Topaz is the closest thing to magic I’ve seen in video upscaling, but, as of a year ago, it’s a finicky beast to get to work. It crashed a LOT, and is just about guaranteed to lose audio sync. They could have fixed it since then, but be prepared to spend a significant amount of time upconverting an episode.
2 Likes
Looked at it; I would have to buy a new rig to cope.
I saw one user saying it upscaled at a rate of one hour processing per minute of video. So for a 90-minute episode that’s either three days’ work or multiple machines.
I take it you’ve used it yourself? Have you tried the photography software also?
1 Like
“twice the information” is a little misleading, as the new information is fake, created by the software. If blow up a picture in photoshop and save a new file that’s twice as big, it doesn’t necessarily mean anything.
2 Likes
1 Like
Pu-Ma-Man
I still enjoyed it.
Pu-Ma-Man
A classic stinker.
Pu-Ma-Man
It’s just a bit sharper.
Pu-Ma-Man
Still flys like $%^&
4 Likes
Hmm. If the question is, “Should MST3K use Kickstarter money to enhance/improve old episodes?” I think my answer would be a cautious no, unless this use was clarified during fundraising. Those dollars are already stretched thin enough. However, I’m absolutely in favor of improving the quality of older episodes, ideally to the point where we could have Blu-ray releases. And I’d be very okay with contributing to a fund to make that happen. It’s really all about transparency.
Thankfully, the process is nowhere near as complex as lifting pieces out of old episodes and reassembling them. Current technology is good enough that that’s not needed.
That said, I trust Joel, Ivan and our other MST3K custodians. If we’re talking about spending 20% of a $6M Kickstarter on this, I think contributors ought to be aware of that beforehand. But that’s clearly not the case here, so if they want to spend a substantially smaller percentage on it, I’m cool with that. My guess is that Pumaman was a test case on several levels – good enough quality to be worthwhile? Worth the money? I have no issue with them spending money to test it out, or spending even more to continue testing it.
Ideally, Shout would cough up some money toward this since they’re the producers and distributors of the physical media, and the ones who presumably get the lion’s share of those sales.
EDIT: The more I think about it, I think who’s behind this matters a lot. If MST was in the hands of some private equity abomination (shiver), we’d be right to scrutinize every move. But everyone involved today is doing it as a labor of love – no one’s getting rich off MST3K today; not Joel, not Shout, not anyone. I don’t see any yachts in their future. So I’m happy to give a lot of slack. They’re looking toward MST’s future because they care about it, and that means a hell of a lot.
5 Likes
I emphatically agree with this, especially because it’s likely that the the episodes chosen will not be my favorites.
If they’re improved with other money, fine.
2 Likes
I would be interested to know how much it cost to do it? As that is my only “concern” within a situation I don’t feel is a concern.
My initial thought is the steps required (get a new best quality print. Edit it to match broadcast digitally add the “shadows” over the new print, and splice in the host segments. Simplified I’m sure) shouldn’t cost too much to make (If Shout has the prints and additional rights aren’t a big concern) compared to the cost of a new full production episode.
If the cost was anywhere between 1 new against 1-4 “updates” I’ll take the new one please and thank you. 1 new against 5-9 updates… that’s decent Gizmoplex content. 1 new against 10 or more updates… now we would just be getting spoiled.
I would also look at it from a question of what and how i’m paying for it. As a Kickstarter backer anything that improves the year I paid for above what was promised during the campaign the better. When time to re-up regardless if “renewing” a gizmoplex subscription, or another round of Kick the Starter, as long as the company is open with how my investment is used, I good with it. If a new campaign was funded for 12 episodes, and suddenly oh it’s 10 and updates… that would be a big nope, but I’m happy with the level of communication from Joel, Matt, Lesley and the others and the extra content like the updates, and the vault picks already make me feel i’m getting more than what I payed for.
3 Likes
Shout does not have the prints nor necessarily the original episode tapes. They’re doing AI upscaling from best digital copy available.
2 Likes
The process wasn’t clear to me. Nor did they give any hint as to the expense. But Joel and Matt both made a point to say in the Pumaman stream that it was indeed to be a test case to see if it was worth the trouble. I should have phrased it better in the original post, but that was the intent of the thread.
I do see a case for using KS money for it. We didn’t just fund Season 13. We funded The Gizmoplex, intended to be an all-encompassing home for the show. And part of how they’re trying to build up the viewer base is having all the 90s episodes available free this year (possibly with ads next year). So touching up the old episodes could potentially help get new people into the show who might otherwise find the video quality too low to bother.
But I agree that we don’t have enough information to make the call. We don’t know how expensive or involved the process is, or what the overall current budget looks like, etc. Joel said the enhanced version has “twice as much information,” but, as noted above, it’s not at all clear what that means. Is he talking resolution? File size? Color depth? And how much of that is actual restoration and how much is AI guesswork?
I’m reminded a bit of Ted Turner’s push to colorize old movies and shows for new audiences. (There are a couple of riffs in the 90s episodes about that.) Some found it tacky and wasteful and actively objected to the practice or mocked him for it. I was a child at the time and found the old black and white archaic and difficult to appreciate, so I actually liked the colorized versions. I thought at the time that it was very impressive that they were able to take the black and white images and figure out what the original colors were.
The poll at the top of the thread, while unscientific, has consistently shown overwhelming support for going ahead with the project. It’s not my thing, personally, and I see there are some other worthy objections, but I also see some advantage to it if others like it.
1 Like
Shout Factory and/or their subsidiary Satellite of Love LLC own the classic episodes and all the masters that Best Brains turned over. If Gizmonic Arts, LLC creates an enhanced version of a classic episode, who owns that version? Does Gizmonic Arts still have to pay Shout Factory for the rights to stream an enhanced version? (I assume that Shout Factory is charging Gizmonic Arts et al. at least a nominal fee for streaming/sales rights.) Would an enhanced episode be considered a derivative work and therefore different legally than the original versions? Especially if the enhanced episode is not created from a master in Shout Factory’s possession. The other question that springs to my mind is why hasn’t Shout invested the time and money for “digital surgery” so they can sell better quality DVD sets?
1 Like
I think this is a good way to look at it. The cost of doing this one digital surgery experiment was part of “build the Gizmoplex”. I’m good with some of my backer money going to this trial. If they were going to do this digital surgery to all (or even a lot of) the old episodes, I think that would be something that I would prefer to have a separate funding source, as I assume that would cost a hefty chunk of cash.
1 Like
I think part of this depends on ownership of the various corporate entities. Is Gizmonic Arts LLC another subsidiary of Shout? (a sub-subsidiary? A su-su-sudio-sidiary?)
2 Likes
I don’t know. But if, as others said above, the process is just a matter of using software to clean up the image with its best guess of what a better print should look like, then the primary investment would be the software itself. Sadly, gone are the days when you could just buy software and own that copy in perpetuity. But even if they bought a 1-year license or something, the major hurdle would be passed. After that, I’d guess it’s a matter of processor time, electricity costs, and maybe having someone review the results to make adjustments and touch-ups.
That’s a lot different from what I’d understood from Joel’s explanation. He was talking about how the show used what they could get at the time, which was often a degraded VHS tape-to-tape transfer, which is different from how accessible digital media is now, and I took that to mean that they were replacing the video with a better copy. But apparently I misunderstood.
1 Like
I can’t be the only one that wants to answer 2 & 4, can I?
1 Like
Gizmonic Arts, LLC is, according to the terms of service at the Gizmoplex Virtual Theater, the owner, operator, and controller of The Gizmoplex. It’s one of (at least) eight LLCs registered in Pennsylvania related to MST3K and the making of the new and live shows.
1 Like
Digital surgery? Why I’m all for it! I’ve had it done, and my fingers never looked lovelier.
5 Likes
Although some of that “fake” information comes from the software’s experience (its training) of what should be visible in a video.
2 Likes
“Fake” is an interesting choice of word. New data is being created where none existed before. That new data is indistinguishable from the original data around it – both are just ones and zeroes. Fundamentally, I don’t think there can be such a thing as fake data. It’s not like Scooby and the gang can pull the mask off a byte and reveal that it’s really a potato underneath. (We can organize data into patterns to create fake/misleading information, but that’s another story.)
I get @Treadwell’s intention, though. Traditionally, adding new data to an image just makes it bigger and fuzzier. I was a skeptic for a very long time, until only recently when the technology became good enough to convince me otherwise.
3 Likes
All the TV shows would have someone say “ENHANCE!” at a computer screen and somehow it would take a low-res image and turn it into a crisp, clear image in which you could read the critical clue in what had been a cluster of four gray pixels. And anyone who had ever worked with a real image editor would roll their eyes knowing that you can’t just magically restore data that wasn’t there. One gray pixel is one gray pixel, and expanding the image just makes it a bigger gray square. The computer has no way of knowing what else was originally in that square.
And yet, somehow, apparently, we’ve gotten to the point where you actually can do that?
5 Likes