[MPlayer-dev-eng] -lavcopts=bit_exact patch
Steven M. Schultz
sms at 2BSD.COM
Fri Feb 6 23:50:55 CET 2004
Greetings -
libavcodec has the option "bit_exact". The comment for that option
(in libavcodec/opts.c) says "use only bit-exact stuff" - HOWEVER,
the 'bit_exact' flags also turns off non-standard behaviour which
causes playback problems with some programs (Apple's Quicktime Player
for example).
I think it would be a good idea to expose that part of libavcodec's
API to mencoder.
I have attached a small patch which implments "-lavcopts=bit_exact".
Is this in good format and otherwise acceptable?
Cheers,
Steven Schultz
-------------- next part --------------
--- libmpcodecs/ve_lavc.c.dist Tue Feb 3 20:47:37 2004
+++ libmpcodecs/ve_lavc.c Fri Feb 6 14:28:24 2004
@@ -120,6 +120,7 @@
static int lavc_param_dia_size= 0;
static int lavc_param_qpel= 0;
static int lavc_param_trell= 0;
+static int lavc_param_bit_exact = 0;
static int lavc_param_aic= 0;
static int lavc_param_aiv= 0;
static int lavc_param_umv= 0;
@@ -230,6 +231,7 @@
#ifdef FF_CMP_VSAD
{"ildctcmp", &lavc_param_ildct_cmp, CONF_TYPE_INT, CONF_RANGE, 0, 2000, NULL},
#endif
+ {"bit_exact", &lavc_param_bit_exact, CONF_TYPE_FLAG, 0, 0, CODEC_FLAG_BITEXACT, NULL},
{"predia", &lavc_param_pre_dia_size, CONF_TYPE_INT, CONF_RANGE, -2000, 2000, NULL},
{"dia", &lavc_param_dia_size, CONF_TYPE_INT, CONF_RANGE, -2000, 2000, NULL},
{"qpel", &lavc_param_qpel, CONF_TYPE_FLAG, 0, 0, CODEC_FLAG_QPEL, NULL},
@@ -560,6 +562,7 @@
#if LIBAVCODEC_BUILD >= 4648
lavc_venc_context->flags|= lavc_param_trell;
#endif
+ lavc_venc_context->flags|= lavc_param_bit_exact;
lavc_venc_context->flags|= lavc_param_aic;
lavc_venc_context->flags|= lavc_param_aiv;
lavc_venc_context->flags|= lavc_param_umv;
More information about the MPlayer-dev-eng
mailing list