[x264-devel] Rare macroblock errors in recent x264...

[BugMaster]

On Tue, 17 Jul 2012 07:43:31 -0700, Jeppe Øland wrote:
> On Tue, Jun 12, 2012 at 4:33 PM, Jeppe Øland <joland at gmail.com> wrote:
>>> I'm thinking some bugs might have crept into the latest version of
>>> x264 (or recently at least).
>>> I have used r2200 for a few weeks now, and very occasionally the
>>> resulting file will have macroblock errors for a few frames.
>>> So far I have only spotted it far into the encoded files (like 1+ hour
>>> in), and the macroblocks are only wrong for 1 or 2 frames.
>>> (And when they are wrong, they don't look crazy bad either)
>>> Input file has been confirmed working fine.
>>>
>>> I'm going to see if I can narrow the problem down ... but just
>>> encoding 30 seconds leading up to a known trouble spots has not
>>> worked.
>>>
>>> Could r2197/2198 be causing problems?
>>
>> Re-ran the test with r2197.
>> Problem still happens - though not in exactly the same place.
>> (I guess that makes sense since r2200 would change the encode).
>>
>> Any ideas for versions to start trying to narrow it down?

> Looks like a few mails to the group disappeared ... repeated here:

> I did some more narrowing down:
>     r2200 = Problem.
>     r2197 = Problem.
>     r2184 = Problem.
>     r2164 = Problem (was much harder to spot with this version).
>     r2145 = Haven't seen a problem yet ... will continue looking.
>    (I don't have a lot of time for it at the moment, so it will take a while)

>> I did some more narrowing down:
>>    r2164 = Problem (was much harder to spot with this version).
>>    r2145 = Haven't seen a problem yet ... will continue looking.

> Is there anywhere I can download Windows in between these so I can
> narrow down the change that causes the problem.
> If not, I guess I'll have to install whatever recommended environment
> needed to build it (unless Linux can cross-compile for Windows).

> Regards,
> -Jeppe

As Jason wrote you last month we need sample. Preferably short source
sample with which we can reproduce your problem. If you can't provide
source sample than at least provide short encoded sample with error so
we can at least see it and analyse encoded stream (what caused error:
high QP, wrong prediction, etc.)
Also your problem can be cause not by encoder but decoder or system
instability (for example RAM errors) due overclocking.