mirror of https://github.com/FFmpeg/FFmpeg.git
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.
718 lines
22 KiB
718 lines
22 KiB
/* |
|
* MidiVid MV30 decoder |
|
* |
|
* Copyright (c) 2020 Paul B Mahol |
|
* |
|
* 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 <stdio.h> |
|
#include <stdlib.h> |
|
#include <string.h> |
|
|
|
#include "libavutil/thread.h" |
|
|
|
#include "avcodec.h" |
|
#include "bytestream.h" |
|
#include "copy_block.h" |
|
#include "mathops.h" |
|
#include "blockdsp.h" |
|
#include "get_bits.h" |
|
#include "internal.h" |
|
#include "aandcttab.h" |
|
|
|
#define CBP_VLC_BITS 9 |
|
|
|
typedef struct MV30Context { |
|
GetBitContext gb; |
|
|
|
int intra_quant; |
|
int inter_quant; |
|
int is_inter; |
|
int mode_size; |
|
int nb_mvectors; |
|
|
|
int block[6][64]; |
|
int16_t *mvectors; |
|
unsigned int mvectors_size; |
|
int16_t *coeffs; |
|
unsigned int coeffs_size; |
|
|
|
int16_t intraq_tab[2][64]; |
|
int16_t interq_tab[2][64]; |
|
|
|
BlockDSPContext bdsp; |
|
AVFrame *prev_frame; |
|
} MV30Context; |
|
|
|
static VLC cbp_tab; |
|
|
|
static const uint8_t luma_tab[] = { |
|
12, 12, 15, 19, 25, 34, 40, 48, |
|
12, 12, 18, 22, 27, 44, 47, 46, |
|
17, 18, 21, 26, 35, 46, 52, 47, |
|
18, 20, 24, 28, 40, 61, 59, 51, |
|
20, 24, 32, 43, 50, 72, 72, 63, |
|
25, 31, 42, 48, 58, 72, 81, 75, |
|
38, 46, 54, 61, 71, 84, 88, 85, |
|
50, 61, 65, 68, 79, 78, 86, 91, |
|
}; |
|
|
|
static const uint8_t chroma_tab[] = { |
|
12, 16, 24, 47, 99, 99, 99, 99, |
|
16, 21, 26, 66, 99, 99, 99, 99, |
|
24, 26, 56, 99, 99, 99, 99, 99, |
|
47, 66, 99, 99, 99, 99, 99, 99, |
|
99, 99, 99, 99, 99, 99, 99, 99, |
|
99, 99, 99, 99, 99, 99, 99, 99, |
|
99, 99, 99, 99, 99, 99, 99, 99, |
|
99, 99, 99, 99, 99, 99, 99, 99, |
|
}; |
|
|
|
static const uint8_t zigzag[] = { |
|
0, 1, 8, 9, 16, 2, 3, 10, |
|
17, 24, 32, 25, 18, 11, 4, 5, |
|
12, 19, 26, 33, 40, 48, 41, 34, |
|
27, 20, 13, 6, 7, 14, 21, 28, |
|
35, 42, 49, 56, 57, 50, 43, 36, |
|
29, 22, 15, 23, 30, 37, 44, 51, |
|
58, 59, 52, 45, 38, 31, 39, 46, |
|
53, 60, 61, 54, 47, 55, 62, 63, |
|
}; |
|
|
|
static void get_qtable(int16_t *table, int quant, const uint8_t *quant_tab) |
|
{ |
|
int factor = quant < 50 ? 5000 / FFMAX(quant, 1) : 200 - FFMIN(quant, 100) * 2; |
|
|
|
for (int i = 0; i < 64; i++) { |
|
table[i] = av_clip((quant_tab[i] * factor + 0x32) / 100, 1, 0x7fff); |
|
table[i] = ((int)ff_aanscales[i] * (int)table[i] + 0x800) >> 12; |
|
} |
|
} |
|
|
|
static inline void idct_1d(unsigned *blk, int step) |
|
{ |
|
const unsigned t0 = blk[0 * step] + blk[4 * step]; |
|
const unsigned t1 = blk[0 * step] - blk[4 * step]; |
|
const unsigned t2 = blk[2 * step] + blk[6 * step]; |
|
const unsigned t3 = ((int)((blk[2 * step] - blk[6 * step]) * 362U) >> 8) - t2; |
|
const unsigned t4 = t0 + t2; |
|
const unsigned t5 = t0 - t2; |
|
const unsigned t6 = t1 + t3; |
|
const unsigned t7 = t1 - t3; |
|
const unsigned t8 = blk[5 * step] + blk[3 * step]; |
|
const unsigned t9 = blk[5 * step] - blk[3 * step]; |
|
const unsigned tA = blk[1 * step] + blk[7 * step]; |
|
const unsigned tB = blk[1 * step] - blk[7 * step]; |
|
const unsigned tC = t8 + tA; |
|
const unsigned tD = (int)((tB + t9) * 473U) >> 8; |
|
const unsigned tE = (((int)(t9 * -669U) >> 8) - tC) + tD; |
|
const unsigned tF = ((int)((tA - t8) * 362U) >> 8) - tE; |
|
const unsigned t10 = (((int)(tB * 277U) >> 8) - tD) + tF; |
|
|
|
blk[0 * step] = t4 + tC; |
|
blk[1 * step] = t6 + tE; |
|
blk[2 * step] = t7 + tF; |
|
blk[3 * step] = t5 - t10; |
|
blk[4 * step] = t5 + t10; |
|
blk[5 * step] = t7 - tF; |
|
blk[6 * step] = t6 - tE; |
|
blk[7 * step] = t4 - tC; |
|
} |
|
|
|
static void idct_put(uint8_t *dst, int stride, int *block) |
|
{ |
|
for (int i = 0; i < 8; i++) { |
|
if ((block[0x08 + i] | |
|
block[0x10 + i] | |
|
block[0x18 + i] | |
|
block[0x20 + i] | |
|
block[0x28 + i] | |
|
block[0x30 + i] | |
|
block[0x38 + i]) == 0) { |
|
block[0x08 + i] = block[i]; |
|
block[0x10 + i] = block[i]; |
|
block[0x18 + i] = block[i]; |
|
block[0x20 + i] = block[i]; |
|
block[0x28 + i] = block[i]; |
|
block[0x30 + i] = block[i]; |
|
block[0x38 + i] = block[i]; |
|
} else { |
|
idct_1d(block + i, 8); |
|
} |
|
} |
|
|
|
for (int i = 0; i < 8; i++) { |
|
idct_1d(block, 1); |
|
for (int j = 0; j < 8; j++) |
|
dst[j] = av_clip_uint8((block[j] >> 5) + 128); |
|
block += 8; |
|
dst += stride; |
|
} |
|
} |
|
|
|
static void idct_add(uint8_t *dst, int stride, |
|
const uint8_t *src, int in_linesize, int *block) |
|
{ |
|
for (int i = 0; i < 8; i++) { |
|
if ((block[0x08 + i] | |
|
block[0x10 + i] | |
|
block[0x18 + i] | |
|
block[0x20 + i] | |
|
block[0x28 + i] | |
|
block[0x30 + i] | |
|
block[0x38 + i]) == 0) { |
|
block[0x08 + i] = block[i]; |
|
block[0x10 + i] = block[i]; |
|
block[0x18 + i] = block[i]; |
|
block[0x20 + i] = block[i]; |
|
block[0x28 + i] = block[i]; |
|
block[0x30 + i] = block[i]; |
|
block[0x38 + i] = block[i]; |
|
} else { |
|
idct_1d(block + i, 8); |
|
} |
|
} |
|
|
|
for (int i = 0; i < 8; i++) { |
|
idct_1d(block, 1); |
|
for (int j = 0; j < 8; j++) |
|
dst[j] = av_clip_uint8((block[j] >> 5) + src[j]); |
|
block += 8; |
|
dst += stride; |
|
src += in_linesize; |
|
} |
|
} |
|
|
|
static inline void idct2_1d(int *blk, int step) |
|
{ |
|
const unsigned int t0 = blk[0 * step]; |
|
const unsigned int t1 = blk[1 * step]; |
|
const unsigned int t2 = (int)(t1 * 473U) >> 8; |
|
const unsigned int t3 = t2 - t1; |
|
const unsigned int t4 = ((int)(t1 * 362U) >> 8) - t3; |
|
const unsigned int t5 = (((int)(t1 * 277U) >> 8) - t2) + t4; |
|
|
|
blk[0 * step] = t1 + t0; |
|
blk[1 * step] = t0 + t3; |
|
blk[2 * step] = t4 + t0; |
|
blk[3 * step] = t0 - t5; |
|
blk[4 * step] = t5 + t0; |
|
blk[5 * step] = t0 - t4; |
|
blk[6 * step] = t0 - t3; |
|
blk[7 * step] = t0 - t1; |
|
} |
|
|
|
static void idct2_put(uint8_t *dst, int stride, int *block) |
|
{ |
|
for (int i = 0; i < 2; i++) { |
|
if ((block[0x08 + i]) == 0) { |
|
block[0x08 + i] = block[i]; |
|
block[0x10 + i] = block[i]; |
|
block[0x18 + i] = block[i]; |
|
block[0x20 + i] = block[i]; |
|
block[0x28 + i] = block[i]; |
|
block[0x30 + i] = block[i]; |
|
block[0x38 + i] = block[i]; |
|
} else { |
|
idct2_1d(block + i, 8); |
|
} |
|
} |
|
|
|
for (int i = 0; i < 8; i++) { |
|
if (block[1] == 0) { |
|
for (int j = 0; j < 8; j++) |
|
dst[j] = av_clip_uint8((block[0] >> 5) + 128); |
|
} else { |
|
idct2_1d(block, 1); |
|
for (int j = 0; j < 8; j++) |
|
dst[j] = av_clip_uint8((block[j] >> 5) + 128); |
|
} |
|
block += 8; |
|
dst += stride; |
|
} |
|
} |
|
|
|
static void idct2_add(uint8_t *dst, int stride, |
|
const uint8_t *src, int in_linesize, |
|
int *block) |
|
{ |
|
for (int i = 0; i < 2; i++) { |
|
if ((block[0x08 + i]) == 0) { |
|
block[0x08 + i] = block[i]; |
|
block[0x10 + i] = block[i]; |
|
block[0x18 + i] = block[i]; |
|
block[0x20 + i] = block[i]; |
|
block[0x28 + i] = block[i]; |
|
block[0x30 + i] = block[i]; |
|
block[0x38 + i] = block[i]; |
|
} else { |
|
idct2_1d(block + i, 8); |
|
} |
|
} |
|
|
|
for (int i = 0; i < 8; i++) { |
|
if (block[1] == 0) { |
|
for (int j = 0; j < 8; j++) |
|
dst[j] = av_clip_uint8((block[0] >> 5) + src[j]); |
|
} else { |
|
idct2_1d(block, 1); |
|
for (int j = 0; j < 8; j++) |
|
dst[j] = av_clip_uint8((block[j] >> 5) + src[j]); |
|
} |
|
block += 8; |
|
dst += stride; |
|
src += in_linesize; |
|
} |
|
} |
|
|
|
static void update_inter_block(uint8_t *dst, int stride, |
|
const uint8_t *src, int in_linesize, |
|
int block) |
|
{ |
|
for (int i = 0; i < 8; i++) { |
|
for (int j = 0; j < 8; j++) |
|
dst[j] = av_clip_uint8(block + src[j]); |
|
dst += stride; |
|
src += in_linesize; |
|
} |
|
} |
|
|
|
static int decode_intra_block(AVCodecContext *avctx, int mode, |
|
GetByteContext *gbyte, int16_t *qtab, |
|
int *block, int *pfill, |
|
uint8_t *dst, int linesize) |
|
{ |
|
MV30Context *s = avctx->priv_data; |
|
int fill; |
|
|
|
switch (mode) { |
|
case 0: |
|
s->bdsp.fill_block_tab[1](dst, 128, linesize, 8); |
|
break; |
|
case 1: |
|
fill = sign_extend(bytestream2_get_ne16(gbyte), 16); |
|
pfill[0] += fill; |
|
block[0] = ((int)((unsigned)pfill[0] * qtab[0]) >> 5) + 128; |
|
s->bdsp.fill_block_tab[1](dst, block[0], linesize, 8); |
|
break; |
|
case 2: |
|
memset(block, 0, sizeof(*block) * 64); |
|
fill = sign_extend(bytestream2_get_ne16(gbyte), 16); |
|
pfill[0] += fill; |
|
block[0] = (unsigned)pfill[0] * qtab[0]; |
|
block[1] = sign_extend(bytestream2_get_ne16(gbyte), 16) * qtab[1]; |
|
block[8] = sign_extend(bytestream2_get_ne16(gbyte), 16) * qtab[8]; |
|
block[9] = sign_extend(bytestream2_get_ne16(gbyte), 16) * qtab[9]; |
|
idct2_put(dst, linesize, block); |
|
break; |
|
case 3: |
|
fill = sign_extend(bytestream2_get_ne16(gbyte), 16); |
|
pfill[0] += fill; |
|
block[0] = (unsigned)pfill[0] * qtab[0]; |
|
for (int i = 1; i < 64; i++) |
|
block[zigzag[i]] = sign_extend(bytestream2_get_ne16(gbyte), 16) * qtab[zigzag[i]]; |
|
idct_put(dst, linesize, block); |
|
break; |
|
} |
|
|
|
return 0; |
|
} |
|
|
|
static int decode_inter_block(AVCodecContext *avctx, int mode, |
|
GetByteContext *gbyte, int16_t *qtab, |
|
int *block, int *pfill, |
|
uint8_t *dst, int linesize, |
|
const uint8_t *src, int in_linesize) |
|
{ |
|
int fill; |
|
|
|
switch (mode) { |
|
case 0: |
|
copy_block8(dst, src, linesize, in_linesize, 8); |
|
break; |
|
case 1: |
|
fill = sign_extend(bytestream2_get_ne16(gbyte), 16); |
|
pfill[0] += fill; |
|
block[0] = (int)((unsigned)pfill[0] * qtab[0]) >> 5; |
|
update_inter_block(dst, linesize, src, in_linesize, block[0]); |
|
break; |
|
case 2: |
|
memset(block, 0, sizeof(*block) * 64); |
|
fill = sign_extend(bytestream2_get_ne16(gbyte), 16); |
|
pfill[0] += fill; |
|
block[0] = (unsigned)pfill[0] * qtab[0]; |
|
block[1] = sign_extend(bytestream2_get_ne16(gbyte), 16) * qtab[1]; |
|
block[8] = sign_extend(bytestream2_get_ne16(gbyte), 16) * qtab[8]; |
|
block[9] = sign_extend(bytestream2_get_ne16(gbyte), 16) * qtab[9]; |
|
idct2_add(dst, linesize, src, in_linesize, block); |
|
break; |
|
case 3: |
|
fill = sign_extend(bytestream2_get_ne16(gbyte), 16); |
|
pfill[0] += fill; |
|
block[0] = (unsigned)pfill[0] * qtab[0]; |
|
for (int i = 1; i < 64; i++) |
|
block[zigzag[i]] = sign_extend(bytestream2_get_ne16(gbyte), 16) * qtab[zigzag[i]]; |
|
idct_add(dst, linesize, src, in_linesize, block); |
|
break; |
|
} |
|
|
|
return 0; |
|
} |
|
|
|
static int decode_coeffs(GetBitContext *gb, int16_t *coeffs, int nb_codes) |
|
{ |
|
memset(coeffs, 0, nb_codes * sizeof(*coeffs)); |
|
|
|
for (int i = 0; i < nb_codes;) { |
|
int value = get_vlc2(gb, cbp_tab.table, CBP_VLC_BITS, 1); |
|
|
|
if (value > 0) { |
|
int x = get_bits(gb, value); |
|
|
|
if (x < (1 << value) / 2) { |
|
x = (1 << (value - 1)) + (x & ((1 << value) - 1 >> 1)); |
|
} else { |
|
x = -(1 << (value - 1)) - (x & ((1 << value) - 1 >> 1)); |
|
} |
|
coeffs[i++] = x; |
|
} else { |
|
int flag = get_bits1(gb); |
|
|
|
i += get_bits(gb, 3 + flag * 3) + 1 + flag * 8; |
|
} |
|
} |
|
|
|
return 0; |
|
} |
|
|
|
static int decode_intra(AVCodecContext *avctx, GetBitContext *gb, AVFrame *frame) |
|
{ |
|
MV30Context *s = avctx->priv_data; |
|
GetBitContext mgb; |
|
uint8_t *dst[6]; |
|
int linesize[6]; |
|
int ret; |
|
|
|
mgb = *gb; |
|
if (get_bits_left(gb) < s->mode_size * 8) |
|
return AVERROR_INVALIDDATA; |
|
|
|
skip_bits_long(gb, s->mode_size * 8); |
|
|
|
linesize[0] = frame->linesize[0]; |
|
linesize[1] = frame->linesize[0]; |
|
linesize[2] = frame->linesize[0]; |
|
linesize[3] = frame->linesize[0]; |
|
linesize[4] = frame->linesize[1]; |
|
linesize[5] = frame->linesize[2]; |
|
|
|
for (int y = 0; y < avctx->height; y += 16) { |
|
GetByteContext gbyte; |
|
int pfill[3][1] = { {0} }; |
|
int nb_codes = get_bits(gb, 16); |
|
|
|
av_fast_padded_malloc(&s->coeffs, &s->coeffs_size, nb_codes * sizeof(*s->coeffs)); |
|
if (!s->coeffs) |
|
return AVERROR(ENOMEM); |
|
ret = decode_coeffs(gb, s->coeffs, nb_codes); |
|
if (ret < 0) |
|
return ret; |
|
|
|
bytestream2_init(&gbyte, (uint8_t *)s->coeffs, nb_codes * sizeof(*s->coeffs)); |
|
|
|
for (int x = 0; x < avctx->width; x += 16) { |
|
dst[0] = frame->data[0] + linesize[0] * y + x; |
|
dst[1] = frame->data[0] + linesize[0] * y + x + 8; |
|
dst[2] = frame->data[0] + linesize[0] * (y + 8) + x; |
|
dst[3] = frame->data[0] + linesize[0] * (y + 8) + x + 8; |
|
dst[4] = frame->data[1] + linesize[4] * (y >> 1) + (x >> 1); |
|
dst[5] = frame->data[2] + linesize[5] * (y >> 1) + (x >> 1); |
|
|
|
for (int b = 0; b < 6; b++) { |
|
int mode = get_bits_le(&mgb, 2); |
|
|
|
ret = decode_intra_block(avctx, mode, &gbyte, s->intraq_tab[b >= 4], |
|
s->block[b], |
|
pfill[(b >= 4) + (b >= 5)], |
|
dst[b], linesize[b]); |
|
if (ret < 0) |
|
return ret; |
|
} |
|
} |
|
} |
|
|
|
return 0; |
|
} |
|
|
|
static int decode_inter(AVCodecContext *avctx, GetBitContext *gb, |
|
AVFrame *frame, AVFrame *prev) |
|
{ |
|
MV30Context *s = avctx->priv_data; |
|
GetBitContext mask; |
|
GetBitContext mgb; |
|
GetByteContext mv; |
|
const int mask_size = ((avctx->height >> 4) * (avctx->width >> 4) * 2 + 7) / 8; |
|
uint8_t *dst[6], *src[6]; |
|
int in_linesize[6]; |
|
int linesize[6]; |
|
int ret, cnt = 0; |
|
int flags = 0; |
|
|
|
in_linesize[0] = prev->linesize[0]; |
|
in_linesize[1] = prev->linesize[0]; |
|
in_linesize[2] = prev->linesize[0]; |
|
in_linesize[3] = prev->linesize[0]; |
|
in_linesize[4] = prev->linesize[1]; |
|
in_linesize[5] = prev->linesize[2]; |
|
|
|
linesize[0] = frame->linesize[0]; |
|
linesize[1] = frame->linesize[0]; |
|
linesize[2] = frame->linesize[0]; |
|
linesize[3] = frame->linesize[0]; |
|
linesize[4] = frame->linesize[1]; |
|
linesize[5] = frame->linesize[2]; |
|
|
|
av_fast_padded_malloc(&s->mvectors, &s->mvectors_size, 2 * s->nb_mvectors * sizeof(*s->mvectors)); |
|
if (!s->mvectors) { |
|
ret = AVERROR(ENOMEM); |
|
goto fail; |
|
} |
|
|
|
mask = *gb; |
|
skip_bits_long(gb, mask_size * 8); |
|
mgb = *gb; |
|
skip_bits_long(gb, s->mode_size * 8); |
|
|
|
ret = decode_coeffs(gb, s->mvectors, 2 * s->nb_mvectors); |
|
if (ret < 0) |
|
goto fail; |
|
|
|
bytestream2_init(&mv, (uint8_t *)s->mvectors, 2 * s->nb_mvectors * sizeof(*s->mvectors)); |
|
|
|
for (int y = 0; y < avctx->height; y += 16) { |
|
GetByteContext gbyte; |
|
int pfill[3][1] = { {0} }; |
|
int nb_codes = get_bits(gb, 16); |
|
|
|
skip_bits(gb, 8); |
|
if (get_bits_left(gb) < 0) { |
|
ret = AVERROR_INVALIDDATA; |
|
goto fail; |
|
} |
|
|
|
av_fast_padded_malloc(&s->coeffs, &s->coeffs_size, nb_codes * sizeof(*s->coeffs)); |
|
if (!s->coeffs) { |
|
ret = AVERROR(ENOMEM); |
|
goto fail; |
|
} |
|
|
|
ret = decode_coeffs(gb, s->coeffs, nb_codes); |
|
if (ret < 0) |
|
goto fail; |
|
|
|
bytestream2_init(&gbyte, (uint8_t *)s->coeffs, nb_codes * sizeof(*s->coeffs)); |
|
|
|
for (int x = 0; x < avctx->width; x += 16) { |
|
if (cnt >= 4) |
|
cnt = 0; |
|
if (cnt == 0) { |
|
if (get_bits_left(&mask) < 8) { |
|
ret = AVERROR_INVALIDDATA; |
|
goto fail; |
|
} |
|
flags = get_bits(&mask, 8); |
|
} |
|
|
|
dst[0] = frame->data[0] + linesize[0] * y + x; |
|
dst[1] = frame->data[0] + linesize[0] * y + x + 8; |
|
dst[2] = frame->data[0] + linesize[0] * (y + 8) + x; |
|
dst[3] = frame->data[0] + linesize[0] * (y + 8) + x + 8; |
|
dst[4] = frame->data[1] + linesize[4] * (y >> 1) + (x >> 1); |
|
dst[5] = frame->data[2] + linesize[5] * (y >> 1) + (x >> 1); |
|
|
|
if ((flags >> (cnt)) & 1) { |
|
int mv_x = sign_extend(bytestream2_get_ne16(&mv), 16); |
|
int mv_y = sign_extend(bytestream2_get_ne16(&mv), 16); |
|
|
|
int px = x + mv_x; |
|
int py = y + mv_y; |
|
|
|
if (px < 0 || px > FFALIGN(avctx->width , 16) - 16 || |
|
py < 0 || py > FFALIGN(avctx->height, 16) - 16) |
|
return AVERROR_INVALIDDATA; |
|
|
|
src[0] = prev->data[0] + in_linesize[0] * py + px; |
|
src[1] = prev->data[0] + in_linesize[0] * py + px + 8; |
|
src[2] = prev->data[0] + in_linesize[0] * (py + 8) + px; |
|
src[3] = prev->data[0] + in_linesize[0] * (py + 8) + px + 8; |
|
src[4] = prev->data[1] + in_linesize[4] * (py >> 1) + (px >> 1); |
|
src[5] = prev->data[2] + in_linesize[5] * (py >> 1) + (px >> 1); |
|
|
|
if ((flags >> (cnt + 4)) & 1) { |
|
for (int b = 0; b < 6; b++) |
|
copy_block8(dst[b], src[b], linesize[b], in_linesize[b], 8); |
|
} else { |
|
for (int b = 0; b < 6; b++) { |
|
int mode = get_bits_le(&mgb, 2); |
|
|
|
ret = decode_inter_block(avctx, mode, &gbyte, s->interq_tab[b >= 4], |
|
s->block[b], |
|
pfill[(b >= 4) + (b >= 5)], |
|
dst[b], linesize[b], |
|
src[b], in_linesize[b]); |
|
if (ret < 0) |
|
goto fail; |
|
} |
|
} |
|
} else { |
|
for (int b = 0; b < 6; b++) { |
|
int mode = get_bits_le(&mgb, 2); |
|
|
|
ret = decode_intra_block(avctx, mode, &gbyte, s->intraq_tab[b >= 4], |
|
s->block[b], |
|
pfill[(b >= 4) + (b >= 5)], |
|
dst[b], linesize[b]); |
|
if (ret < 0) |
|
goto fail; |
|
} |
|
} |
|
|
|
cnt++; |
|
} |
|
} |
|
|
|
fail: |
|
return ret; |
|
} |
|
|
|
static int decode_frame(AVCodecContext *avctx, void *data, |
|
int *got_frame, AVPacket *avpkt) |
|
{ |
|
MV30Context *s = avctx->priv_data; |
|
GetBitContext *gb = &s->gb; |
|
AVFrame *frame = data; |
|
int ret; |
|
|
|
if ((ret = init_get_bits8(gb, avpkt->data, avpkt->size)) < 0) |
|
return ret; |
|
|
|
if ((ret = ff_get_buffer(avctx, frame, AV_GET_BUFFER_FLAG_REF)) < 0) |
|
return ret; |
|
|
|
s->intra_quant = get_bits(gb, 8); |
|
s->inter_quant = s->intra_quant + get_sbits(gb, 8); |
|
s->is_inter = get_bits_le(gb, 16); |
|
s->mode_size = get_bits_le(gb, 16); |
|
if (s->is_inter) |
|
s->nb_mvectors = get_bits_le(gb, 16); |
|
|
|
get_qtable(s->intraq_tab[0], s->intra_quant, luma_tab); |
|
get_qtable(s->intraq_tab[1], s->intra_quant, chroma_tab); |
|
|
|
frame->key_frame = s->is_inter == 0; |
|
|
|
if (frame->key_frame) { |
|
ret = decode_intra(avctx, gb, frame); |
|
if (ret < 0) |
|
return ret; |
|
} else { |
|
get_qtable(s->interq_tab[0], s->inter_quant, luma_tab); |
|
get_qtable(s->interq_tab[1], s->inter_quant, chroma_tab); |
|
|
|
if (!s->prev_frame->data[0]) { |
|
av_log(avctx, AV_LOG_ERROR, "Missing reference frame.\n"); |
|
return AVERROR_INVALIDDATA; |
|
} |
|
|
|
ret = decode_inter(avctx, gb, frame, s->prev_frame); |
|
if (ret < 0) |
|
return ret; |
|
} |
|
|
|
av_frame_unref(s->prev_frame); |
|
if ((ret = av_frame_ref(s->prev_frame, frame)) < 0) |
|
return ret; |
|
|
|
*got_frame = 1; |
|
|
|
return avpkt->size; |
|
} |
|
|
|
static const uint8_t cbp_bits[] = { |
|
2, 2, 3, 3, 3, 4, 5, 6, 7, 8, 9, 9, |
|
}; |
|
|
|
static av_cold void init_static_data(void) |
|
{ |
|
INIT_VLC_STATIC_FROM_LENGTHS(&cbp_tab, CBP_VLC_BITS, FF_ARRAY_ELEMS(cbp_bits), |
|
cbp_bits, 1, NULL, 0, 0, 0, 0, 1 << CBP_VLC_BITS); |
|
} |
|
|
|
static av_cold int decode_init(AVCodecContext *avctx) |
|
{ |
|
MV30Context *s = avctx->priv_data; |
|
static AVOnce init_static_once = AV_ONCE_INIT; |
|
|
|
avctx->pix_fmt = AV_PIX_FMT_YUV420P; |
|
avctx->color_range = AVCOL_RANGE_JPEG; |
|
|
|
ff_blockdsp_init(&s->bdsp, avctx); |
|
|
|
s->prev_frame = av_frame_alloc(); |
|
if (!s->prev_frame) |
|
return AVERROR(ENOMEM); |
|
|
|
ff_thread_once(&init_static_once, init_static_data); |
|
|
|
return 0; |
|
} |
|
|
|
static void decode_flush(AVCodecContext *avctx) |
|
{ |
|
MV30Context *s = avctx->priv_data; |
|
|
|
av_frame_unref(s->prev_frame); |
|
} |
|
|
|
static av_cold int decode_close(AVCodecContext *avctx) |
|
{ |
|
MV30Context *s = avctx->priv_data; |
|
|
|
av_frame_free(&s->prev_frame); |
|
av_freep(&s->coeffs); |
|
s->coeffs_size = 0; |
|
av_freep(&s->mvectors); |
|
s->mvectors_size = 0; |
|
|
|
return 0; |
|
} |
|
|
|
const AVCodec ff_mv30_decoder = { |
|
.name = "mv30", |
|
.long_name = NULL_IF_CONFIG_SMALL("MidiVid 3.0"), |
|
.type = AVMEDIA_TYPE_VIDEO, |
|
.id = AV_CODEC_ID_MV30, |
|
.priv_data_size = sizeof(MV30Context), |
|
.init = decode_init, |
|
.close = decode_close, |
|
.decode = decode_frame, |
|
.flush = decode_flush, |
|
.capabilities = AV_CODEC_CAP_DR1, |
|
.caps_internal = FF_CODEC_CAP_INIT_THREADSAFE | |
|
FF_CODEC_CAP_INIT_CLEANUP, |
|
};
|
|
|