KoTH x Eva: decode vs. traditional capture comparison?
 

A bit of a mirror from the videohelp post I made, but this is the closest I can do to a "proper-ish" traditional and decode comparison using the King of The Hill x Neon Genesis Evangelion mashup video created by Gamblor. Bet it would be considered outdated and invalid anyway since the decode video is a year old, oh well..

Is it perfect? No, the decode sample isn't mine and was sourced from a YouTube upload back in 2024. And my sample wasn't done using a VHS tape and VCR, but a Video8 tape and a Hi8 Camcorder. Though I intentionally recorded the footage onto video8 instead of Hi8 since the original YouTube upload said it recorded to an old Disney VHS tape, not S-VHS.

I digitized the video8 tape using a Sony CCD-TRV87 Hi8 Camcorder (it was recorded on a CCD-TRV66) Cypress CDM 640 Multisystem Converter and Pinnacle 510 USB capture card. For the Pinnacle's Proc Amp, Brightness was set to 10 and contrast was set to -18. The Cypress had it's saturation turned down by three stages because by default it is oversaturated, turning it down by 3 stages makes it more in line with how the colors look with the Cypress out of the chain.

Here are some image comparisons between mine, the YouTuber's decode capture and the original source video (for the first link only). All I did to my sample here was crop it so the proportions closely match the Decode version.

And here are video samples. Yes they were encoded to mp4s using VirtualDub2 due to file size limits, and for my sample the only restoration work I did was resizing and removing the blended fields from my capture, which was caused due to the source video being at 25 FPS and I was playing it back on an NTSC PlayStation 3 & Camcorder.

From the thread you are talking about it’s kinda crazy that even people like Brad have trouble making a comparison that everyone agrees is fair. I have seen threads where Nicholas Serra had trouble to and I’m thinking if these people are having problems then how tough is it to get working? Those are both very knowledgeable people.

I've theorized that the recording of a digital file to an analog tape and then capture from that would be a fair way to provide a more or less identical starting source for different capture comparisons. Issue is, it is hard to standardize how to create the starting tape and much of it will depend on the way that you made the digital file into analog signal to begin with which likely wasn't the same for both of you. My solution to that would be to have the original file as a downloadable DVD image and the DVD player itself is what handles the analog conversion. Problem there is there could be a fair degree of variance on how analog output can vary on different makes/models of DVD players. Further, there'd be variances in how well the deck making the recording does so. I would say that the best chance a player has is to play back a tape of a recording that it made itself as long as the heads aren't worn, etc.

Really the only way to do a fair comparison would be to do VHS_Decode with the same exact physical tape. it would be even "more fair" if the same VHS player was also used to play that tape for both Decode and standard capture, but it'd also be reasonable to use Decode on an inexpensive VHS player without a TBC and then compare that to a higher end SVHS machine with line TBC as that's the typical stated use case of Decode - to get "better" results than high end equipment. Tradeoff is still processing time, difficulty of initial setup, possibility of things no longer working after a software update (happened to me), and large files for intermediate stages.

My guess is the reason you didn't do your own Decode-capture is because it is not particularly user friendly to get going initially. The pitch used to be simpler where it was just do do some easy/light mods to a CX card, but then clockgen mods and other things became "standard" recommendations, so with each of those comes additional complexity. I would have done more comparisons myself, but I haven't been able to get it working on recent versions of MacOS and my windows machines are older and probably not up to the task.

Yeah I apologize, I know it isn't fully fair. I just don't want to bother getting a decode capture setup, I'm not the kind of guy that likes to physically handle electronics. Can use a laptop, but I don't deal with needing to add or replace parts for example.

That and I feel like I'm being able to get quality results with the equipment I have anyway, though if anybody wants to take a crack at this, feel free to do so

And I'd say the comparison Brad did make is so far the most proper fair comparison out there, Lollo did a great job looking at it. Main take away being that the decode sample had artifacts the traditional did not (ghosting/ringing, most noticeable around the disclaimer text) and the text itself being sharper with the traditional capture. The CGI screenshot is debatable. Latreche34 said that the decode sample has an extra layer of noise that disguises itself as more detail.

Ringing and haloing is still a problem, yeah it can be tweaked but, is it a problem that the artifact happens to begin with? Doesn't happen with conventional captures unless the ringing is baked into the tape, ie a TV recording with poor reception.

Heck, Brad himself said the ringing around text was why he didn't bother showing screenshots of the decode captures he did of an EP Mode TV Recording and Camcorder tape.

The bigger the sample size the better.

It's analog. Nothing will ever be 100% "fair" because there's too many variables. So there will always be people (or "experts") that try to cherry-pick data to justify the conclusion they reached beforehand. Let obvious visuals be your guide.

I'm reminded of one of my favorite quote, from the movie JFK. "Theoretical physics can also prove that an elephant can hang off a cliff with its tail tied to a daisy! But use your eyes, your common sense."

- Most people will see your sample as the obvious proof that it is: vhs-decode is lacking.
- Others will use mental gymnastics to insist vhs-decode is better.

Do you mean a "longer" sample? :hmm:

I'd say comparing a frame to the original video is also plenty to show that my capture didn't lose that much in terms of fine details. Yes it was sourced from a 720p YouTube Upload and is animation, which just has less fine details compared to live action.

I meant it’s better if possible to base a conclusion off more than one tape. The larger the number of tapes that conclusion is based on the better. In this case there aren’t that many people that have recommended refurbished gear that is in good shape that also have decode figured out so it can’t be based off of a lot of tapes.

Okay, I know this isn't from the same source video (this time being Evangelion x Tribute by Tenacious D) and isn't about a decode comparison, and more so a comparison between the original 1080p YouTube video and my capture of it on Hi8 tape.

But this to me says "what can decode bring to this quality wise that my traditional setup can not achieve?" I don't think much, if at all.

https://imgsli.com/Mzk3ODM2/2/3

https://imgsli.com/Mzk3ODM3

I don't trust imgsli.com for the long-term, so always attach pictures on this forum as well.