Some Newbie Questions About Video Capture
 

Hi all!

I'm new to the whole video capture setup, so apologies in advance if I ask any trivial questions. The FAQs on this site have so far been very helpful!

After discovering that some of our family 8 mm camcorder tapes are already starting to degrade, I've embarked on a mission to digitize it all. Before I sit through 30+ hours of home video though, I really want to nail my setup so I do it right the first go round.

My current setup:
I am playing back our 8 mm cassettes via the original Sharp camcorder component output. This part of my setup is unlikely to change, as it seems the head on the camcorder is misaligned just enough so that the tapes don't play back correctly on other (properly aligned) 8 mm players. I have a Hauppage WinTV-PVR-150 PCI card for video capture, which is working fine on Windows 10 using their WinTV program. Based on my research, this card has a hardware MPEG-2 encoder which spits out MPEG-2 TS files to my hard drive. I am using Handbrake to crop the picture slightly (...tracking issues) and re-encoding it as H.264 for storage and ease of playback (I will keep the original TS files though).

My questions:

• Do you think there will be a significant increase in quality by investing in an ATI/other alternative capture hardware? I assume the benefit will be that I can capture a lossless stream (whereas the WinTV-PVR-150 is a lossy hardware MPEG-2 stream). The source media isn't exactly the greatest quality though, and resolution-wise the PVR-150's captures seem alright - just a bit desaturated, which I can fix in Final Cut.
• Likewise, is my Hauppage card likely better or worse than these $30-50 USB dongles available on Amazon? I could seen investing up to $100-150 to improve my capture quality.
• Lastly - might be the wrong forum for this - but I want to pump the saturation a little bit in post, given that the output from the PVR-150 is a little grey. However, Final Cut doesn't import MPEG-2 natively. I have tried Premiere Pro, which does, but ends up with some really weird looking sharpening artefacts after import. Any experience with this? Or will I just have to continue converting to H.264 as an intermediary?

Sorry - realize that was pretty wordy. My basic question: is the WinTV-PVR-150's MPEG-2 quality considered reasonable in 2018? Or with a budget of ~$100 is it worth investing in some new hardware to convert my old 8mm tapes?

Some sample images (original MPEG-2 TS file left; after processing, sharpening, cropping, and re-encoding with Final Cut on right - I realize I probably went a little far with the saturation now): https://imgur.com/a/9A8zroT

Welcome to digitalfaq :salute:

I echo your sentiments and I'm glad you find that the quality approach and pro-inspired methods that are the target of the guides have helped you in your video projects. While some of the guides are a bit dated in spots (hardware and software have changed somewhat nowadays), the basic principles have remained the same.

However I'm confused about which guides you refer to. I'm not aware of any Faqs, posts, or threads that recommend your hardware, software, or working media. In fact, the guides and threads I've seen over the years in this and other forums advise against many of your choices.

Considering the alignment issue, it appears you've made a sensible choice in playback gear.

Undoubtedly, lossless media would be an improvement.

You seem to imply that low quality source deserves only low quality processing that makes the results look as poor as the source, or worse. I disagree. My main misgiving about your posted images is that they look like SD source blown up with software to HD proportions, which does not look like HD.

Again, you have me wondering which of our guides you refer to. We have recommended a few updated capture devices at less than $100 that satisfy our criteria, are used by many of our members who've posted samples here, and which compare favorably with the legacy ATI devices we've recommended. What our users don't recommend are the cheap copies and EasyCap (aka Easy-Crap) ripoffs of those products.

If you have followed this forum or browsed previous threads you'd see that no one here has ever recommend Premiere Pro for anything more than cut-and-join and color correction. It is a poor performer in many respects, does simple tasks like colorspace conversions and resizing poorly, has no talent for deinterlace or denoising or other analog source fixes, and has never been recommended as a restoration platform. FCP is a little better in some respects, but it has similar limitations and is recommended for cut/edits, ciolor work, and final encoding only.

As forf modifying saturation, it appears to be largely a matter of preference. Doing so seems to have blown away some highlights in some of your images, but no too badly. Still, have you considered the effects of the color matrix for your HD output as opposed to the SD matrix used in your source? Capturing your YPbPr source to MPEG cost you 50% of your original color resolution. Color matrix and other issues can't really be analyzed well with lossy processed video and further processing of still images. One would need an unaltered .ts sample of the original to get into meaningful analysis.

Your mention of using h.264 as an intermediate working codec made me shudder with thoughts of the damage you've inflicted on your already lossy captures. MPEG2 as a capture format was never designed for edits or re-encoding. It is strictly a lossy interframe final delivery format. I still have no idea why you haven't chanced on better software and codec recommendations at digitalfaq for your purpose.

It wouldn't have been considered ideal for further processing or improvement in previous years, nor would it be recommended today if further processing is intended. Meanwhile, thank you for your posted images. We prefer video samples over still images, as this forum is focused primarily on video effects, meaning that analysis and comments are limited when based solely on still images. You appear to have controlled your capture signal levels and color balance fairly well, which are area where newcomers usually blow away their efforts. Strong side lighting obscures considerable shadow detail in the images, which if they were video would prompt me to suggest some contrast masking to open up the shadows a bit. Unfortunately all that lossy processing sharply increases noise in darker areas.