virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-28 20:53:54 +02:00
Code Issues Releases Activity
6c694074e1
FFmpeg/libavcodec/hqx.c
Andreas Rheinhardt 02220b88fc avcodec/thread: Don't use ThreadFrame when unnecessary 
The majority of frame-threaded decoders (mainly the intra-only)
need exactly one part of ThreadFrame: The AVFrame. They don't
need the owners nor the progress, yet they had to use it because
ff_thread_(get|release)_buffer() requires it.

This commit changes this and makes these functions work with ordinary
AVFrames; the decoders that need the extra fields for progress
use ff_thread_(get|release)_ext_buffer() which work exactly
as ff_thread_(get|release)_buffer() used to do.

This also avoids some unnecessary allocations of progress AVBuffers,
namely for H.264 and HEVC film grain frames: These frames are not
used for synchronization and therefore don't need a ThreadFrame.

Also move the ThreadFrame structure as well as ff_thread_ref_frame()
to threadframe.h, the header for frame-threaded decoders with
inter-frame dependencies.

Reviewed-by: Anton Khirnov <anton@khirnov.net>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
2022-02-09 17:22:35 +01:00

552 lines
18 KiB
C
Raw Blame History

/*
* Canopus HQX decoder
*
* 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
*/
#include <inttypes.h>
#include "libavutil/imgutils.h"
#include "libavutil/intreadwrite.h"
#include "avcodec.h"
#include "canopus.h"
#include "get_bits.h"
#include "internal.h"
#include "thread.h"
#include "hqx.h"
#include "hqxdsp.h"
/* HQX has four modes - 422, 444, 422alpha and 444alpha - all 12-bit */
enum HQXFormat {
HQX_422 = 0,
HQX_444,
HQX_422A,
HQX_444A,
};
#define HQX_HEADER_SIZE 59
/* macroblock selects a group of 4 possible quants and
* a block can use any of those four quantisers
* one column is powers of 2, the other one is powers of 2 * 3,
* then there is the special one, powers of 2 * 5 */
static const int hqx_quants[16][4] = {
{ 0x1, 0x2, 0x4, 0x8 }, { 0x1, 0x3, 0x6, 0xC },
{ 0x2, 0x4, 0x8, 0x10 }, { 0x3, 0x6, 0xC, 0x18 },
{ 0x4, 0x8, 0x10, 0x20 }, { 0x6, 0xC, 0x18, 0x30 },
{ 0x8, 0x10, 0x20, 0x40 },
{ 0xA, 0x14, 0x28, 0x50 },
{ 0xC, 0x18, 0x30, 0x60 },
{ 0x10, 0x20, 0x40, 0x80 }, { 0x18, 0x30, 0x60, 0xC0 },
{ 0x20, 0x40, 0x80, 0x100 }, { 0x30, 0x60, 0xC0, 0x180 },
{ 0x40, 0x80, 0x100, 0x200 }, { 0x60, 0xC0, 0x180, 0x300 },
{ 0x80, 0x100, 0x200, 0x400 }
};
static const uint8_t hqx_quant_luma[64] = {
16, 16, 16, 19, 19, 19, 42, 44,
16, 16, 19, 19, 19, 38, 43, 45,
16, 19, 19, 19, 40, 41, 45, 48,
19, 19, 19, 40, 41, 42, 46, 49,
19, 19, 40, 41, 42, 43, 48, 101,
19, 38, 41, 42, 43, 44, 98, 104,
42, 43, 45, 46, 48, 98, 109, 116,
44, 45, 48, 49, 101, 104, 116, 123,
};
static const uint8_t hqx_quant_chroma[64] = {
16, 16, 19, 25, 26, 26, 42, 44,
16, 19, 25, 25, 26, 38, 43, 91,
19, 25, 26, 27, 40, 41, 91, 96,
25, 25, 27, 40, 41, 84, 93, 197,
26, 26, 40, 41, 84, 86, 191, 203,
26, 38, 41, 84, 86, 177, 197, 209,
42, 43, 91, 93, 191, 197, 219, 232,
44, 91, 96, 197, 203, 209, 232, 246,
};
static inline void put_blocks(HQXContext *ctx, int plane,
int x, int y, int ilace,
int16_t *block0, int16_t *block1,
const uint8_t *quant)
{
int fields = ilace ? 2 : 1;
int lsize = ctx->pic->linesize[plane];
uint8_t *p = ctx->pic->data[plane] + x * 2;
ctx->hqxdsp.idct_put((uint16_t *)(p + y * lsize),
lsize * fields, block0, quant);
ctx->hqxdsp.idct_put((uint16_t *)(p + (y + (ilace ? 1 : 8)) * lsize),
lsize * fields, block1, quant);
}
static inline void hqx_get_ac(GetBitContext *gb, const HQXAC *ac,
int *run, int *lev)
{
int val;
val = show_bits(gb, ac->lut_bits);
if (ac->lut[val].bits == -1) {
GetBitContext gb2 = *gb;
skip_bits(&gb2, ac->lut_bits);
val = ac->lut[val].lev + show_bits(&gb2, ac->extra_bits);
}
*run = ac->lut[val].run;
*lev = ac->lut[val].lev;
skip_bits(gb, ac->lut[val].bits);
}
static int decode_block(GetBitContext *gb, VLC *vlc,
const int *quants, int dcb,
int16_t block[64], int *last_dc)
{
int q, dc;
int ac_idx;
int run, lev, pos = 1;
memset(block, 0, 64 * sizeof(*block));
dc = get_vlc2(gb, vlc->table, HQX_DC_VLC_BITS, 2);
*last_dc += dc;
block[0] = sign_extend(*last_dc << (12 - dcb), 12);
q = quants[get_bits(gb, 2)];
if (q >= 128)
ac_idx = HQX_AC_Q128;
else if (q >= 64)
ac_idx = HQX_AC_Q64;
else if (q >= 32)
ac_idx = HQX_AC_Q32;
else if (q >= 16)
ac_idx = HQX_AC_Q16;
else if (q >= 8)
ac_idx = HQX_AC_Q8;
else
ac_idx = HQX_AC_Q0;
do {
hqx_get_ac(gb, &ff_hqx_ac[ac_idx], &run, &lev);
pos += run;
if (pos >= 64)
break;
block[ff_zigzag_direct[pos++]] = lev * q;
} while (pos < 64);
return 0;
}
static int hqx_decode_422(HQXContext *ctx, int slice_no, int x, int y)
{
HQXSlice *slice = &ctx->slice[slice_no];
GetBitContext *gb = &slice->gb;
const int *quants;
int flag;
int last_dc;
int i, ret;
if (ctx->interlaced)
flag = get_bits1(gb);
else
flag = 0;
quants = hqx_quants[get_bits(gb, 4)];
for (i = 0; i < 8; i++) {
int vlc_index = ctx->dcb - 9;
if (i == 0 || i == 4 || i == 6)
last_dc = 0;
ret = decode_block(gb, &ctx->dc_vlc[vlc_index], quants,
ctx->dcb, slice->block[i], &last_dc);
if (ret < 0)
return ret;
}
put_blocks(ctx, 0, x, y, flag, slice->block[0], slice->block[2], hqx_quant_luma);
put_blocks(ctx, 0, x + 8, y, flag, slice->block[1], slice->block[3], hqx_quant_luma);
put_blocks(ctx, 2, x >> 1, y, flag, slice->block[4], slice->block[5], hqx_quant_chroma);
put_blocks(ctx, 1, x >> 1, y, flag, slice->block[6], slice->block[7], hqx_quant_chroma);
return 0;
}
static int hqx_decode_422a(HQXContext *ctx, int slice_no, int x, int y)
{
HQXSlice *slice = &ctx->slice[slice_no];
GetBitContext *gb = &slice->gb;
const int *quants;
int flag = 0;
int last_dc;
int i, ret;
int cbp;
cbp = get_vlc2(gb, ctx->cbp_vlc.table, HQX_CBP_VLC_BITS, 1);
for (i = 0; i < 12; i++)
memset(slice->block[i], 0, sizeof(**slice->block) * 64);
for (i = 0; i < 12; i++)
slice->block[i][0] = -0x800;
if (cbp) {
if (ctx->interlaced)
flag = get_bits1(gb);
quants = hqx_quants[get_bits(gb, 4)];
cbp |= cbp << 4; // alpha CBP
if (cbp & 0x3) // chroma CBP - top
cbp |= 0x500;
if (cbp & 0xC) // chroma CBP - bottom
cbp |= 0xA00;
for (i = 0; i < 12; i++) {
if (i == 0 || i == 4 || i == 8 || i == 10)
last_dc = 0;
if (cbp & (1 << i)) {
int vlc_index = ctx->dcb - 9;
ret = decode_block(gb, &ctx->dc_vlc[vlc_index], quants,
ctx->dcb, slice->block[i], &last_dc);
if (ret < 0)
return ret;
}
}
}
put_blocks(ctx, 3, x, y, flag, slice->block[ 0], slice->block[ 2], hqx_quant_luma);
put_blocks(ctx, 3, x + 8, y, flag, slice->block[ 1], slice->block[ 3], hqx_quant_luma);
put_blocks(ctx, 0, x, y, flag, slice->block[ 4], slice->block[ 6], hqx_quant_luma);
put_blocks(ctx, 0, x + 8, y, flag, slice->block[ 5], slice->block[ 7], hqx_quant_luma);
put_blocks(ctx, 2, x >> 1, y, flag, slice->block[ 8], slice->block[ 9], hqx_quant_chroma);
put_blocks(ctx, 1, x >> 1, y, flag, slice->block[10], slice->block[11], hqx_quant_chroma);
return 0;
}
static int hqx_decode_444(HQXContext *ctx, int slice_no, int x, int y)
{
HQXSlice *slice = &ctx->slice[slice_no];
GetBitContext *gb = &slice->gb;
const int *quants;
int flag;
int last_dc;
int i, ret;
if (ctx->interlaced)
flag = get_bits1(gb);
else
flag = 0;
quants = hqx_quants[get_bits(gb, 4)];
for (i = 0; i < 12; i++) {
int vlc_index = ctx->dcb - 9;
if (i == 0 || i == 4 || i == 8)
last_dc = 0;
ret = decode_block(gb, &ctx->dc_vlc[vlc_index], quants,
ctx->dcb, slice->block[i], &last_dc);
if (ret < 0)
return ret;
}
put_blocks(ctx, 0, x, y, flag, slice->block[0], slice->block[ 2], hqx_quant_luma);
put_blocks(ctx, 0, x + 8, y, flag, slice->block[1], slice->block[ 3], hqx_quant_luma);
put_blocks(ctx, 2, x, y, flag, slice->block[4], slice->block[ 6], hqx_quant_chroma);
put_blocks(ctx, 2, x + 8, y, flag, slice->block[5], slice->block[ 7], hqx_quant_chroma);
put_blocks(ctx, 1, x, y, flag, slice->block[8], slice->block[10], hqx_quant_chroma);
put_blocks(ctx, 1, x + 8, y, flag, slice->block[9], slice->block[11], hqx_quant_chroma);
return 0;
}
static int hqx_decode_444a(HQXContext *ctx, int slice_no, int x, int y)
{
HQXSlice *slice = &ctx->slice[slice_no];
GetBitContext *gb = &slice->gb;
const int *quants;
int flag = 0;
int last_dc;
int i, ret;
int cbp;
cbp = get_vlc2(gb, ctx->cbp_vlc.table, HQX_CBP_VLC_BITS, 1);
for (i = 0; i < 16; i++)
memset(slice->block[i], 0, sizeof(**slice->block) * 64);
for (i = 0; i < 16; i++)
slice->block[i][0] = -0x800;
if (cbp) {
if (ctx->interlaced)
flag = get_bits1(gb);
quants = hqx_quants[get_bits(gb, 4)];
cbp |= cbp << 4; // alpha CBP
cbp |= cbp << 8; // chroma CBP
for (i = 0; i < 16; i++) {
if (i == 0 || i == 4 || i == 8 || i == 12)
last_dc = 0;
if (cbp & (1 << i)) {
int vlc_index = ctx->dcb - 9;
ret = decode_block(gb, &ctx->dc_vlc[vlc_index], quants,
ctx->dcb, slice->block[i], &last_dc);
if (ret < 0)
return ret;
}
}
}
put_blocks(ctx, 3, x, y, flag, slice->block[ 0], slice->block[ 2], hqx_quant_luma);
put_blocks(ctx, 3, x + 8, y, flag, slice->block[ 1], slice->block[ 3], hqx_quant_luma);
put_blocks(ctx, 0, x, y, flag, slice->block[ 4], slice->block[ 6], hqx_quant_luma);
put_blocks(ctx, 0, x + 8, y, flag, slice->block[ 5], slice->block[ 7], hqx_quant_luma);
put_blocks(ctx, 2, x, y, flag, slice->block[ 8], slice->block[10], hqx_quant_chroma);
put_blocks(ctx, 2, x + 8, y, flag, slice->block[ 9], slice->block[11], hqx_quant_chroma);
put_blocks(ctx, 1, x, y, flag, slice->block[12], slice->block[14], hqx_quant_chroma);
put_blocks(ctx, 1, x + 8, y, flag, slice->block[13], slice->block[15], hqx_quant_chroma);
return 0;
}
static const int shuffle_16[16] = {
0, 5, 11, 14, 2, 7, 9, 13, 1, 4, 10, 15, 3, 6, 8, 12
};
static int decode_slice(HQXContext *ctx, int slice_no)
{
int mb_w = (ctx->width + 15) >> 4;
int mb_h = (ctx->height + 15) >> 4;
int grp_w = (mb_w + 4) / 5;
int grp_h = (mb_h + 4) / 5;
int grp_h_edge = grp_w * (mb_w / grp_w);
int grp_v_edge = grp_h * (mb_h / grp_h);
int grp_v_rest = mb_w - grp_h_edge;
int grp_h_rest = mb_h - grp_v_edge;
int num_mbs = mb_w * mb_h;
int num_tiles = (num_mbs + 479) / 480;
int std_tile_blocks = num_mbs / (16 * num_tiles);
int g_tile = slice_no * num_tiles;
int blk_addr, loc_addr, mb_x, mb_y, pos, loc_row, i;
int tile_blocks, tile_limit, tile_no;
for (tile_no = 0; tile_no < num_tiles; tile_no++, g_tile++) {
tile_blocks = std_tile_blocks;
tile_limit = -1;
if (g_tile < num_mbs - std_tile_blocks * 16 * num_tiles) {
tile_limit = num_mbs / (16 * num_tiles);
tile_blocks++;
}
for (i = 0; i < tile_blocks; i++) {
if (i == tile_limit)
blk_addr = g_tile + 16 * num_tiles * i;
else
blk_addr = tile_no + 16 * num_tiles * i +
num_tiles * shuffle_16[(i + slice_no) & 0xF];
loc_row = grp_h * (blk_addr / (grp_h * mb_w));
loc_addr = blk_addr % (grp_h * mb_w);
if (loc_row >= grp_v_edge) {
mb_x = grp_w * (loc_addr / (grp_h_rest * grp_w));
pos = loc_addr % (grp_h_rest * grp_w);
} else {
mb_x = grp_w * (loc_addr / (grp_h * grp_w));
pos = loc_addr % (grp_h * grp_w);
}
if (mb_x >= grp_h_edge) {
mb_x += pos % grp_v_rest;
mb_y = loc_row + (pos / grp_v_rest);
} else {
mb_x += pos % grp_w;
mb_y = loc_row + (pos / grp_w);
}
ctx->decode_func(ctx, slice_no, mb_x * 16, mb_y * 16);
}
}
return 0;
}
static int decode_slice_thread(AVCodecContext *avctx, void *arg,
int slice_no, int threadnr)
{
HQXContext *ctx = avctx->priv_data;
uint32_t *slice_off = ctx->slice_off;
int ret;
if (slice_off[slice_no] < HQX_HEADER_SIZE ||
slice_off[slice_no] >= slice_off[slice_no + 1] ||
slice_off[slice_no + 1] > ctx->data_size) {
av_log(avctx, AV_LOG_ERROR, "Invalid slice size %d.\n", ctx->data_size);
return AVERROR_INVALIDDATA;
}
ret = init_get_bits8(&ctx->slice[slice_no].gb,
ctx->src + slice_off[slice_no],
slice_off[slice_no + 1] - slice_off[slice_no]);
if (ret < 0)
return ret;
return decode_slice(ctx, slice_no);
}
static int hqx_decode_frame(AVCodecContext *avctx, void *data,
int *got_picture_ptr, AVPacket *avpkt)
{
HQXContext *ctx = avctx->priv_data;
AVFrame *const frame = data;
uint8_t *src = avpkt->data;
uint32_t info_tag;
int data_start;
int i, ret;
if (avpkt->size < 4 + 4) {
av_log(avctx, AV_LOG_ERROR, "Frame is too small %d.\n", avpkt->size);
return AVERROR_INVALIDDATA;
}
info_tag = AV_RL32(src);
if (info_tag == MKTAG('I', 'N', 'F', 'O')) {
uint32_t info_offset = AV_RL32(src + 4);
if (info_offset > INT_MAX || info_offset + 8 > avpkt->size) {
av_log(avctx, AV_LOG_ERROR,
"Invalid INFO header offset: 0x%08"PRIX32" is too large.\n",
info_offset);
return AVERROR_INVALIDDATA;
}
ff_canopus_parse_info_tag(avctx, src + 8, info_offset);
info_offset += 8;
src += info_offset;
}
data_start = src - avpkt->data;
ctx->data_size = avpkt->size - data_start;
ctx->src = src;
ctx->pic = data;
if (ctx->data_size < HQX_HEADER_SIZE) {
av_log(avctx, AV_LOG_ERROR, "Frame too small.\n");
return AVERROR_INVALIDDATA;
}
if (src[0] != 'H' || src[1] != 'Q') {
av_log(avctx, AV_LOG_ERROR, "Not an HQX frame.\n");
return AVERROR_INVALIDDATA;
}
ctx->interlaced = !(src[2] & 0x80);
ctx->format = src[2] & 7;
ctx->dcb = (src[3] & 3) + 8;
ctx->width = AV_RB16(src + 4);
ctx->height = AV_RB16(src + 6);
for (i = 0; i < 17; i++)
ctx->slice_off[i] = AV_RB24(src + 8 + i * 3);
if (ctx->dcb == 8) {
av_log(avctx, AV_LOG_ERROR, "Invalid DC precision %d.\n", ctx->dcb);
return AVERROR_INVALIDDATA;
}
ret = av_image_check_size(ctx->width, ctx->height, 0, avctx);
if (ret < 0) {
av_log(avctx, AV_LOG_ERROR, "Invalid stored dimensions %dx%d.\n",
ctx->width, ctx->height);
return AVERROR_INVALIDDATA;
}
avctx->coded_width = FFALIGN(ctx->width, 16);
avctx->coded_height = FFALIGN(ctx->height, 16);
avctx->width = ctx->width;
avctx->height = ctx->height;
avctx->bits_per_raw_sample = 10;
//The minimum size is 2bit per macroblock
// hqx_decode_422 & hqx_decode_444 have a unconditionally stored 4bits hqx_quants index
// hqx_decode_422a & hqx_decode_444a use cbp_vlc which has a minimum length of 2 bits for its VLCs
// The code rejects slices overlapping in their input data
if (avctx->coded_width / 16 * (avctx->coded_height / 16) *
(100 - avctx->discard_damaged_percentage) / 100 > 4LL * avpkt->size)
return AVERROR_INVALIDDATA;
switch (ctx->format) {
case HQX_422:
avctx->pix_fmt = AV_PIX_FMT_YUV422P16;
ctx->decode_func = hqx_decode_422;
break;
case HQX_444:
avctx->pix_fmt = AV_PIX_FMT_YUV444P16;
ctx->decode_func = hqx_decode_444;
break;
case HQX_422A:
avctx->pix_fmt = AV_PIX_FMT_YUVA422P16;
ctx->decode_func = hqx_decode_422a;
break;
case HQX_444A:
avctx->pix_fmt = AV_PIX_FMT_YUVA444P16;
ctx->decode_func = hqx_decode_444a;
break;
default:
av_log(avctx, AV_LOG_ERROR, "Invalid format: %d.\n", ctx->format);
return AVERROR_INVALIDDATA;
}
ret = ff_thread_get_buffer(avctx, frame, 0);
if (ret < 0)
return ret;
avctx->execute2(avctx, decode_slice_thread, NULL, NULL, 16);
ctx->pic->key_frame = 1;
ctx->pic->pict_type = AV_PICTURE_TYPE_I;
*got_picture_ptr = 1;
return avpkt->size;
}
static av_cold int hqx_decode_close(AVCodecContext *avctx)
{
int i;
HQXContext *ctx = avctx->priv_data;
ff_free_vlc(&ctx->cbp_vlc);
for (i = 0; i < 3; i++) {
ff_free_vlc(&ctx->dc_vlc[i]);
}
return 0;
}
static av_cold int hqx_decode_init(AVCodecContext *avctx)
{
HQXContext *ctx = avctx->priv_data;
ff_hqxdsp_init(&ctx->hqxdsp);
return ff_hqx_init_vlcs(ctx);
}
const AVCodec ff_hqx_decoder = {
.name = "hqx",
.long_name = NULL_IF_CONFIG_SMALL("Canopus HQX"),
.type = AVMEDIA_TYPE_VIDEO,
.id = AV_CODEC_ID_HQX,
.priv_data_size = sizeof(HQXContext),
.init = hqx_decode_init,
.decode = hqx_decode_frame,
.close = hqx_decode_close,
.capabilities = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_SLICE_THREADS |
AV_CODEC_CAP_FRAME_THREADS,
.caps_internal = FF_CODEC_CAP_INIT_THREADSAFE |
FF_CODEC_CAP_INIT_CLEANUP,
};
Reference in New Issue View Git Blame Copy Permalink