Some updates (Patreon)
Content
Hey everyone, nothing pressing to report, but I realized nothing was stopping me from just letting you know what's got production hung up for the moment, and hell, maybe someone will have some input on it. I seem to just keep stumbling into bad subject after bad subject, at least in terms of nothing being straightforward
Way back in December I was working on a script about backplane-based PCs when I bogged down for... Honestly, reasons I can hardly even remember at this point. Oh right, it's because it turned out the chassis I decided to pull the trigger on and spend 300 bucks for ended up having a bizarre PICMG board inside - one made on a little used and quickly forgotten revision of the standard that incidentally does not fit 99% of the SBCs out there. SBCs are not easy to come by, and I've bought two on eBay that haven't ended up working. I have all of... one, I think, that works in this board, and I've determined that even if I tried buying others, they wouldn't fit either.
My only real option is to desolder and replace one of the PCI slots on the backplane, which is... not appealing, and may destroy the board, leaving me with nothing. It doesn't help that the whole subject is divided into "dully straightforward" and "a big mess of barely-related gadget-anecdotes" that I had not figured out how to turn into a narrative by the time this problem became apparent. So I had to shelve that until I figure it out.
While licking my wounds on that one, I received the 24 frame PCs I wrote about a couple posts back, and decided to pivot to that topic, which was also not easy. First I had to rebuild one of the machines completely because the motherboard was dead; then I had to do a whole bunch of examination and experimentation with the data on the hard drives; then I had to obtain a monitor that supported the weird mode it outputs.
In the middle of this process, I realize that the guy who built these things has a website and an email address. I get in touch, expecting no reply or "quit bothering me kid" or maybe even "The contents of those drives is the property of such and such studio, cease and desist and destroy them immediately." Instead he replies tersely and just says, yep, you found me.
Not sure what to do with that, I reply as politely as I can, but include a list of questions - carefully chosen to, let's be frank, massage his ego as an engineer. If you do a project and 30 years later some kid who wasn't even born when you started your career starts pestering you with questions, you're going to be a lot more inclined to answer if he acknowledges that what you did is kind of cool and impressive, which it is, so this wasn't a problem.
Couple days go by. No reply. Finally I email him back and basically say, "hey man, I understand if you're busy or just don't feel like answering any of this, no problem at all, just let me know if you can or can't." Couple days later he replies and says he might get back to me, if and when he feels like it. He's not being rude, he just very obviously does not prioritize my needs, which is perfectly reasonable, the guy being retired and me being a rando.
But where's this leave me? In a lurch. This is the kind of video that I would have filled with pure speculation 5 years ago, and paid the price - I'd have gotten called out on in the comments by people who knew better, and probably spread some misinformation. So here, I have the opportunity to do exactly the right thing and I take it - and it leaves me in an interminable holding pattern. Worse, even if the guy replies someday, I don't think I would get very much out of it. The couple answers he gave me already are... Vague, to put it mildly.
But he IS out there, so what do I do? I can't say I couldn't reach him, that's a lie. I can't just ignore the fact that the guy is findable because then I'll put out the video and 150 people will assume I never thought to try getting in touch, and pester him with even more questions. So I absolutely have to acknowledge that I got in touch with him, but somehow brush off the fact that I got no information from that exchange without being insulting.
So I was paralyzed, totally unable to take another step on the script because anything I wrote might end up having to get removed if he ever got back to me. Finally yesterday Daria talks me into throwing my caution to the wind, putting a very precisely worded "Don't worry about it" in the script (which pretty much solved the problem as far as I'm concerned,) and then I put in four or five hours of work, feeling very productive... until I ran into a new and catastrophic problem.
The whole time I've been thinking about these things, it was under the belief that you cannot put a CRT in front of a standard 24 FPS movie camera and get a clean picture unless the CRT is running at a multiple of 24 Hz (in this case, 48hz, though 120 will also work.) I swear I have heard this assertion in various places; obviously it's why these machines were modified the way they are; I am also aware that there are lots of 48hz-modified TVs out there; and I know the Mac Plus in Star Trek IV was modified to do the same thing (shout out to Ron's Computer Videos for covering that topic, it was pretty damn cool and I will be citing that video in mine w/ his permission.) So this is a common practice, yet I still found myself thinking that it couldn't be the sole, or even typical answer.
I started flipping through a bunch of movies - Independence Day, Hackers and Blade to name a few - and looking at just how many TVs and computer monitors are in every scene. They're all over the place, sometimes simply as minor background elements, and I just can't believe every single one of them was modified for 48hz and was being driven by a 48hz computer or VCR. It'd cost an unholy fortune even by hollywood standards - and what about low-budget productions? I've seen really cheaply filmed movies that still had working CRTs. There HAS to be another explanation.
So finally I do what I should have done in the first place and just Google "how to shoot CRT on 24 FPS" and sure enough, there's people on a filmmaking forum talking about how to adjust the Panasonic DVX100 (lol) to do it. That's a camcorder but hey, the advice turns out to be the same - you just set the shutter angle to 144°, which (in film terms) means 1/60th of a second, long enough to cover a full screen refresh on a standard NTSC CRT.
Now I knew this, logically - 1/60th is a subset of 1/24th, so obviously it has to be possible to adjust a standard movie camera to shoot at that speed. But why, then, would you bother hacking together 48hz displays at all?
Obviously at that point I'm kinda back to the drawing board, I have to rewrite the whole script to remove the basic, flawed assumptions I started with. But replace them with what? What is the ultimate answer here? I have some theories - but I don't have anywhere to go to find someone who can put it all together, which is frustrating. For what it's worth, here are those theories:
First, it seems that not all film cameras had adjustable shutters. Older and low-end models appear to have typically been locked at 180°, which is exactly 1/48th of a second. It is thus impossible to film a 59.94hz (or even 50hz) tube on such a machine and get a clean result. As far as I can tell there were a LOT of cameras like this, so 48hz displays are the only solution that would result in a clean picture. Terrific.
But then, why would a big-budget, late-era production like Star Trek IV care? They absolutely had access to adjustable-shutter cameras, even if they only used them for the 20-second sequence containing the Mac.
Okay so, second reason: It sucks to be stuck at 144°. What if you don't like the motion characteristics that gets you, or - for some reason - you just can't get the exposure you need that way? Well, that's something, but it feels weak to me, especially for the exact shot we're talking about, which is a locked-off closeup without a ton of motion and where they could have just lowered the lights to get the exposure they needed.
Then there's the sync issue: sure, you can set your shutter to a 1/60th interval, but each shot begins at the top of a 1/24th cadence, and that's unavoidable. You can't genlock a 59.94hz TV to a 24hz pulse, I don't think, you'd just be asking for too much latitude, the set would be constantly wobbling as it was violently resynced. This is a total guess, since I have no way to test it or anyone to ask, but it seems valid. The problem is that with that level of confidence it's just... really really hard to say anything. Any way I phrase it, it comes off as the XKCD joke about "up to 25% or more"; basically I'd be saying nothing of substance, yet talking anyway. I hate that.
Worse: in a lot of cases it wouldn't matter. Sure, the screen might not technically be synced to the camera, but you're still going to get what looks like a full-screen image. It'll have a split in the middle, which will jump all over the place in every frame, but since the edges of the split touch, it won't be obvious.
Now, to be fair, this sucks for high-motion content - if you're filming a TV with full-motion video on it, like a camera feed, you're going to see tearing artifacts. But for, say, the scene in Star Trek IV that consists of no more than 40 distinct still-frames across 20 seconds, would that really have mattered? It's a brief sequence with heavy comical overtones that - in practice - doesn't come close to looking realistic anyway, and is also being shown to an audience that's overwhelmingly never used a computer, let alone one with a GUI. Would a little bit of tearing really matter?
Well... I mean, yes, in the sense that some directors might look at that and go "no, that's just not the right way to do it, we need to do better," - but I feel like a lot would just shrug at it. Who cares? The audience won't. So I'd guess in practice 48hz displays were only used for really important closeups, or by picky directors and DPs... but is that really true? I don't know and I have no way to find out. I could go track down someone and ask, but I can't survey the whole industry circa the mid 80s or 90s, so what kind of firm answers can I give? None, really. I could even trying reading the SMPTE journals or trying to find marketing ephemera but all it's going to say is, "if you care about this, here is how to solve it." It's not going to tell me who cared and why.
For what it's worth, I did some experiments and confirmed at least some of the ground truths. Using my Blackmagic camera, and a monitor running at 60 Hz, I can get it to look right at 144°, but almost any other setting produces awful flicker or black/bright bars. Some values above 180° are usable-ish, but my understanding is that most film cameras top out at 180° for some reason? Don't really understand that and have no idea how to begin investigating it but there it is.
With the monitor at 48hz however, I can get perfect sync from about 75° to 180° - and at various angles above that, up to almost 360°, which would have been impossible (afaik) on a film camera but is worth mentioning anyway. At any angle other than 180° I get a black bar artifact of one size or another, but I can fix that by adjusting the phase of the camera (in practice: cycling it out of 24fps mode and back repeatedly, lol.)
The ideal solution would be to genlock the PC monitor to the camera, but none of my cameras can output a sync signal - and apparently real movie cameras often had shutter phase adjustments, even back in the day, that would have solved this for most cases. Like, it was fiddly, but sometimes filmmaking is; and it wouldn't have worked in scenes with multiple CRTs in shot at once, but you could solve that by just genlocking all the tubes to some common reference, then phasing the camera to match it.
But what do I do with any of this? I can give speculation, but I can't give answers, and I really don't know how to fix that. And to be frank, this is what my last... I don't know. 15 productions have been like? Total collapse of confidence in any of my conclusions after weeks or months of effort. This just seems to be what my job is nowadays. Nothing's easy, nothings certain - I'm never sure if I actually have the answers or I'm just making crap up.
The stress is absurd. I never know if I'm going to get dogpiled (rightfully) for getting things wrong, I never know if I could have changed my whole understanding of the thing if I had just sent one email or DM or waited another two days. Every time something goes wrong like this, I start thinking about how long it's been since my last video and how much longer it's going to be now. At this point it's been what, two months, and I have no idea when I'll have another video out.
I realize this is all my fault - my whole product, as far as I'm concerned, is covering stuff that nobody else covers. That means the information about it just isn't readily available, or what is available is not clear cut, so every single project is going to be like this forever. I still love the job, I still don't want to do anything else, but man, there is just a certain unpleasant flavor to the feeling of never knowing if you're making the right decisions that gets to you eventually, especially when you keep getting knocked back to square one when it turns out you were wrong.
Anyway. Thanks for sticking around. If anyone has input on the realities of filming CRTs 40 years ago feel free to comment. Hopefully I'll have something soon.