You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

634 lines
18 KiB

/*
* H.261 decoder
* Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
* Copyright (c) 2004 Maarten Daniels
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg 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
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* H.261 decoder.
*/
#include "libavutil/avassert.h"
#include "libavutil/thread.h"
#include "avcodec.h"
#include "codec_internal.h"
#include "decode.h"
#include "mpeg_er.h"
#include "mpegutils.h"
#include "mpegvideo.h"
#include "mpegvideodec.h"
#include "h261.h"
#define H261_MBA_VLC_BITS 8
#define H261_MTYPE_VLC_BITS 6
#define H261_MV_VLC_BITS 7
#define H261_CBP_VLC_BITS 9
#define TCOEFF_VLC_BITS 9
#define MBA_STUFFING 33
#define MBA_STARTCODE 34
static VLCElem h261_mba_vlc[540];
static VLCElem h261_mtype_vlc[80];
static VLCElem h261_mv_vlc[144];
static VLCElem h261_cbp_vlc[512];
typedef struct H261DecContext {
MpegEncContext s;
H261Context common;
int current_mba;
int mba_diff;
int current_mv_x;
int current_mv_y;
int gob_number;
int gob_start_code_skipped; // 1 if gob start code is already read before gob header is read
} H261DecContext;
static av_cold void h261_decode_init_static(void)
{
VLC_INIT_STATIC_TABLE(h261_mba_vlc, H261_MBA_VLC_BITS, 35,
ff_h261_mba_bits, 1, 1,
ff_h261_mba_code, 1, 1, 0);
VLC_INIT_STATIC_SPARSE_TABLE(h261_mtype_vlc, H261_MTYPE_VLC_BITS, 10,
ff_h261_mtype_bits, 1, 1,
ff_h261_mtype_code, 1, 1,
ff_h261_mtype_map, 2, 2, 0);
VLC_INIT_STATIC_TABLE(h261_mv_vlc, H261_MV_VLC_BITS, 17,
&ff_h261_mv_tab[0][1], 2, 1,
&ff_h261_mv_tab[0][0], 2, 1, 0);
VLC_INIT_STATIC_TABLE(h261_cbp_vlc, H261_CBP_VLC_BITS, 63,
&ff_h261_cbp_tab[0][1], 2, 1,
&ff_h261_cbp_tab[0][0], 2, 1, 0);
INIT_FIRST_VLC_RL(ff_h261_rl_tcoeff, 552);
}
static av_cold int h261_decode_init(AVCodecContext *avctx)
{
static AVOnce init_static_once = AV_ONCE_INIT;
H261DecContext *const h = avctx->priv_data;
MpegEncContext *const s = &h->s;
avcodec/mpegpicture: Make MPVPicture refcounted Up until now, an initialized MpegEncContext had an array of MPVPictures (way more than were ever needed) and the MPVPicture* contained in the MPVWorkPictures as well as the input_picture and reordered_input_picture arrays (for the encoder) pointed into this array. Several of the pointers could point to the same slot and because there was no reference counting involved, one had to check for aliasing before unreferencing. Furthermore, given that these pointers were not ownership pointers the pointers were often simply reset without unreferencing the slot (happened e.g. for the RV30 and RV40 decoders) or there were moved without resetting the src pointer (happened for the encoders where the entries in the input_picture and reordered_input_picture arrays were not reset). Instead actually releasing these pictures was performed by looping over the whole array and checking which one of the entries needed to be kept. Given that the array had way too many slots (36), this meant that more than 30 MPVPictures have been unnecessarily unreferenced in every ff_mpv_frame_start(); something similar happened for the encoder. This commit changes this by making the MPVPictures refcounted via the RefStruct API. The MPVPictures itself are part of a pool so that this does not entail constant allocations; instead, the amount of allocations actually goes down, because the earlier code used such a large array of MPVPictures (36 entries) and allocated an AVFrame for every one of these on every ff_mpv_common_init(). In fact, the pool is only freed when closing the codec, so that reinitializations don't lead to new allocations (this avoids having to sync the pool in update_thread_context). Making MPVPictures refcounted also has another key benefit: It makes it possible to directly share them across threads (when using frame-threaded decoding), eliminating ugly code with underlying av_frame_ref()'s; sharing these pictures can't fail any more. The pool is allocated in ff_mpv_decode_init() for decoders, which therefore can fail now. This and the fact that the pool is not unreferenced in ff_mpv_common_end() also necessitated to mark several mpegvideo-decoders with the FF_CODEC_CAP_INIT_CLEANUP flag. *: This also means that there is no good reason any more for ff_mpv_common_frame_size_change() to exist. Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
1 year ago
int ret;
avctx->framerate = (AVRational) { 30000, 1001 };
s->private_ctx = &h->common;
// set defaults
avcodec/mpegpicture: Make MPVPicture refcounted Up until now, an initialized MpegEncContext had an array of MPVPictures (way more than were ever needed) and the MPVPicture* contained in the MPVWorkPictures as well as the input_picture and reordered_input_picture arrays (for the encoder) pointed into this array. Several of the pointers could point to the same slot and because there was no reference counting involved, one had to check for aliasing before unreferencing. Furthermore, given that these pointers were not ownership pointers the pointers were often simply reset without unreferencing the slot (happened e.g. for the RV30 and RV40 decoders) or there were moved without resetting the src pointer (happened for the encoders where the entries in the input_picture and reordered_input_picture arrays were not reset). Instead actually releasing these pictures was performed by looping over the whole array and checking which one of the entries needed to be kept. Given that the array had way too many slots (36), this meant that more than 30 MPVPictures have been unnecessarily unreferenced in every ff_mpv_frame_start(); something similar happened for the encoder. This commit changes this by making the MPVPictures refcounted via the RefStruct API. The MPVPictures itself are part of a pool so that this does not entail constant allocations; instead, the amount of allocations actually goes down, because the earlier code used such a large array of MPVPictures (36 entries) and allocated an AVFrame for every one of these on every ff_mpv_common_init(). In fact, the pool is only freed when closing the codec, so that reinitializations don't lead to new allocations (this avoids having to sync the pool in update_thread_context). Making MPVPictures refcounted also has another key benefit: It makes it possible to directly share them across threads (when using frame-threaded decoding), eliminating ugly code with underlying av_frame_ref()'s; sharing these pictures can't fail any more. The pool is allocated in ff_mpv_decode_init() for decoders, which therefore can fail now. This and the fact that the pool is not unreferenced in ff_mpv_common_end() also necessitated to mark several mpegvideo-decoders with the FF_CODEC_CAP_INIT_CLEANUP flag. *: This also means that there is no good reason any more for ff_mpv_common_frame_size_change() to exist. Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
1 year ago
ret = ff_mpv_decode_init(s, avctx);
if (ret < 0)
return ret;
s->out_format = FMT_H261;
s->low_delay = 1;
avctx->pix_fmt = AV_PIX_FMT_YUV420P;
ff_thread_once(&init_static_once, h261_decode_init_static);
return 0;
}
static inline void h261_init_dest(MpegEncContext *s)
{
const unsigned block_size = 8 >> s->avctx->lowres;
ff_init_block_index(s);
s->dest[0] += 2 * block_size;
s->dest[1] += block_size;
s->dest[2] += block_size;
}
/**
* Decode the group of blocks header or slice header.
* @return <0 if an error occurred
*/
static int h261_decode_gob_header(H261DecContext *h)
{
unsigned int val;
MpegEncContext *const s = &h->s;
if (!h->gob_start_code_skipped) {
/* Check for GOB Start Code */
val = show_bits(&s->gb, 15);
if (val)
return -1;
/* We have a GBSC */
skip_bits(&s->gb, 16);
}
h->gob_start_code_skipped = 0;
h->gob_number = get_bits(&s->gb, 4); /* GN */
s->qscale = get_bits(&s->gb, 5); /* GQUANT */
/* Check if gob_number is valid */
if (s->mb_height == 18) { // CIF
if ((h->gob_number <= 0) || (h->gob_number > 12))
return -1;
} else { // QCIF
if ((h->gob_number != 1) && (h->gob_number != 3) &&
(h->gob_number != 5))
return -1;
}
/* GEI */
if (skip_1stop_8data_bits(&s->gb) < 0)
return AVERROR_INVALIDDATA;
if (s->qscale == 0) {
av_log(s->avctx, AV_LOG_ERROR, "qscale has forbidden 0 value\n");
if (s->avctx->err_recognition & (AV_EF_BITSTREAM | AV_EF_COMPLIANT))
return -1;
}
/* For the first transmitted macroblock in a GOB, MBA is the absolute
* address. For subsequent macroblocks, MBA is the difference between
* the absolute addresses of the macroblock and the last transmitted
* macroblock. */
h->current_mba = 0;
h->mba_diff = 0;
return 0;
}
/**
* Decode skipped macroblocks.
* @return 0
*/
static int h261_decode_mb_skipped(H261DecContext *h, int mba1, int mba2)
{
MpegEncContext *const s = &h->s;
int i;
s->mb_intra = 0;
for (i = mba1; i < mba2; i++) {
int j, xy;
s->mb_x = ((h->gob_number - 1) % 2) * 11 + i % 11;
s->mb_y = ((h->gob_number - 1) / 2) * 3 + i / 11;
xy = s->mb_x + s->mb_y * s->mb_stride;
h261_init_dest(s);
for (j = 0; j < 6; j++)
s->block_last_index[j] = -1;
s->mv_dir = MV_DIR_FORWARD;
s->mv_type = MV_TYPE_16X16;
s->cur_pic.mb_type[xy] = MB_TYPE_SKIP | MB_TYPE_16x16 | MB_TYPE_FORWARD_MV;
s->mv[0][0][0] = 0;
s->mv[0][0][1] = 0;
s->mb_skipped = 1;
h->common.mtype &= ~MB_TYPE_H261_FIL;
if (s->cur_pic.motion_val[0]) {
int b_stride = 2*s->mb_width + 1;
int b_xy = 2 * s->mb_x + (2 * s->mb_y) * b_stride;
s->cur_pic.motion_val[0][b_xy][0] = s->mv[0][0][0];
s->cur_pic.motion_val[0][b_xy][1] = s->mv[0][0][1];
}
ff_mpv_reconstruct_mb(s, s->block);
}
return 0;
}
static int decode_mv_component(GetBitContext *gb, int v)
{
int mv_diff = get_vlc2(gb, h261_mv_vlc, H261_MV_VLC_BITS, 2);
/* check if mv_diff is valid */
if (mv_diff < 0)
return v;
if (mv_diff && get_bits1(gb))
mv_diff = -mv_diff;
v += mv_diff;
if (v <= -16)
v += 32;
else if (v >= 16)
v -= 32;
return v;
}
/**
* Decode a macroblock.
* @return <0 if an error occurred
*/
static int h261_decode_block(H261DecContext *h, int16_t *block, int n, int coded)
{
MpegEncContext *const s = &h->s;
int level, i, j, run;
const RLTable *rl = &ff_h261_rl_tcoeff;
const uint8_t *scan_table;
const int qmul = s->qscale << 1, qadd = (s->qscale - 1) | 1;
/* For the variable length encoding there are two code tables, one being
* used for the first transmitted LEVEL in INTER, INTER + MC and
* INTER + MC + FIL blocks, the second for all other LEVELs except the
* first one in INTRA blocks which is fixed length coded with 8 bits.
* NOTE: The two code tables only differ in one VLC so we handle that
* manually. */
scan_table = s->intra_scantable.permutated;
if (s->mb_intra) {
/* DC coef */
level = get_bits(&s->gb, 8);
// 0 (00000000b) and -128 (10000000b) are FORBIDDEN
if ((level & 0x7F) == 0) {
av_log(s->avctx, AV_LOG_ERROR, "illegal dc %d at %d %d\n",
level, s->mb_x, s->mb_y);
return -1;
}
/* The code 1000 0000 is not used, the reconstruction level of 1024
* being coded as 1111 1111. */
if (level == 255)
level = 128;
block[0] = level * s->y_dc_scale;
i = 1;
} else if (coded) {
// Run Level Code
// EOB Not possible for first level when cbp is available (that's why the table is different)
// 0 1 1s
// * * 0*
int check = show_bits(&s->gb, 2);
i = 0;
if (check & 0x2) {
skip_bits(&s->gb, 2);
block[0] = qmul + qadd;
block[0] *= (check & 0x1) ? -1 : 1;
i = 1;
}
} else {
i = 0;
}
if (!coded) {
s->block_last_index[n] = i - 1;
return 0;
}
{
OPEN_READER(re, &s->gb);
i--; // offset by -1 to allow direct indexing of scan_table
for (;;) {
UPDATE_CACHE(re, &s->gb);
GET_RL_VLC(level, run, re, &s->gb, rl->rl_vlc[0], TCOEFF_VLC_BITS, 2, 0);
if (run == 66) {
if (level) {
CLOSE_READER(re, &s->gb);
av_log(s->avctx, AV_LOG_ERROR, "illegal ac vlc code at %dx%d\n",
s->mb_x, s->mb_y);
return -1;
}
/* escape */
/* The remaining combinations of (run, level) are encoded with a
* 20-bit word consisting of 6 bits escape, 6 bits run and 8 bits
* level. */
run = SHOW_UBITS(re, &s->gb, 6) + 1;
SKIP_CACHE(re, &s->gb, 6);
level = SHOW_SBITS(re, &s->gb, 8);
if (level > 0)
level = level * qmul + qadd;
else if (level < 0)
level = level * qmul - qadd;
SKIP_COUNTER(re, &s->gb, 6 + 8);
} else if (level == 0) {
break;
} else {
level = level * qmul + qadd;
if (SHOW_UBITS(re, &s->gb, 1))
level = -level;
SKIP_COUNTER(re, &s->gb, 1);
}
i += run;
if (i >= 64) {
CLOSE_READER(re, &s->gb);
av_log(s->avctx, AV_LOG_ERROR, "run overflow at %dx%d\n",
s->mb_x, s->mb_y);
return -1;
}
j = scan_table[i];
block[j] = level;
}
CLOSE_READER(re, &s->gb);
}
s->block_last_index[n] = i;
return 0;
}
static int h261_decode_mb(H261DecContext *h)
{
MpegEncContext *const s = &h->s;
H261Context *const com = &h->common;
int i, cbp, xy;
cbp = 63;
// Read mba
do {
h->mba_diff = get_vlc2(&s->gb, h261_mba_vlc,
H261_MBA_VLC_BITS, 2);
/* Check for slice end */
/* NOTE: GOB can be empty (no MB data) or exist only of MBA_stuffing */
if (h->mba_diff == MBA_STARTCODE) { // start code
h->gob_start_code_skipped = 1;
return SLICE_END;
}
} while (h->mba_diff == MBA_STUFFING); // stuffing
if (h->mba_diff < 0) {
if (get_bits_left(&s->gb) <= 7)
return SLICE_END;
av_log(s->avctx, AV_LOG_ERROR, "illegal mba at %d %d\n", s->mb_x, s->mb_y);
return SLICE_ERROR;
}
h->mba_diff += 1;
h->current_mba += h->mba_diff;
if (h->current_mba > MBA_STUFFING)
return SLICE_ERROR;
s->mb_x = ((h->gob_number - 1) % 2) * 11 + ((h->current_mba - 1) % 11);
s->mb_y = ((h->gob_number - 1) / 2) * 3 + ((h->current_mba - 1) / 11);
xy = s->mb_x + s->mb_y * s->mb_stride;
h261_init_dest(s);
// Read mtype
com->mtype = get_vlc2(&s->gb, h261_mtype_vlc, H261_MTYPE_VLC_BITS, 2);
if (com->mtype < 0) {
av_log(s->avctx, AV_LOG_ERROR, "Invalid mtype index\n");
return SLICE_ERROR;
}
// Read mquant
if (IS_QUANT(com->mtype))
ff_set_qscale(s, get_bits(&s->gb, 5));
s->mb_intra = IS_INTRA4x4(com->mtype);
// Read mv
if (IS_16X16(com->mtype)) {
/* Motion vector data is included for all MC macroblocks. MVD is
* obtained from the macroblock vector by subtracting the vector
* of the preceding macroblock. For this calculation the vector
* of the preceding macroblock is regarded as zero in the
* following three situations:
* 1) evaluating MVD for macroblocks 1, 12 and 23;
* 2) evaluating MVD for macroblocks in which MBA does not represent a difference of 1;
* 3) MTYPE of the previous macroblock was not MC. */
if ((h->current_mba == 1) || (h->current_mba == 12) ||
(h->current_mba == 23) || (h->mba_diff != 1)) {
h->current_mv_x = 0;
h->current_mv_y = 0;
}
h->current_mv_x = decode_mv_component(&s->gb, h->current_mv_x);
h->current_mv_y = decode_mv_component(&s->gb, h->current_mv_y);
} else {
h->current_mv_x = 0;
h->current_mv_y = 0;
}
// Read cbp
if (HAS_CBP(com->mtype))
cbp = get_vlc2(&s->gb, h261_cbp_vlc, H261_CBP_VLC_BITS, 1) + 1;
if (s->mb_intra) {
s->cur_pic.mb_type[xy] = MB_TYPE_INTRA;
goto intra;
}
//set motion vectors
s->mv_dir = MV_DIR_FORWARD;
s->mv_type = MV_TYPE_16X16;
s->cur_pic.mb_type[xy] = MB_TYPE_16x16 | MB_TYPE_FORWARD_MV;
s->mv[0][0][0] = h->current_mv_x * 2; // gets divided by 2 in motion compensation
s->mv[0][0][1] = h->current_mv_y * 2;
if (s->cur_pic.motion_val[0]) {
int b_stride = 2*s->mb_width + 1;
int b_xy = 2 * s->mb_x + (2 * s->mb_y) * b_stride;
s->cur_pic.motion_val[0][b_xy][0] = s->mv[0][0][0];
s->cur_pic.motion_val[0][b_xy][1] = s->mv[0][0][1];
}
intra:
/* decode each block */
if (s->mb_intra || HAS_CBP(com->mtype)) {
s->bdsp.clear_blocks(s->block[0]);
for (i = 0; i < 6; i++) {
if (h261_decode_block(h, s->block[i], i, cbp & 32) < 0)
return SLICE_ERROR;
cbp += cbp;
}
} else {
for (i = 0; i < 6; i++)
s->block_last_index[i] = -1;
}
ff_mpv_reconstruct_mb(s, s->block);
return SLICE_OK;
}
/**
* Decode the H.261 picture header.
* @return <0 if no startcode found
*/
static int h261_decode_picture_header(H261DecContext *h)
{
MpegEncContext *const s = &h->s;
int format, i;
uint32_t startcode = 0;
for (i = get_bits_left(&s->gb); i > 24; i -= 1) {
startcode = ((startcode << 1) | get_bits(&s->gb, 1)) & 0x000FFFFF;
if (startcode == 0x10)
break;
}
if (startcode != 0x10) {
av_log(s->avctx, AV_LOG_ERROR, "Bad picture start code\n");
return -1;
}
/* temporal reference */
skip_bits(&s->gb, 5); /* picture timestamp */
/* PTYPE starts here */
skip_bits1(&s->gb); /* split screen off */
skip_bits1(&s->gb); /* camera off */
skip_bits1(&s->gb); /* freeze picture release off */
format = get_bits1(&s->gb);
// only 2 formats possible
if (format == 0) { // QCIF
s->width = 176;
s->height = 144;
} else { // CIF
s->width = 352;
s->height = 288;
}
skip_bits1(&s->gb); /* still image mode off */
skip_bits1(&s->gb); /* Reserved */
/* PEI */
if (skip_1stop_8data_bits(&s->gb) < 0)
return AVERROR_INVALIDDATA;
/* H.261 has no I-frames, but if we pass AV_PICTURE_TYPE_I for the first
* frame, the codec crashes if it does not contain all I-blocks
* (e.g. when a packet is lost). */
s->pict_type = AV_PICTURE_TYPE_P;
h->gob_number = 0;
return 0;
}
static int h261_decode_gob(H261DecContext *h)
{
MpegEncContext *const s = &h->s;
ff_set_qscale(s, s->qscale);
/* decode mb's */
while (h->current_mba <= MBA_STUFFING) {
int ret;
/* DCT & quantize */
ret = h261_decode_mb(h);
if (ret < 0) {
if (ret == SLICE_END) {
h261_decode_mb_skipped(h, h->current_mba, 33);
return 0;
}
av_log(s->avctx, AV_LOG_ERROR, "Error at MB: %d\n",
s->mb_x + s->mb_y * s->mb_stride);
return -1;
}
h261_decode_mb_skipped(h,
h->current_mba - h->mba_diff,
h->current_mba - 1);
}
return -1;
}
/**
* returns the number of bytes consumed for building the current frame
*/
static int get_consumed_bytes(MpegEncContext *s, int buf_size)
{
int pos = get_bits_count(&s->gb) >> 3;
if (pos == 0)
pos = 1; // avoid infinite loops (i doubt that is needed but ...)
if (pos + 10 > buf_size)
pos = buf_size; // oops ;)
return pos;
}
static int h261_decode_frame(AVCodecContext *avctx, AVFrame *pict,
int *got_frame, AVPacket *avpkt)
{
H261DecContext *const h = avctx->priv_data;
const uint8_t *buf = avpkt->data;
int buf_size = avpkt->size;
MpegEncContext *s = &h->s;
int ret;
ff_dlog(avctx, "*****frame %"PRId64" size=%d\n", avctx->frame_num, buf_size);
ff_dlog(avctx, "bytes=%x %x %x %x\n", buf[0], buf[1], buf[2], buf[3]);
h->gob_start_code_skipped = 0;
init_get_bits(&s->gb, buf, buf_size * 8);
ret = h261_decode_picture_header(h);
/* skip if the header was thrashed */
if (ret < 0) {
av_log(s->avctx, AV_LOG_ERROR, "header damaged\n");
return -1;
}
if (s->width != avctx->coded_width || s->height != avctx->coded_height) {
ff_mpv_common_end(s);
}
if (!s->context_initialized) {
if ((ret = ff_mpv_common_init(s)) < 0)
return ret;
ret = ff_set_dimensions(avctx, s->width, s->height);
if (ret < 0)
return ret;
}
if ((avctx->skip_frame >= AVDISCARD_NONREF && s->pict_type == AV_PICTURE_TYPE_B) ||
(avctx->skip_frame >= AVDISCARD_NONKEY && s->pict_type != AV_PICTURE_TYPE_I) ||
avctx->skip_frame >= AVDISCARD_ALL)
return buf_size;
if (ff_mpv_frame_start(s, avctx) < 0)
return -1;
ff_mpeg_er_frame_start(s);
/* decode each macroblock */
s->mb_x = 0;
s->mb_y = 0;
while (h->gob_number < (s->mb_height == 18 ? 12 : 5)) {
if (h261_decode_gob_header(h) < 0)
break;
h261_decode_gob(h);
}
ff_mpv_frame_end(s);
avcodec/mpegpicture: Split MPVPicture into WorkPicture and ordinary Pic There are two types of MPVPictures: Three (cur_pic, last_pic, next_pic) that are directly part of MpegEncContext and an array of MPVPictures that are separately allocated and are mostly accessed via pointers (cur|last|next)_pic_ptr; they are also used to store AVFrames in the encoder (necessary due to B-frames). As the name implies, each of the former is directly associated with one of the _ptr pointers: They actually share the same underlying buffers, but the ones that are part of the context can have their data pointers offset and their linesize doubled for field pictures. Up until now, each of these had their own references; in particular, there was an underlying av_frame_ref() to sync cur_pic and cur_pic_ptr etc. This is wasteful. This commit changes this relationship: cur_pic, last_pic and next_pic now become MPVWorkPictures; this structure does not have an AVFrame at all any more, but only the cached values of data and linesize. It also contains a pointer to the corresponding MPVPicture, establishing a more natural relationsship between the two. This already means that creating the context-pictures from the pointers can no longer fail. What has not been changed is the fact that the MPVPicture* pointers are not ownership pointers and that the MPVPictures are part of an array of MPVPictures that is owned by a single AVCodecContext. Doing so will be done in a latter commit. Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
1 year ago
av_assert0(s->pict_type == s->cur_pic.ptr->f->pict_type);
avcodec/mpegpicture: Split MPVPicture into WorkPicture and ordinary Pic There are two types of MPVPictures: Three (cur_pic, last_pic, next_pic) that are directly part of MpegEncContext and an array of MPVPictures that are separately allocated and are mostly accessed via pointers (cur|last|next)_pic_ptr; they are also used to store AVFrames in the encoder (necessary due to B-frames). As the name implies, each of the former is directly associated with one of the _ptr pointers: They actually share the same underlying buffers, but the ones that are part of the context can have their data pointers offset and their linesize doubled for field pictures. Up until now, each of these had their own references; in particular, there was an underlying av_frame_ref() to sync cur_pic and cur_pic_ptr etc. This is wasteful. This commit changes this relationship: cur_pic, last_pic and next_pic now become MPVWorkPictures; this structure does not have an AVFrame at all any more, but only the cached values of data and linesize. It also contains a pointer to the corresponding MPVPicture, establishing a more natural relationsship between the two. This already means that creating the context-pictures from the pointers can no longer fail. What has not been changed is the fact that the MPVPicture* pointers are not ownership pointers and that the MPVPictures are part of an array of MPVPictures that is owned by a single AVCodecContext. Doing so will be done in a latter commit. Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
1 year ago
if ((ret = av_frame_ref(pict, s->cur_pic.ptr->f)) < 0)
return ret;
avcodec/mpegpicture: Split MPVPicture into WorkPicture and ordinary Pic There are two types of MPVPictures: Three (cur_pic, last_pic, next_pic) that are directly part of MpegEncContext and an array of MPVPictures that are separately allocated and are mostly accessed via pointers (cur|last|next)_pic_ptr; they are also used to store AVFrames in the encoder (necessary due to B-frames). As the name implies, each of the former is directly associated with one of the _ptr pointers: They actually share the same underlying buffers, but the ones that are part of the context can have their data pointers offset and their linesize doubled for field pictures. Up until now, each of these had their own references; in particular, there was an underlying av_frame_ref() to sync cur_pic and cur_pic_ptr etc. This is wasteful. This commit changes this relationship: cur_pic, last_pic and next_pic now become MPVWorkPictures; this structure does not have an AVFrame at all any more, but only the cached values of data and linesize. It also contains a pointer to the corresponding MPVPicture, establishing a more natural relationsship between the two. This already means that creating the context-pictures from the pointers can no longer fail. What has not been changed is the fact that the MPVPicture* pointers are not ownership pointers and that the MPVPictures are part of an array of MPVPictures that is owned by a single AVCodecContext. Doing so will be done in a latter commit. Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
1 year ago
ff_print_debug_info(s, s->cur_pic.ptr, pict);
*got_frame = 1;
return get_consumed_bytes(s, buf_size);
}
const FFCodec ff_h261_decoder = {
.p.name = "h261",
CODEC_LONG_NAME("H.261"),
.p.type = AVMEDIA_TYPE_VIDEO,
.p.id = AV_CODEC_ID_H261,
.priv_data_size = sizeof(H261DecContext),
.init = h261_decode_init,
FF_CODEC_DECODE_CB(h261_decode_frame),
.close = ff_mpv_decode_close,
.p.capabilities = AV_CODEC_CAP_DR1,
.p.max_lowres = 3,
};