[MEncoder-users] [mpeg4 @ 0x86f46cc]Error: 2pass curve failed to converge

Corey Hickey bugfood-ml at fatooh.org
Tue Oct 24 23:17:01 CEST 2006


Trent Piepho wrote:
>>> -    if(fabs(expected_bits/all_available_bits - 1.0) > 0.01 ){
>>> +    if(abs(expected_bits/all_available_bits - 1.0) > 0.01 ){
>> But fabs are very different from abs. For the first, the result of abs
>> is int, then you shouldn't compare it with 0.01, it should be 0. And
>> the parameters of abs should be integer also. Then, I think probably
>> the if will be always evaluated as false.
> 
> The original if statement is equivalent to this:
> 
> if(expected_bits >= all_available_bits*2){
> 
> ie.g expected_bits must be double all_available_bits or more.  That will
> pretty much never happen.  Basically, the original "2pass curve" check
> was disabled and that's why it didn't cause any problems.
> 
> I think it would make more sense to change to check from an error to a
> warning.  Something about how the maximum allowed quality level doesn't use
> all available bits.

Are my mails going to /dev/null or are you guys just not paying 
attention? I don't mean to sound rude, but I've already explained 
several times that _this_problem_has_been_fixed_in_SVN_already! :)

...meanwhile, between the second and third time I said the problem is 
fixed in SVN, 1.0rc1 was released, with the fix...

Read the SVN log message!
http://lists.mplayerhq.hu/pipermail/ffmpeg-cvslog/2006-September/003471.html

[cut]
+    if (toobig == 0) {
+        av_log(s->avctx, AV_LOG_INFO,
+            "[lavc rc] Using all of requested bitrate is not "
+            "necessary for this video with these parameters.\n");
[cut]

Sheesh. :)

-Corey



More information about the MEncoder-users mailing list