mirror of https://git.ffmpeg.org/ffmpeg.git
screenpresso: Correctly handle keyframes
The first byte contains compression level together with keyframe status. When a frame is not interpreted correctly, its data is summed to the reference, and would degrade over time, producing an incorrect result. Signed-off-by: Vittorio Giovara <vittorio.giovara@gmail.com>
This commit is contained in:
parent
95db8c757c
commit
ec8a69fab6
|
@ -115,10 +115,9 @@ static int screenpresso_decode_frame(AVCodecContext *avctx, void *data,
|
|||
return AVERROR_INVALIDDATA;
|
||||
}
|
||||
|
||||
/* Basic sanity check, but not really harmful */
|
||||
if (avpkt->data[0] != 0x73 && avpkt->data[0] != 0x72)
|
||||
av_log(avctx, AV_LOG_WARNING, "Unknown header 0x%02X\n", avpkt->data[0]);
|
||||
keyframe = (avpkt->data[0] == 0x73);
|
||||
/* Compression level (4 bits) and keyframe information (1 bit) */
|
||||
av_log(avctx, AV_LOG_DEBUG, "Compression level %d\n", avpkt->data[0] >> 4);
|
||||
keyframe = avpkt->data[0] & 1;
|
||||
|
||||
/* Pixel size */
|
||||
component_size = ((avpkt->data[1] >> 2) & 0x03) + 1;
|
||||
|
|
Loading…
Reference in New Issue