mirror of
https://github.com/FFmpeg/FFmpeg.git
synced 2024-12-18 03:19:31 +02:00
def97856de
Express bitfields more simply. Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
294 lines
8.9 KiB
C
294 lines
8.9 KiB
C
/*
|
|
* Discworld II BMV video decoder
|
|
* Copyright (c) 2011 Konstantin Shishkov
|
|
*
|
|
* This file is part of Libav.
|
|
*
|
|
* Libav 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.
|
|
*
|
|
* Libav 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 Libav; if not, write to the Free Software
|
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
|
*/
|
|
|
|
#include "libavutil/common.h"
|
|
|
|
#include "avcodec.h"
|
|
#include "bytestream.h"
|
|
#include "internal.h"
|
|
|
|
enum BMVFlags{
|
|
BMV_NOP = 0,
|
|
BMV_END,
|
|
BMV_DELTA,
|
|
BMV_INTRA,
|
|
|
|
BMV_SCROLL = 0x04,
|
|
BMV_PALETTE = 0x08,
|
|
BMV_COMMAND = 0x10,
|
|
BMV_AUDIO = 0x20,
|
|
BMV_EXT = 0x40,
|
|
BMV_PRINT = 0x80
|
|
};
|
|
|
|
#define SCREEN_WIDE 640
|
|
#define SCREEN_HIGH 429
|
|
|
|
typedef struct BMVDecContext {
|
|
AVCodecContext *avctx;
|
|
|
|
uint8_t *frame, frame_base[SCREEN_WIDE * (SCREEN_HIGH + 1)];
|
|
uint32_t pal[256];
|
|
const uint8_t *stream;
|
|
} BMVDecContext;
|
|
|
|
#define NEXT_BYTE(v) v = forward ? v + 1 : v - 1;
|
|
|
|
static int decode_bmv_frame(const uint8_t *source, int src_len, uint8_t *frame, int frame_off)
|
|
{
|
|
unsigned val, saved_val = 0;
|
|
int tmplen = src_len;
|
|
const uint8_t *src, *source_end = source + src_len;
|
|
uint8_t *frame_end = frame + SCREEN_WIDE * SCREEN_HIGH;
|
|
uint8_t *dst, *dst_end;
|
|
int len, mask;
|
|
int forward = (frame_off <= -SCREEN_WIDE) || (frame_off >= 0);
|
|
int read_two_nibbles, flag;
|
|
int advance_mode;
|
|
int mode = 0;
|
|
int i;
|
|
|
|
if (src_len <= 0)
|
|
return AVERROR_INVALIDDATA;
|
|
|
|
if (forward) {
|
|
src = source;
|
|
dst = frame;
|
|
dst_end = frame_end;
|
|
} else {
|
|
src = source + src_len - 1;
|
|
dst = frame_end - 1;
|
|
dst_end = frame - 1;
|
|
}
|
|
for (;;) {
|
|
int shift = 0;
|
|
flag = 0;
|
|
|
|
/* The mode/len decoding is a bit strange:
|
|
* values are coded as variable-length codes with nibble units,
|
|
* code end is signalled by two top bits in the nibble being nonzero.
|
|
* And since data is bytepacked and we read two nibbles at a time,
|
|
* we may get a nibble belonging to the next code.
|
|
* Hence this convoluted loop.
|
|
*/
|
|
if (!mode || (tmplen == 4)) {
|
|
if (src < source || src >= source_end)
|
|
return AVERROR_INVALIDDATA;
|
|
val = *src;
|
|
read_two_nibbles = 1;
|
|
} else {
|
|
val = saved_val;
|
|
read_two_nibbles = 0;
|
|
}
|
|
if (!(val & 0xC)) {
|
|
for (;;) {
|
|
if (!read_two_nibbles) {
|
|
if (src < source || src >= source_end)
|
|
return AVERROR_INVALIDDATA;
|
|
shift += 2;
|
|
val |= *src << shift;
|
|
if (*src & 0xC)
|
|
break;
|
|
}
|
|
// two upper bits of the nibble is zero,
|
|
// so shift top nibble value down into their place
|
|
read_two_nibbles = 0;
|
|
shift += 2;
|
|
mask = (1 << shift) - 1;
|
|
val = ((val >> 2) & ~mask) | (val & mask);
|
|
NEXT_BYTE(src);
|
|
if ((val & (0xC << shift))) {
|
|
flag = 1;
|
|
break;
|
|
}
|
|
}
|
|
} else if (mode) {
|
|
flag = tmplen != 4;
|
|
}
|
|
if (flag) {
|
|
tmplen = 4;
|
|
} else {
|
|
saved_val = val >> (4 + shift);
|
|
tmplen = 0;
|
|
val &= (1 << (shift + 4)) - 1;
|
|
NEXT_BYTE(src);
|
|
}
|
|
advance_mode = val & 1;
|
|
len = (val >> 1) - 1;
|
|
mode += 1 + advance_mode;
|
|
if (mode >= 4)
|
|
mode -= 3;
|
|
if (len <= 0 || FFABS(dst_end - dst) < len)
|
|
return AVERROR_INVALIDDATA;
|
|
switch (mode) {
|
|
case 1:
|
|
if (forward) {
|
|
if (dst - frame + SCREEN_WIDE < frame_off ||
|
|
dst - frame + SCREEN_WIDE + frame_off < 0 ||
|
|
frame_end - dst < frame_off + len ||
|
|
frame_end - dst < len)
|
|
return AVERROR_INVALIDDATA;
|
|
for (i = 0; i < len; i++)
|
|
dst[i] = dst[frame_off + i];
|
|
dst += len;
|
|
} else {
|
|
dst -= len;
|
|
if (dst - frame + SCREEN_WIDE < frame_off ||
|
|
dst - frame + SCREEN_WIDE + frame_off < 0 ||
|
|
frame_end - dst < frame_off + len ||
|
|
frame_end - dst < len)
|
|
return AVERROR_INVALIDDATA;
|
|
for (i = len - 1; i >= 0; i--)
|
|
dst[i] = dst[frame_off + i];
|
|
}
|
|
break;
|
|
case 2:
|
|
if (forward) {
|
|
if (source + src_len - src < len)
|
|
return AVERROR_INVALIDDATA;
|
|
memcpy(dst, src, len);
|
|
dst += len;
|
|
src += len;
|
|
} else {
|
|
if (src - source < len)
|
|
return AVERROR_INVALIDDATA;
|
|
dst -= len;
|
|
src -= len;
|
|
memcpy(dst, src, len);
|
|
}
|
|
break;
|
|
case 3:
|
|
val = forward ? dst[-1] : dst[1];
|
|
if (forward) {
|
|
memset(dst, val, len);
|
|
dst += len;
|
|
} else {
|
|
dst -= len;
|
|
memset(dst, val, len);
|
|
}
|
|
break;
|
|
default:
|
|
break;
|
|
}
|
|
if (dst == dst_end)
|
|
return 0;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
static int decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
|
|
AVPacket *pkt)
|
|
{
|
|
BMVDecContext * const c = avctx->priv_data;
|
|
AVFrame *frame = data;
|
|
int type, scr_off;
|
|
int i, ret;
|
|
uint8_t *srcptr, *outptr;
|
|
|
|
c->stream = pkt->data;
|
|
type = bytestream_get_byte(&c->stream);
|
|
if (type & BMV_AUDIO) {
|
|
int blobs = bytestream_get_byte(&c->stream);
|
|
if (pkt->size < blobs * 65 + 2) {
|
|
av_log(avctx, AV_LOG_ERROR, "Audio data doesn't fit in frame\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
c->stream += blobs * 65;
|
|
}
|
|
if (type & BMV_COMMAND) {
|
|
int command_size = (type & BMV_PRINT) ? 8 : 10;
|
|
if (c->stream - pkt->data + command_size > pkt->size) {
|
|
av_log(avctx, AV_LOG_ERROR, "Command data doesn't fit in frame\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
c->stream += command_size;
|
|
}
|
|
if (type & BMV_PALETTE) {
|
|
if (c->stream - pkt->data > pkt->size - 768) {
|
|
av_log(avctx, AV_LOG_ERROR, "Palette data doesn't fit in frame\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
for (i = 0; i < 256; i++)
|
|
c->pal[i] = bytestream_get_be24(&c->stream);
|
|
}
|
|
if (type & BMV_SCROLL) {
|
|
if (c->stream - pkt->data > pkt->size - 2) {
|
|
av_log(avctx, AV_LOG_ERROR, "Screen offset data doesn't fit in frame\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
scr_off = (int16_t)bytestream_get_le16(&c->stream);
|
|
} else if ((type & BMV_INTRA) == BMV_INTRA) {
|
|
scr_off = -640;
|
|
} else {
|
|
scr_off = 0;
|
|
}
|
|
|
|
if ((ret = ff_get_buffer(avctx, frame, 0)) < 0) {
|
|
av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n");
|
|
return ret;
|
|
}
|
|
|
|
if (decode_bmv_frame(c->stream, pkt->size - (c->stream - pkt->data), c->frame, scr_off)) {
|
|
av_log(avctx, AV_LOG_ERROR, "Error decoding frame data\n");
|
|
return AVERROR_INVALIDDATA;
|
|
}
|
|
|
|
memcpy(frame->data[1], c->pal, AVPALETTE_SIZE);
|
|
frame->palette_has_changed = type & BMV_PALETTE;
|
|
|
|
outptr = frame->data[0];
|
|
srcptr = c->frame;
|
|
|
|
for (i = 0; i < avctx->height; i++) {
|
|
memcpy(outptr, srcptr, avctx->width);
|
|
srcptr += avctx->width;
|
|
outptr += frame->linesize[0];
|
|
}
|
|
|
|
*got_frame = 1;
|
|
|
|
/* always report that the buffer was completely consumed */
|
|
return pkt->size;
|
|
}
|
|
|
|
static av_cold int decode_init(AVCodecContext *avctx)
|
|
{
|
|
BMVDecContext * const c = avctx->priv_data;
|
|
|
|
c->avctx = avctx;
|
|
avctx->pix_fmt = AV_PIX_FMT_PAL8;
|
|
|
|
c->frame = c->frame_base + 640;
|
|
|
|
return 0;
|
|
}
|
|
|
|
AVCodec ff_bmv_video_decoder = {
|
|
.name = "bmv_video",
|
|
.long_name = NULL_IF_CONFIG_SMALL("Discworld II BMV video"),
|
|
.type = AVMEDIA_TYPE_VIDEO,
|
|
.id = AV_CODEC_ID_BMV_VIDEO,
|
|
.priv_data_size = sizeof(BMVDecContext),
|
|
.init = decode_init,
|
|
.decode = decode_frame,
|
|
.capabilities = AV_CODEC_CAP_DR1,
|
|
};
|