Wrongly captured PAL VHS at 50fps... no chance to deinterlace?
 

As many around the forum, I have already behind me countless hours of trial & error, searching for tips, knowledge and guides regarding the unthankful adventure of capturing VHS for archival and sharing purposes.

I got the capture part sort of covered (good JVC SVHS, TBC), and was very happy with using the IOData GV-USB2 capture device.

After getting through around 15 tapes, I had the need to encode one of them for sharing purposes. To my dismay, I realized I had captured in 50fps instead of 25fps.

I had not realized that the drivers for the capture device and AmaRecTV combination I used led to the option of capturing in 25fps simply not appearing. I finally found a guide informing about the situation and how to avoid it here: https://aaproductions.net/gvusb2.htm

So I tried just as a test to capture one tape in 25fps. As expected, I got an interlaced outcome, which could "easily" be deinterlaced following the guidelines for Hybrid / QTGMC and outputing to smooth 50fps (bob).

My problem: I have no idea how to "save" my 50fps captures. Using the same encoding method as above, the final encoded file has 100fps - my eyes tell me it looks good, I have no double frames when frame skipping after encoding, but I guess this is totally wrong.

Is there a way to "overwrite" somehow the captured files and "make" them 25fps without losing information? From what I tried so far, I either convert the files via VirtualDub to 25fps and get a "slow motion" video, or I convert to 25fps by "decimating by 2", leaving out half the information I had.

Can these 50fps interlaced videos be treated somehow as 25fps for further processing?

Apologies in advance for the lack of knowledge!

Those "50 fps interlaced" videos are actually 50 fps progressive, i.e. they are 50 full frames per second.

As you've described, when you captured at 50 fps you captured progressive and not interlaced. It has done what a deinterlacer such as QTGMC would have done by converting 50 fields per second (which add up to 25 full frames per second when interlaced together) into 50 full frames per second by interpolating information that wasn't there to generate twice as much frame data and also remove the "jagged edges" effect you can see when playing back interlaced video on a progressive display with a player that doesn't properly deinterlace the video. The jagged edge removal is something QTGMC would do. Although, you may be seeing jagged edges baked into the 50 fps progressive file your capture card generated.

The analog video on the tape is 50 fields per second. A field is half of a full frame. The fields alternate at an even/odd or odd/even cadence, depending on the field order. Each even/odd pair represent the even/odd numbered rows of video in an interlaced frame. You can observe this in Virtualdub with an Avisynth script that calls SeparateFields(). This will show you each of the fields captured. You'll notice the video is now twice as many "frames" in the virtualdub counter, but the video is half as tall.

If I were you, I'd recapture at 25fps to get interlaced. QTGMC would be better at deinterlacing than your capture card, if you wanted progressive video in the end. And if you are archiving these videos in a high bitrate or lossless format, I'd want them archived in interlaced format if I ever wanted to work on them again in the future as technology progresses.

A forum search will show examples of using SeparateFields().SelectEvery(4,0,3) on the output from QTGMC to return it back to an interlaced video at the interlaced frame rate of 25 interlaced vs 50 progressive. This is done to get the benefits of QTGMC cleanup, but ultimately returning the video to it's original frame rate and interlaced format. You could do this on your 50 fps progressive video to get back an interlaced 25 fps video. But this is 25fps interlaced in spec only. Any destructive operations done by your capture has already been baked into the file, even if you separate it back into "interlaced".

Hope this helps. Best of luck to you.

Hello keaton, and thanks first of all for the detailed answer.

I will go through your suggestions in more detail, but I can already say that the jagged edges / combing effect are in fact "backed it".

My capture device did not have any "real-time" deinterlacing setting activated during capturing, but you must be right in a sense that, if the vcr can only "output" 25fps / 50 fields per second, then my capture device had to "interpolate" and insert two frames for each second of capture, no?

Your last paragraph points me in a good direction, which would be to "process" the video and discard the frames. What I think I will do, to really understand if I am happy with this "processing back to 25fps interlaced", is maybe do it in a sample and then recapture the same piece of video properly at 25fps, and compare both.

Thank you once more!

You're welcome. I find it is essential to at least understand the basics when I'm working with something. I find it's best to communicate those basics before getting into answers or opinions. That gives you the opportunity to decide for yourself what's important and how you want to proceed. Ultimately it is each individual's opinion what is "good enough" and how much detail to get into. You can really get into details when doing this kind of work. Some will want to go very deep into details and have the time and dedication to do so. Some will want to do things as expeditiously as possible.

Correct, the capture device has to be inserting an additional field for every field it scanned from the VCR to generate 50 full frames per second. Every 20 milliseconds, a field has finished being transmitted from the VCR's output. With so little time between fields, perhaps it is just line doubling each field, i.e. duplicating the even rows as odd rows or vice versa and putting those two fields together for a frame. That would be the simplest form of interpolation. Although video players can use fancier deinterlacing methods during real time playback of interlaced video. But perhaps that's done in the video graphics card drivers, which have much more processing power than a capture device. Not familiar with the hardware details. Of course, QTGMC is much more processing intensive and cannot be done in real time.

That's a good idea to compare a native 25 fps interlaced capture with a 50 fps progressive capture that is then "re-interlaced" back to 25 fps. If you find that using something like SeparateFields().SelectEvery(4,0,3) produces comparable results to what a native 25 fps capture does, and your happy with it, then you've saved yourself a lot of work. But if you find something that you don't like such as baked in combing that cannot be undone, then you understand why you need to recapture.

I ran into a similar bug with my GV-USB2, but I was capturing NTSC so it was capturing at 59.94fps instead of 50, and each frame was duplicated. To fix it, I typed in 29.97 into the UI. Strangely, when go back to check, it shows 59.94 again, but at least it captures correctly, and you only have to do it once.

I think decimating every other frame is the correct fix for already captured video, but I recaptured anyway.