Deinterlacing, compressing VHS captures for distribution?

[Spektre]

Hello all,

It’s been over a decade, but a family member gave me a batch of tapes they would like digitized. My brain’s got a lot of rust.

My setup is:
A Windows 10 2019 LTSC build
1T NVME capture drive
10T storage server to move to after each capture is complete.
ATI 600 USB stick
AG-1980P deck (That I hope is still OK after all the inactivity. I am not qualified to tear it down and clean/lubricate)
Ambery TBC-1 (I know these get a bad rap, but it has worked flawlessly over a few hundred hours of tape for me in the past.)

I downloaded Virtualdub and did a couple quick captures to make sure I remembered how to do all that.

A few questions come to mind.

1. In the intervening decade has virtualdub added the ability to listen to audio without terribly dropping frames? (Without doing some form of hardware splitting to listen to it from another pathway)

I am starting with Lagarith encoded 720x480 VHS captures.

Although everyone watches these “on a TV”, they come from a source (PC, media player) that outputs progressive video, thus requiring deinterlacing instructions to be handed out to the consumers.

I am therefore going to distribute this batch in deinterlaced, lossy compressed format (probably H.264).

2. What is a good deinterlacing algorithm that does not have to be performed in realtime that I can apply to these (Yadif looks visually acceptable to me when done at playback on VLC polayer)

3. Does Virtualdub do the deinterlacing and compression or is that an avisynth job (or some other tool, I’m not familiar with avisynth either)

Thanks in advance.

PS: Thinking about distribution media, what kind of compression rates does one usually see with H.264?

[Aya_Rei]

Quote:

Originally Posted by Spektre

A few questions come to mind.

1. In the intervening decade has virtualdub added the ability to listen to audio without terribly dropping frames? (Without doing some form of hardware splitting to listen to it from another pathway)

2. What is a good deinterlacing algorithm that does not have to be performed in realtime that I can apply to these (Yadif looks visually acceptable to me when done at playback on VLC polayer)

3. Does Virtualdub do the deinterlacing and compression or is that an avisynth job (or some other tool, I’m not familiar with avisynth either)

Thanks in advance.

PS: Thinking about distribution media, what kind of compression rates does one usually see with H.264?

I'll try and answer this best I can

1. Sadly no, while I've heard dropping frames with audio playback enable is card dependent, it's best to keep it audio playback while capturing disable. (Of course, keep capture audio enabled) Having the volume mixer enabled seems to be fair game, if you want to at least check you are getting audio from the source device.

2. You can apply QTGMC deinterlacing to your captured files after the fact, either using Avisynth itself or running the filters by using Hybrid from Selur. It is best to keep your capture files natively interlaced anyway.

3. It can, but it'd be better off to do all of that using Hybrid. Can do deinterlacing, compressing to mp4 (or any file/codec of your choosing) and use avisynth/vapoursynth filters.

[aramkolt]

Do you get dropped frames or audio sync issues when removing the Ambery TBC from the chain? Or is there any other difference you notice if not using it?

[Spektre]

Quote:

Originally Posted by aramkolt

Do you get dropped frames or audio sync issues when removing the Ambery TBC from the chain? Or is there any other difference you notice if not using it?

During my original run a decade or so ago, I had 2 problems tapes that caused dropped frames that did not drop frames once added.

The combination of this and the line TBC on the VHS deck (I don't know which fixed it) corrected some flag waving on other tapes.

It is also fairly adept and removing copy protection.

[Spektre]

Quote:

Originally Posted by Aya_Rei

I'll try and answer this best I can

1. Sadly no, while I've heard dropping frames with audio playback enable is card dependent, it's best to keep it audio playback while capturing disable. (Of course, keep capture audio enabled) Having the volume mixer enabled seems to be fair game, if you want to at least check you are getting audio from the source device.

2. You can apply QTGMC deinterlacing to your captured files after the fact, either using Avisynth itself or running the filters by using Hybrid from Selur. It is best to keep your capture files natively interlaced anyway.

3. It can, but it'd be better off to do all of that using Hybrid. Can do deinterlacing, compressing to mp4 (or any file/codec of your choosing) and use avisynth/vapoursynth filters.

Thanks. Appreciate it. I'll go take a look at Hybrid. A new tool for me.

[Aya_Rei]

Glad I could help, Hybrid is rather daunting and can appear to be rather scary and complicated, as it is for more advanced users as Selur states on his site.

Perhaps this tutorial from Video Capture Guides can help you ease into using it.

[vwestlife]

Are you really sure you need to de-interlace it? As long as the source video is properly marked in the file's metadata as interlaced, any media player program (Windows Media Player, QuickTime, VLC, etc.) should automatically de-interlace it, and usually does an excellent job of it (including full 50/60 fps motion) because it uses your graphics chip's hardware-based de-interlacing. I've compared this hardware de-interlacing to software-based de-interlacing, and it's better than everything except QTGMC, which is comparable to or slightly better than it.

If you use software-based de-interlacing, you're forever locked into whatever method you used, but if you keep it interlaced, you can always go back and do it again with better-quality methods as they become available.

[Spektre]

Quote:

Originally Posted by vwestlife

Are you really sure you need to de-interlace it? As long as the source video is properly marked in the file's metadata as interlaced, any media player program (Windows Media Player, QuickTime, VLC, etc.) should automatically de-interlace it, and usually does an excellent job of it (including full 50/60 fps motion) because it uses your graphics chip's hardware-based de-interlacing. I've compared this hardware de-interlacing to software-based de-interlacing, and it's better than everything except QTGMC, which is comparable to or slightly better than it.

If you use software-based de-interlacing, you're forever locked into whatever method you used, but if you keep it interlaced, you can always go back and do it again with better-quality methods as they become available.

Yes, I'm sure.

[Gary34]

Quote:

If you use software-based de-interlacing, you're forever locked into whatever method you used, but if you keep it interlaced, you can always go back and do it again with better-quality methods as they become available.

We aren’t ever really stuck because we capture interlaced losslessly compressed. We can always go back to the Huffy file and deinterlace, filter, then encode.

[Spektre]

Quote:

Originally Posted by Aya_Rei

Glad I could help, Hybrid is rather daunting and can appear to be rather scary and complicated, as it is for more advanced users as Selur states on his site.

Perhaps this tutorial from Video Capture Guides can help you ease into using it.

I am following a tutorial on Hybrid. but when I load in my video file, all the controls on the Deinterlace tab disappear.

Have you seen this before? Any ideas?

PS: Hybrid seems to think the file from my ATI 600 card from virtualdub is progressive.

[Aya_Rei]

Hybrid defaults to progressive, click on the Overwrite input scan type checkbox and set it to top field first

if working with Mini DV tapes, set it to bottom field first.