On Thu, Apr 04, 2024 at 09:51:50AM -0300, James Almer wrote:
> On 4/4/2024 9:29 AM, Michael Niedermayer wrote:
> > On Thu, Apr 04, 2024 at 08:50:12AM -0300, James Almer wrote:
> > > On 4/4/2024 8:47 AM, Michael Niedermayer wrote:
> > > > On Thu, Apr 04, 2024 at 09:12:04AM +0200, Anton Khirnov wrote:
> > > > > Quoting Michael Niedermayer (2024-04-04 00:57:39)
> > > > > > Hi
> > > > > > 
> > > > > > I will try to make the 7.0 release from the release branch in the next 48h
> > > > > > (with some luck but easy possible it will get delayed)
> > > > > > i will not try to fix any issues marked as blocking 7.0 on trac, IIRC
> > > > > 
> > > > > IMO people are overly eager to mark everything as 'blocking'.
> > > > 
> > > > No, thats not true, very few bugs are marked as blocking, we have 2837
> > > > open/new bugs and 1 marked as blocking. And we had maybe 2-3 marked as
> > > > blocking maximum at any time over the last weeks
> > > 
> > > The bug marked as blocking has no available sample, and the commit you say
> > 
> > https://samples.ffmpeg.org/ffmpeg-bugs/trac/ticket4161/
> > 
> > I dont understand why google hasnt indexed (or finding that) though
> > 
> > My testcase:
> > -i VR_MOVIE_GuyMartinsSpitfireBcast169\ qsf\ lappyAspectNoChnge_extract.mpg -bitexact -vcodec libx264 -an -vsync cfr file-16:9.ts
> > 
> > 
> > > generated the regression just moved a failure path around, so how can it
> > > change the reported aspect ratio of the input?
> > 
> > I have not investigated it, I just see that there are various circular features
> > in the video that are not circular anymore
> 
> I can't find a difference between 6.1 and master:
[...]
> Am i doing something wrong?

This bug is cursed, be carefull

git checkout  924a6f3cc7788e2d258348b1547a49805091ea2d^
make -j32
./ffmpeg -i ~/tickets/4161/VR_MOVIE_GuyMartinsSpitfireBcast169\ qsf\ lappyAspectNoChnge_extract.mpg -bitexact -vcodec libx264 -an -vsync cfr file-169-924a6f3cc7788e2d258348b1547a49805091ea2d^.ts

git checkout  924a6f3cc7788e2d258348b1547a49805091ea2d
make -j32
./ffmpeg -i ~/tickets/4161/VR_MOVIE_GuyMartinsSpitfireBcast169\ qsf\ lappyAspectNoChnge_extract.mpg -bitexact -vcodec libx264 -an -vsync cfr file-169-924a6f3cc7788e2d258348b1547a49805091ea2d.ts

ls -alF file-169-924a6f3cc7788e2d258348b1547a49805091ea2d.ts file-169-924a6f3cc7788e2d258348b1547a49805091ea2d^.ts
-rw-r----- 1 michael michael 1255652 Apr  4 18:32 'file-169-924a6f3cc7788e2d258348b1547a49805091ea2d^.ts'
-rw-r----- 1 michael michael 1255652 Apr  4 18:33  file-169-924a6f3cc7788e2d258348b1547a49805091ea2d.ts

md5sum  file-169-924a6f3cc7788e2d258348b1547a49805091ea2d.ts file-169-924a6f3cc7788e2d258348b1547a49805091ea2d^.ts
9ec90b2497787f7d5d76aa1d34cde7eb  file-169-924a6f3cc7788e2d258348b1547a49805091ea2d.ts
4361f0f20181c39a3a9b3a7f35619520  file-169-924a6f3cc7788e2d258348b1547a49805091ea2d^.ts

now if i play these with ffplay

./ffplay file-169-924a6f3cc7788e2d258348b1547a49805091ea2d.ts
./ffplay file-169-924a6f3cc7788e2d258348b1547a49805091ea2d^.ts

and one of these simply has the wrong aspect ratio
i never looked at anything else but i now also see in the ffplay output

Stream #0:0[0x100]: Video: h264 (High) ([27][0][0][0] / 0x001B), yuv420p(progressive), 720x576 [SAR 16:15 DAR 4:3], 25 fps, 25 tbr, 90k tbn
vs.
Stream #0:0[0x100]: Video: h264 (High) ([27][0][0][0] / 0x001B), yuv420p(progressive), 720x576 [SAR 64:45 DAR 16:9], 25 fps, 25 tbr, 90k tbn

thx

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Asymptotically faster algorithms should always be preferred if you have
asymptotical amounts of data