View Full Version : Help determining the type of source
kalehrl
3rd September 2024, 12:39
I have a cable capture that seems to be progressive for the first part and then after commercial break it turns to interlaced according to my eyes. Can you have a look? I cut the first part before the commercial break and after it. After indexing it, you can see "mouse teeth" in the second part. Thanks.
https://www.mediafire.com/file/1oarjrev9mrtwwd/Otpisani+E01+(2).mkv/file
johnmeyer
3rd September 2024, 15:51
I see no change. It does, however, seem to be sped up by more that I would expect from the 24 fps to 25 fps film to PAL conversion. I saw no interlacing
poisondeathray
3rd September 2024, 16:21
After the fade to black, the fade in has a single orphaned field of film content paired with a black field (ie it' s missing the other film field) - that causes the combing . Likely it was edited incorrectly.
kalehrl
3rd September 2024, 16:40
As I wrote, there is a commercial break where that black frame is. I used mkvtoolnix to cut out the commercials. This is a PAL TV series, just remastered and aired for the first time on TV. Weird because I can definitely see interlacing after the black frame in the rest of the footage.
poisondeathray
3rd September 2024, 17:22
The differences are likely the source filter . I used DGDecNV
The version of LSmash I'm using has problems with this source and I seem combing in almost all the frames after as well
The version of ffms2 I'm using is more similar to DGDecNV in this case, but there are some block decoding errors, or error correction issues
kalehrl
3rd September 2024, 17:50
I'm also using lsmash works. The one which comes with MeGUI. I'll try ffms2.
kalehrl
3rd September 2024, 19:17
With ffms2 the file is indexed properly and there are no "fake" interlacing artifacts.
The video is pure progressive, 25fps.
poisondeathray
3rd September 2024, 19:32
With ffms2 the file is indexed properly and there are no "fake" interlacing artifacts.
The video is pure progressive, 25fps.
There is an orphaned field in your sample, thus 1 frame is combed . ffms2 is like DGDecNV in that regard . But the ffms2 version I'm using has some decoding errors - frame is not 100% clean, some blocks. The presence of 1 combed frame suggests there might be orphaned fields from the way it was edited. Look around the commercial breaks on a longer sample
If it were me, I would use DGSource with TFM(pp=0)
kalehrl
3rd September 2024, 20:25
Yes, that frame is from cutting with mkvtoolnix gui which isn't really frame accurate even though I chose key frames.
It must have been from the leftover commercial break.
I first tried using avs cutter in MeGUI but since I was using the default indexer lsmash which created a bad index, cutting with avs cutter was even less accurate than with mkvtoolnix.
Now I indexed with ffms2 and avs cutter is frame accurate so that bad frame is no longer present.
DGSource is for Nvidia cards? I don't have a discreet graphics card.
vBulletin® v3.8.11, Copyright ©2000-2026, vBulletin Solutions Inc.