OK, guys, test this:

http://rationalqm.us/misc/DGIndexNV_crop_fix.exe

Aha, the problem is not with cropping at all. The frame is not being displayed properly in the DGIndexNV window. In other words, it is a preview issue. It's a regression from the new code that supported main window resizing. I have a fix working but have to regression test before giving a test versi...

Guest 2 wrote: ↑

Wed Feb 10, 2021 3:18 am

All the streams but chapters.txt are present. Could you fix it?

Sure.

Please be more specific about the stream type and what is missing.

Looks like autocropping may have suffered a regression. Investigating...

I don't understand. You mean when serving with DGSource()? You have to either export the cropping to your script (see user manual) or give the cropping manually in your DGSource() invocation.

Or maybe you are saying that autocrop is having an issue with it?

Marking resolved.

In the case when the user, when indexing the source file, demultiplexes the video stream, most likely, he needs to work with the extracted video stream directly, and not with the original container file Please justify this claim. In any case doing this would be highly nontrivial so it's unlikely th...

Please re-download slipstream 226 and try again. The WAV file was not being closed properly.

Ah. I see the problem. Will fix it ASAP. Thank you for the report.

Again, I do not know what you mean by MPC-HC "cannot output it". You mean it cannot play it? I say again also that the demuxed WAV from your sample plays fine for me in MPC-HC.

Post your Avisynth script.

Maybe you should post your demuxed WAV so I can see if it is the same as mine.

Works fine for me. Program stream PCM gets demuxed to WAV. The resulting WAV plays in MPC-HC. I don't know what you mean by "cannot be output". What audio source filter are you using with Avisynth and what is the error?

That is normal. It's a CLI application, not a GUI. Don't you read documents? ;) From Notes.txt: ----- 1. This tool requires a donation to obtain the required license. Please use the Donate button on the DGDecNV page. To get your machine ID on a non-Nvidia system for licensing purposes, run this in a...

Thank you. Released as slipstream 226.

DGDecNV slipstream 226: * Fixed so that the PAR is consistently returned as the aspect ratio for all video types. :salute: Vraelon * Fix handling of streams with coded height 1088. This was a regression of the new cropping. :salute: Guest 2 * WAV files demuxed from program streams (e.g., VOB) were n...

Thank you, Sir!

I updated the Vraelon test build with the fix for this issue. Please test and advise.

http://rationalqm.us/misc/DGIndexNV_vraelon.rar

The cropping is misbehaving of course with coded 1088, but worse is that the black bar should not appear at all at the bottom (the video fills 1080). I have both fixed locally and will release with the aspect ratio stuff this evening or tomorrow morning.

Can't believe this wasn't noticed until now.

Thank you for your testing. I'll mark this resolved and make a slipstream after addressing Guest 2's cropping issue.

Please re-download the fix candidate and retest. Things should be OK with your sample.

Thank you. Downloading... Next time a smaller sample will be preferred.

Please link an unprocessed source sample that behaves that way.

Please don't make me ask every time for a sample.

Yes, complications from 1088 versus 1080. I'll do something to make the handling sensible for cropping. Thank you for the report.

Here is a candidate fix for all the aspect ratio stuff. We were doing different things for the different video types. Also, the stream syntax elements differ for the different video types. This candidate fix revises all the video types so that: 1) the Info dialog shows the PAR (pixel aspect ratio), ...

Got it. Gonna fix it tomorrow maybe with an explanation.

So hard to be perfect. Fall short and they'll be taking X-rays.