats. For example, it used IMGFMT_YV12 for 4:2:0 YUV, which was defined to the string 'YV12' interpreted as unsigned int. Additionally, it used to encode information into the numeric values of some formats. The RGB formats had their bit depth and endian encoded into the least significant byte. Extended planar formats (420P10 etc.) had chroma shift, endian, and component bit depth encoded. (This has been removed in recent commits.) Replace the FourCC mess with a simple enum. Remove all the redundant formats like YV12/I420/IYUV. Replace some image format names by something more intuitive, most importantly IMGFMT_YV12 -> IMGFMT_420P. Add img_fourcc.h, which contains the old IDs for code that actually uses FourCCs. Change the way demuxers, that output raw video, identify the video format: they set either MP_FOURCC_RAWVIDEO or MP_FOURCC_IMGFMT to request the rawvideo decoder, and sh_video->imgfmt specifies the pixel format. Like the previous hack, this is supposed to avoid the need for a complete codecs.cfg entry per format, or other lookup tables. (Note that the RGB raw video FourCCs mostly rely on ffmpeg's mappings for NUT raw video, but this is still considered better than adding a raw video decoder - even if trivial, it would be full of annoying lookup tables.) The TV code has not been tested. Some corrective changes regarding endian and other image format flags creep in.
2012-12-23 19:03:30 +00:00
{
desc.flags |= MP_IMGFLAG_LE | MP_IMGFLAG_BE;
} else {
desc.flags |= (pd->flags & AV_PIX_FMT_FLAG_BE)
? MP_IMGFLAG_BE : MP_IMGFLAG_LE;
video: decouple internal pixel formats from FourCCs mplayer's video chain traditionally used FourCCs for pixel formats. For example, it used IMGFMT_YV12 for 4:2:0 YUV, which was defined to the string 'YV12' interpreted as unsigned int. Additionally, it used to encode information into the numeric values of some formats. The RGB formats had their bit depth and endian encoded into the least significant byte. Extended planar formats (420P10 etc.) had chroma shift, endian, and component bit depth encoded. (This has been removed in recent commits.) Replace the FourCC mess with a simple enum. Remove all the redundant formats like YV12/I420/IYUV. Replace some image format names by something more intuitive, most importantly IMGFMT_YV12 -> IMGFMT_420P. Add img_fourcc.h, which contains the old IDs for code that actually uses FourCCs. Change the way demuxers, that output raw video, identify the video format: they set either MP_FOURCC_RAWVIDEO or MP_FOURCC_IMGFMT to request the rawvideo decoder, and sh_video->imgfmt specifies the pixel format. Like the previous hack, this is supposed to avoid the need for a complete codecs.cfg entry per format, or other lookup tables. (Note that the RGB raw video FourCCs mostly rely on ffmpeg's mappings for NUT raw video, but this is still considered better than adding a raw video decoder - even if trivial, it would be full of annoying lookup tables.) The TV code has not been tested. Some corrective changes regarding endian and other image format flags creep in.
2012-12-23 19:03:30 +00:00
}
desc.plane_bits = planedepth[0];
if (mpfmt == IMGFMT_XYZ12_LE || mpfmt == IMGFMT_XYZ12_BE) {
desc.flags |= MP_IMGFLAG_XYZ;
} else if (!(pd->flags & AV_PIX_FMT_FLAG_RGB) &&
fmt != AV_PIX_FMT_MONOBLACK &&
fmt != AV_PIX_FMT_PAL8)
{
desc.flags |= MP_IMGFLAG_YUV;
} else {
desc.flags |= MP_IMGFLAG_RGB;
}
if (pd->flags & AV_PIX_FMT_FLAG_ALPHA)
desc.flags |= MP_IMGFLAG_ALPHA;
if (mpfmt >= IMGFMT_RGB0_START && mpfmt <= IMGFMT_RGB0_END)
desc.flags &= ~MP_IMGFLAG_ALPHA;
if (desc.num_planes == pd->nb_components)
desc.flags |= MP_IMGFLAG_PLANAR;
if (!(pd->flags & AV_PIX_FMT_FLAG_HWACCEL) &&
!(pd->flags & AV_PIX_FMT_FLAG_BITSTREAM))
{
desc.flags |= MP_IMGFLAG_BYTE_ALIGNED;
for (int p = 0; p < desc.num_planes; p++)
desc.bytes[p] = desc.bpp[p] / 8;
}
// PSEUDOPAL is a complete braindeath nightmare, however it seems various
// parts of FFmpeg expect that it has a palette allocated.
if (pd->flags & (AV_PIX_FMT_FLAG_PAL | AV_PIX_FMT_FLAG_PSEUDOPAL))
desc.flags |= MP_IMGFLAG_PAL;
if ((desc.flags & MP_IMGFLAG_YUV) && (desc.flags & MP_IMGFLAG_BYTE_ALIGNED))
{
bool same_depth = true;
for (int p = 0; p < desc.num_planes; p++) {
same_depth &= planedepth[p] == planedepth[0] &&
desc.bpp[p] == desc.bpp[0];
}
if (same_depth && pd->nb_components == desc.num_planes)
desc.flags |= MP_IMGFLAG_YUV_P;
}
for (int p = 0; p < desc.num_planes; p++) {
desc.xs[p] = (p == 1 || p == 2) ? desc.chroma_xs : 0;
desc.ys[p] = (p == 1 || p == 2) ? desc.chroma_ys : 0;
}
desc.align_x = 1 << desc.chroma_xs;
desc.align_y = 1 << desc.chroma_ys;
if ((desc.bpp[0] % 8) != 0)
desc.align_x = 8 / desc.bpp[0]; // expect power of 2
return desc;
}
// Find a format that is MP_IMGFLAG_YUV_P with the following configuration.
int mp_imgfmt_find_yuv_planar(int xs, int ys, int planes, int component_bits)
{
for (int n = IMGFMT_START + 1; n < IMGFMT_END; n++) {
struct mp_imgfmt_desc desc = mp_imgfmt_get_desc(n);
if (desc.id && (desc.flags & MP_IMGFLAG_YUV_P)) {
if (desc.num_planes == planes && desc.chroma_xs == xs &&
desc.chroma_ys == ys && desc.plane_bits == component_bits &&
(desc.flags & MP_IMGFLAG_NE))
return desc.id;
}
}
return 0;
}