mirror of
https://github.com/FFmpeg/FFmpeg.git
synced 2025-07-01 05:31:04 +08:00
correct quantizer bias
Originally committed as revision 438 to svn://svn.ffmpeg.org/ffmpeg/trunk
This commit is contained in:
@ -245,6 +245,7 @@ int mjpeg_init(MpegEncContext *s)
|
||||
|
||||
s->min_qcoeff=-1023;
|
||||
s->max_qcoeff= 1023;
|
||||
s->intra_quant_bias= 1<<(QUANT_BIAS_SHIFT-1); //(a + x/2)/x
|
||||
|
||||
/* build all the huffman tables */
|
||||
build_huffman_codes(m->huff_size_dc_luminance,
|
||||
|
Reference in New Issue
Block a user