USB device for composite video capture at full res/interlaced?

[wonkey_monkey]

I recently ordered the cheapest USB/composite video capture device I could find, thinking at the very least it might just give me raw video (it claimed it supported "MJPEG and YUV422"), but it turned out to be terrible - only recording one field, not seeming to provide full res (720x576) to third-party software, and so on. And 90% of other listings just seem to be clones of the same one.

So can anyone recommend a decent USB capture device that'll work on Windows 10 and preserve interlacing? Preferably raw or MJPG rather than MPEG.

[Emulgator]

Hauppauge USB-Live 2, 8-bit 422 uncompressed possible, runs here on Win10.
https://forum.doom9.org/showthread.php?t=183936

[lollo2]

Yes, Hauppauge USB-Live 2 and IOData GV-USB2 are two excellent cards, available new and working flawless with modern OSs!

[x264N00b]

I compared my cheap APLIC UBS grabber with the Hauppauge USB-Live 2 because it's recommended in forums so often. Captured in FFV1 YUV 4:2:0 8bit at ~50Mbps with OBS there was almost no difference in PQ. (PAL)
It think I did the Hauppauge capture with YUV 4:2:2 8bit, not sure. The ALIC uses STK1160, SAA7115H and ALC655 chipsets.

Unfortunately my VCR was only able to output compostite video and not s-video. Maybe that was the bottleneck?

[lollo2]

Originally Posted by x264N00b

Unfortunately my VCR was only able to output compostite video and not s-video. Maybe that was the bottleneck?

The comb filter of Hauppauge USB-Live 2 is poor, so it is not reccomended to use this card with a composite input. However, when fed with a Y/C signal is excellent.

Always remember that the most important element is the player (not the capture card), which should be a high end S-VHS VCR with TBC.

Capture YUV 4:2:2 lossless interlaced from Y/C output with AmarecTV (recommended) or VirtualDub.

[rgr]

Originally Posted by x264N00b

I compared my cheap APLIC UBS grabber with the Hauppauge USB-Live 2 because it's recommended in forums so often. Captured in FFV1 YUV 4:2:0 8bit at ~50Mbps with OBS there was almost no difference in PQ. (PAL)
It think I did the Hauppauge capture with YUV 4:2:2 8bit, not sure. The ALIC uses STK1160, SAA7115H and ALC655 chipsets.

Unfortunately my VCR was only able to output compostite video and not s-video. Maybe that was the bottleneck?

First of all, you should not use OBS for capturing.

[x264N00b]

Can you please explain why? It's not very helpful when people just posting some advice without any explanation. Initially I used virtualdub but I always got big audio synchronization problems, doesn't matter what I tried. So I switched to OBS and it worked very well. I think there was also no difference in picture quality.

[rgr]

There's been so much written about this.. . VirtualDub and AmarecTV give you full control over the captured image, especially with advanced diagnostics. If you experienced desync in VirtualDub, it was probably due to settings or other issues. OBS (I'm guessing here) is probably skipping frames or duplicates them to mask the problem. It's hard to say, as it's a poor program designed for streaming; it's completely unsuitable for VHS capture.

I tested VirtualDub with Hauptpage and using a VCR with TBC, I had perfect results - 100% captured frames k with perfect synchronization. Link below.

[rgr]

Originally Posted by wonkey_monkey

I recently ordered the cheapest USB/composite video capture device I could find, thinking at the very least it might just give me raw video (it claimed it supported "MJPEG and YUV422"), but it turned out to be terrible - only recording one field, not seeming to provide full res (720x576) to third-party software, and so on. And 90% of other listings just seem to be clones of the same one.

So can anyone recommend a decent USB capture device that'll work on Windows 10 and preserve interlacing? Preferably raw or MJPG rather than MPEG.

https://forum.videohelp.com/threads/...-August-VGB300

If you have a VCR with only a composite output (or without TBC), a Digital8/miniDV camcorder with a video input will be the optimal choice.

[lollo2]

We checked OBS captures in VideoHelo forums many times, and always found weird video architectures, that's why is not suggested as capture software. Search that forum for evidences and facts!

If you have asynch in VirtualDub, first of all you should insure that you have a time base corrected signal. If, despite TBCs in the workflow, you still have asynch, you can switch to AmarecTV, because VirtualDub with modern OSs and modern cards tends to introduce a/v asynch, which can simetimes be mitigated by playing with its setting. Sometimes.

OBS may keep marginally synch by dropping/inserting frames, which is bad. Forget about it!

[rgr]

VirtualDub doesn't introduce any desynchronization, and if it does, it's due to incorrect settings. And if it persists despite these changes, you need to investigate your system further, as the problem lies deeper. Using a different program won't fix anything; it will only mask the problem. In that thread, I provided the correct settings for Hauptpage and what needs to be changed for the UGREEN HDMI grabber.

[rgr]

And lack of TBC doesn't cause any A/V desync issues. It only reduces frame drops (even to 0), and improves a few other things (long story).

[lollo2]

Originally Posted by rgr

VirtualDub doesn't introduce any desynchronization, and if it does, it's due to incorrect settings. And if it persists despite these changes, you need to investigate your system further, as the problem lies deeper.

False, there are many many evidences of that!

Originally Posted by rgr

And lack of TBC doesn't cause any A/V desync issues. It only reduces frame drops (even to 0), and improves a few other things (long story).

Absolutely false. Dropped frames (not compensated by the capture software) because lack of TBC is the main reason for a/v asynch. Since decades!

[rgr]

You're wrong.
I did about 10-20 test captures with VirtualDub, with various devices, from VCRs with and without TBC, using a DVD recorder as TBC, from its analog sources, and from HDMI. The slight offsets were only detectable on a frame-by-frame basis, completely undetectable by ear.
I encourage you to familiarize yourself with the test results instead of repeating nonsense.

[lollo2]

The only non sense are from you. I did hundred and hundred of captures and tests, so I report the facts.
I encourage you to study and practice the subject, to avoid spreading false information!

[rgr]

It doesn't matter how many tests you run – what matters is whether they are solid, repeatable, and reliable.

As you can see, yours weren't...

[lollo2]

The opposite.

As everybody can conclude by studying and reading and practicing about the subject, is your test that is flawed; without any doubt!

[rgr]

Originally Posted by lollo2

The opposite.

As everybody can conclude by studying and reading and practicing about the subject, is your test that is flawed; without any doubt!

The tests are flawed.
- But why?
- Because it's bad.
- But why??
- Everyone can see it's terrible!

I know that kind of argument; you're not improving.

[lollo2]

Sure!