596067ea35
Previous fix (b17734598d) tried to fix this, but it seems that
depending on CPU and video width, ffmpeg might need at least 64
byte alignment, even if CPU we're running on is only AVX2 (32 byte
alignment). That is because some other parts of ffmpeg code
statically pick 64 or 32 byte alignment internally, depending on whether
AVX512 support is even compiled in (even if CPU might not have it).
I have checked whether this does not negatively affect a platform where
SIMD alignment is always 16 (Mac), and it does not, everything works as
expected.
Pull Request: https://projects.blender.org/blender/blender/pulls/128107
370 lines
12 KiB
C
370 lines
12 KiB
C
/* SPDX-FileCopyrightText: 2011 Peter Schlaile
|
|
*
|
|
* SPDX-License-Identifier: GPL-2.0-or-later */
|
|
|
|
/** \file
|
|
* \ingroup ffmpeg
|
|
*
|
|
* Compatibility macros to make every FFMPEG installation appear
|
|
* like the most current installation (wrapping some functionality sometimes)
|
|
* it also includes all FFMPEG header files at once, no need to do it
|
|
* separately.
|
|
*/
|
|
|
|
#ifndef __FFMPEG_COMPAT_H__
|
|
#define __FFMPEG_COMPAT_H__
|
|
|
|
#include <libavcodec/avcodec.h>
|
|
#include <libavformat/avformat.h>
|
|
#include <libavutil/cpu.h>
|
|
#include <libswscale/swscale.h>
|
|
|
|
/* Check if our ffmpeg is new enough, avoids user complaints.
|
|
* Minimum supported version is currently 3.2.0 which mean the following library versions:
|
|
* libavutil > 55.30
|
|
* libavcodec > 57.60
|
|
* libavformat > 57.50
|
|
*
|
|
* We only check for one of these as they are usually updated in tandem.
|
|
*/
|
|
#if (LIBAVFORMAT_VERSION_MAJOR < 57) || \
|
|
((LIBAVFORMAT_VERSION_MAJOR == 57) && (LIBAVFORMAT_VERSION_MINOR <= 50))
|
|
# error "FFmpeg 3.2.0 or newer is needed, Upgrade your FFmpeg or disable it"
|
|
#endif
|
|
/* end sanity check */
|
|
|
|
/* visual studio 2012 does not define inline for C */
|
|
#ifdef _MSC_VER
|
|
# define FFMPEG_INLINE static __inline
|
|
#else
|
|
# define FFMPEG_INLINE static inline
|
|
#endif
|
|
|
|
#if (LIBAVFORMAT_VERSION_MAJOR < 59)
|
|
/* For versions older than ffmpeg 5.0, use the old channel layout variables.
|
|
* We intend to only keep this workaround for around two releases (3.5, 3.6).
|
|
* If it sticks around any longer, then we should consider refactoring this.
|
|
*/
|
|
# define FFMPEG_USE_OLD_CHANNEL_VARS
|
|
#endif
|
|
|
|
/* Threaded sws_scale_frame was added in ffmpeg 5.0 (swscale version 6.1). */
|
|
#if (LIBSWSCALE_VERSION_INT >= AV_VERSION_INT(6, 1, 100))
|
|
# define FFMPEG_SWSCALE_THREADING
|
|
#endif
|
|
|
|
/* AV_CODEC_CAP_AUTO_THREADS was renamed to AV_CODEC_CAP_OTHER_THREADS with
|
|
* upstream commit
|
|
* github.com/FFmpeg/FFmpeg/commit/7d09579190def3ef7562399489e628f3b65714ce
|
|
* (lavc 58.132.100) and removed with commit
|
|
* github.com/FFmpeg/FFmpeg/commit/10c9a0874cb361336237557391d306d26d43f137
|
|
* for ffmpeg 6.0.
|
|
*/
|
|
#if LIBAVCODEC_VERSION_INT < AV_VERSION_INT(58, 132, 100)
|
|
# define AV_CODEC_CAP_OTHER_THREADS AV_CODEC_CAP_AUTO_THREADS
|
|
#endif
|
|
|
|
#if (LIBAVFORMAT_VERSION_MAJOR < 58) || \
|
|
((LIBAVFORMAT_VERSION_MAJOR == 58) && (LIBAVFORMAT_VERSION_MINOR < 76))
|
|
# define FFMPEG_USE_DURATION_WORKAROUND 1
|
|
|
|
/* Before ffmpeg 4.4, package duration calculation used depricated variables to calculate the
|
|
* packet duration. Use the function from commit
|
|
* github.com/FFmpeg/FFmpeg/commit/1c0885334dda9ee8652e60c586fa2e3674056586
|
|
* to calculate the correct framerate for ffmpeg < 4.4.
|
|
*/
|
|
|
|
FFMPEG_INLINE
|
|
void my_guess_pkt_duration(AVFormatContext *s, AVStream *st, AVPacket *pkt)
|
|
{
|
|
if (pkt->duration < 0 && st->codecpar->codec_type != AVMEDIA_TYPE_SUBTITLE) {
|
|
av_log(s,
|
|
AV_LOG_WARNING,
|
|
"Packet with invalid duration %" PRId64 " in stream %d\n",
|
|
pkt->duration,
|
|
pkt->stream_index);
|
|
pkt->duration = 0;
|
|
}
|
|
|
|
if (pkt->duration) {
|
|
return;
|
|
}
|
|
|
|
switch (st->codecpar->codec_type) {
|
|
case AVMEDIA_TYPE_VIDEO:
|
|
if (st->avg_frame_rate.num > 0 && st->avg_frame_rate.den > 0) {
|
|
pkt->duration = av_rescale_q(1, av_inv_q(st->avg_frame_rate), st->time_base);
|
|
}
|
|
else if (st->time_base.num * 1000LL > st->time_base.den) {
|
|
pkt->duration = 1;
|
|
}
|
|
break;
|
|
case AVMEDIA_TYPE_AUDIO: {
|
|
int frame_size = av_get_audio_frame_duration2(st->codecpar, pkt->size);
|
|
if (frame_size && st->codecpar->sample_rate) {
|
|
pkt->duration = av_rescale_q(
|
|
frame_size, (AVRational){1, st->codecpar->sample_rate}, st->time_base);
|
|
}
|
|
break;
|
|
}
|
|
default:
|
|
break;
|
|
}
|
|
}
|
|
#endif
|
|
|
|
FFMPEG_INLINE
|
|
int64_t timestamp_from_pts_or_dts(int64_t pts, int64_t dts)
|
|
{
|
|
/* Some videos do not have any pts values, use dts instead in those cases if
|
|
* possible. Usually when this happens dts can act as pts because as all frames
|
|
* should then be presented in their decoded in order. IE pts == dts. */
|
|
if (pts == AV_NOPTS_VALUE) {
|
|
return dts;
|
|
}
|
|
return pts;
|
|
}
|
|
|
|
FFMPEG_INLINE
|
|
int64_t av_get_pts_from_frame(AVFrame *picture)
|
|
{
|
|
return timestamp_from_pts_or_dts(picture->pts, picture->pkt_dts);
|
|
}
|
|
|
|
/* Duration of the frame, in the same units as pts. 0 if unknown. */
|
|
FFMPEG_INLINE
|
|
int64_t av_get_frame_duration_in_pts_units(const AVFrame *picture)
|
|
{
|
|
#if LIBAVUTIL_VERSION_INT < AV_VERSION_INT(57, 30, 100)
|
|
return picture->pkt_duration;
|
|
#else
|
|
return picture->duration;
|
|
#endif
|
|
}
|
|
|
|
FFMPEG_INLINE size_t ffmpeg_get_buffer_alignment()
|
|
{
|
|
/* Note: even if av_frame_get_buffer suggests to pass 0 for alignment,
|
|
* as of ffmpeg 6.1/7.0 it does not use correct alignment for AVX512
|
|
* CPU (frame.c get_video_buffer ends up always using 32 alignment,
|
|
* whereas it should have used 64). Reported upstream:
|
|
* https://trac.ffmpeg.org/ticket/11116 and the fix on their code
|
|
* side is to use 64 byte alignment as soon as AVX512 is compiled
|
|
* in (even if CPU might not support it). So play safe and
|
|
* use at least 64 byte alignment here too. Currently larger than
|
|
* 64 alignment would not happen anywhere, but keep on querying
|
|
* av_cpu_max_align just in case some future platform might. */
|
|
size_t align = av_cpu_max_align();
|
|
if (align < 64) {
|
|
align = 64;
|
|
}
|
|
return align;
|
|
}
|
|
|
|
/* -------------------------------------------------------------------- */
|
|
/** \name Deinterlace code block
|
|
*
|
|
* NOTE: The code in this block are from FFmpeg 2.6.4, which is licensed by LGPL.
|
|
* \{ */
|
|
|
|
#define MAX_NEG_CROP 1024
|
|
|
|
#define times4(x) x, x, x, x
|
|
#define times256(x) times4(times4(times4(times4(times4(x)))))
|
|
|
|
static const uint8_t ff_compat_crop_tab[256 + 2 * MAX_NEG_CROP] = {
|
|
times256(0x00), 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0A,
|
|
0x0B, 0x0C, 0x0D, 0x0E, 0x0F, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16,
|
|
0x17, 0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F, 0x20, 0x21, 0x22,
|
|
0x23, 0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2A, 0x2B, 0x2C, 0x2D, 0x2E,
|
|
0x2F, 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3A,
|
|
0x3B, 0x3C, 0x3D, 0x3E, 0x3F, 0x40, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46,
|
|
0x47, 0x48, 0x49, 0x4A, 0x4B, 0x4C, 0x4D, 0x4E, 0x4F, 0x50, 0x51, 0x52,
|
|
0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x5A, 0x5B, 0x5C, 0x5D, 0x5E,
|
|
0x5F, 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, 0x6A,
|
|
0x6B, 0x6C, 0x6D, 0x6E, 0x6F, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76,
|
|
0x77, 0x78, 0x79, 0x7A, 0x7B, 0x7C, 0x7D, 0x7E, 0x7F, 0x80, 0x81, 0x82,
|
|
0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, 0x8A, 0x8B, 0x8C, 0x8D, 0x8E,
|
|
0x8F, 0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, 0x99, 0x9A,
|
|
0x9B, 0x9C, 0x9D, 0x9E, 0x9F, 0xA0, 0xA1, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6,
|
|
0xA7, 0xA8, 0xA9, 0xAA, 0xAB, 0xAC, 0xAD, 0xAE, 0xAF, 0xB0, 0xB1, 0xB2,
|
|
0xB3, 0xB4, 0xB5, 0xB6, 0xB7, 0xB8, 0xB9, 0xBA, 0xBB, 0xBC, 0xBD, 0xBE,
|
|
0xBF, 0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, 0xC8, 0xC9, 0xCA,
|
|
0xCB, 0xCC, 0xCD, 0xCE, 0xCF, 0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6,
|
|
0xD7, 0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF, 0xE0, 0xE1, 0xE2,
|
|
0xE3, 0xE4, 0xE5, 0xE6, 0xE7, 0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE,
|
|
0xEF, 0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, 0xF8, 0xF9, 0xFA,
|
|
0xFB, 0xFC, 0xFD, 0xFE, 0xFF, times256(0xFF)};
|
|
|
|
#undef times4
|
|
#undef times256
|
|
|
|
/* filter parameters: [-1 4 2 4 -1] // 8 */
|
|
FFMPEG_INLINE
|
|
void deinterlace_line(uint8_t *dst,
|
|
const uint8_t *lum_m4,
|
|
const uint8_t *lum_m3,
|
|
const uint8_t *lum_m2,
|
|
const uint8_t *lum_m1,
|
|
const uint8_t *lum,
|
|
int size)
|
|
{
|
|
const uint8_t *cm = ff_compat_crop_tab + MAX_NEG_CROP;
|
|
int sum;
|
|
|
|
for (; size > 0; size--) {
|
|
sum = -lum_m4[0];
|
|
sum += lum_m3[0] << 2;
|
|
sum += lum_m2[0] << 1;
|
|
sum += lum_m1[0] << 2;
|
|
sum += -lum[0];
|
|
dst[0] = cm[(sum + 4) >> 3];
|
|
lum_m4++;
|
|
lum_m3++;
|
|
lum_m2++;
|
|
lum_m1++;
|
|
lum++;
|
|
dst++;
|
|
}
|
|
}
|
|
|
|
FFMPEG_INLINE
|
|
void deinterlace_line_inplace(
|
|
uint8_t *lum_m4, uint8_t *lum_m3, uint8_t *lum_m2, uint8_t *lum_m1, uint8_t *lum, int size)
|
|
{
|
|
const uint8_t *cm = ff_compat_crop_tab + MAX_NEG_CROP;
|
|
int sum;
|
|
|
|
for (; size > 0; size--) {
|
|
sum = -lum_m4[0];
|
|
sum += lum_m3[0] << 2;
|
|
sum += lum_m2[0] << 1;
|
|
lum_m4[0] = lum_m2[0];
|
|
sum += lum_m1[0] << 2;
|
|
sum += -lum[0];
|
|
lum_m2[0] = cm[(sum + 4) >> 3];
|
|
lum_m4++;
|
|
lum_m3++;
|
|
lum_m2++;
|
|
lum_m1++;
|
|
lum++;
|
|
}
|
|
}
|
|
|
|
/* deinterlacing : 2 temporal taps, 3 spatial taps linear filter. The
|
|
* top field is copied as is, but the bottom field is deinterlaced
|
|
* against the top field. */
|
|
FFMPEG_INLINE
|
|
void deinterlace_bottom_field(
|
|
uint8_t *dst, int dst_wrap, const uint8_t *src1, int src_wrap, int width, int height)
|
|
{
|
|
const uint8_t *src_m2, *src_m1, *src_0, *src_p1, *src_p2;
|
|
int y;
|
|
|
|
src_m2 = src1;
|
|
src_m1 = src1;
|
|
src_0 = &src_m1[src_wrap];
|
|
src_p1 = &src_0[src_wrap];
|
|
src_p2 = &src_p1[src_wrap];
|
|
for (y = 0; y < (height - 2); y += 2) {
|
|
memcpy(dst, src_m1, width);
|
|
dst += dst_wrap;
|
|
deinterlace_line(dst, src_m2, src_m1, src_0, src_p1, src_p2, width);
|
|
src_m2 = src_0;
|
|
src_m1 = src_p1;
|
|
src_0 = src_p2;
|
|
src_p1 += 2 * src_wrap;
|
|
src_p2 += 2 * src_wrap;
|
|
dst += dst_wrap;
|
|
}
|
|
memcpy(dst, src_m1, width);
|
|
dst += dst_wrap;
|
|
/* do last line */
|
|
deinterlace_line(dst, src_m2, src_m1, src_0, src_0, src_0, width);
|
|
}
|
|
|
|
FFMPEG_INLINE
|
|
int deinterlace_bottom_field_inplace(uint8_t *src1, int src_wrap, int width, int height)
|
|
{
|
|
uint8_t *src_m1, *src_0, *src_p1, *src_p2;
|
|
int y;
|
|
uint8_t *buf = (uint8_t *)av_malloc(width);
|
|
if (!buf) {
|
|
return AVERROR(ENOMEM);
|
|
}
|
|
|
|
src_m1 = src1;
|
|
memcpy(buf, src_m1, width);
|
|
src_0 = &src_m1[src_wrap];
|
|
src_p1 = &src_0[src_wrap];
|
|
src_p2 = &src_p1[src_wrap];
|
|
for (y = 0; y < (height - 2); y += 2) {
|
|
deinterlace_line_inplace(buf, src_m1, src_0, src_p1, src_p2, width);
|
|
src_m1 = src_p1;
|
|
src_0 = src_p2;
|
|
src_p1 += 2 * src_wrap;
|
|
src_p2 += 2 * src_wrap;
|
|
}
|
|
/* do last line */
|
|
deinterlace_line_inplace(buf, src_m1, src_0, src_0, src_0, width);
|
|
av_free(buf);
|
|
return 0;
|
|
}
|
|
|
|
FFMPEG_INLINE
|
|
int av_image_deinterlace(
|
|
AVFrame *dst, const AVFrame *src, enum AVPixelFormat pix_fmt, int width, int height)
|
|
{
|
|
int i, ret;
|
|
|
|
if (pix_fmt != AV_PIX_FMT_YUV420P && pix_fmt != AV_PIX_FMT_YUVJ420P &&
|
|
pix_fmt != AV_PIX_FMT_YUV422P && pix_fmt != AV_PIX_FMT_YUVJ422P &&
|
|
pix_fmt != AV_PIX_FMT_YUV444P && pix_fmt != AV_PIX_FMT_YUV411P &&
|
|
pix_fmt != AV_PIX_FMT_GRAY8)
|
|
{
|
|
return -1;
|
|
}
|
|
if ((width & 3) != 0 || (height & 3) != 0) {
|
|
return -1;
|
|
}
|
|
|
|
for (i = 0; i < 3; i++) {
|
|
if (i == 1) {
|
|
switch (pix_fmt) {
|
|
case AV_PIX_FMT_YUVJ420P:
|
|
case AV_PIX_FMT_YUV420P:
|
|
width >>= 1;
|
|
height >>= 1;
|
|
break;
|
|
case AV_PIX_FMT_YUV422P:
|
|
case AV_PIX_FMT_YUVJ422P:
|
|
width >>= 1;
|
|
break;
|
|
case AV_PIX_FMT_YUV411P:
|
|
width >>= 2;
|
|
break;
|
|
default:
|
|
break;
|
|
}
|
|
if (pix_fmt == AV_PIX_FMT_GRAY8) {
|
|
break;
|
|
}
|
|
}
|
|
if (src == dst) {
|
|
ret = deinterlace_bottom_field_inplace(dst->data[i], dst->linesize[i], width, height);
|
|
if (ret < 0) {
|
|
return ret;
|
|
}
|
|
}
|
|
else {
|
|
deinterlace_bottom_field(
|
|
dst->data[i], dst->linesize[i], src->data[i], src->linesize[i], width, height);
|
|
}
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
/** \} Deinterlace code block */
|
|
|
|
#endif
|