I did a quick test on a frame from “Points of Departure,” and the answer is “not as easy as I hoped.” I’ve tried this technique before, matching and blending the same film from two different releases (and, apparently, two different digital scans), and I’ve also run into this problem of the film not scanning perfectly flat, and having some amount of distortion. I’d have to do further research to see if there are tools to address this that I have access to,¹It feels like a combination of technologies should be able to solve this; I’ve got two slightly different sources of the same ground-truth. The HD frame shows exactly how 3/4th of the SD frame should look after it’s upscaled, so it should be possible to have a computer program compare the two, adjust the color, size, rotation, and proportion of the SD frame to match the HD one as closely as possible, apply some ML upscaling (ideally using the very shot that’s being upscaled as training data so the computer can extrapolate the lost detail at the sides of the frame from the existing detail in the middle), and blend the two sources. I know all these features exist independently, but I have no idea how to get them all into one tool. There might also be some complications with the DVDs needing to be detelecined and having scanline issues in the raw data. so this is just a quick test and comparison using Photoshop. The vertical framing of the DVD and Blu-Ray don’t match exactly, and the DVD’s colors are a little more contrasty, so I had to make some adjustments to match better.

First off, the test image merging the two sources. I upscaled the DVD frame using an AI tool, color-corrected it as described above, and added some grain, while softening the border between the HD frame and the widescreen one:

It could be worse!

Now, the quality comparisons. Here is a straight split image, comparing the DVD and the Blu-Ray (the DVD is on the left, if the pillarboxing on the right doesn’t give it away). Because I am cruel to old technology, for these images I scaled the DVD screencap up without any interpolation, so it’ll look pixelated rather than just blurry.

Sheridan is noticeably taller on the Blu-Ray. It turns out the widescreen version (at least in this shot) also expanded the frame vertically. This next one zooms in the DVD screencap a bit more, so they line up properly, even though it crops out some of the DVD picture. Check out the fine detail on the HD side, like the EA logo on Sheridan’s belt buckle, or the ribbing on the light panels on the wall.

And, finally, I uncropped the above image, eliminated the split-screen and put a red outline around the Blu-Ray screencap, so you can see how the framing is different. To be clear, I’ve done this with one frame of one episode, I won’t say this exact difference applies to every shot in every episode, some may have more information on the bottom in the widescreen version, some may even be cropped in vertically on the DVD, as an ancient website analyzing the first widescreen releases of B5 proposed.

As a bonus, here are uncompressed PNGs of the source frames at their original resolution:²More or less, the concept of “original resolution” is a little wacky with the anamorphic DVD format.

I recently started rebuilding a new, PBR-based set of models in Lightwave 2020, mostly thanks to the models of Chris Kuhn, Marc Bell, and Alexander Klemm, but I hadn’t really done anything with them yet. This new model was a good chance to, and since it hadn’t been officially converted into Lightwave, I got to dip back into my roots a little when I had to convert anything I wanted to use. Downloading a cool new model that was just mesh and textures, and really digging into it to get it to look right. I spent the weekend building out the lighting rig and doing various minor modifications and tweaks, like breaking out the formation lights and impulse glows so they could be animated, and found picked an angle to run a test render (a good thing, too, I found a tiny sliver of window-box sticking out of the hull). I thought it looked pretty good, so I did a final version adding a basic Sun/Earth/Moon three-point light setup, and that was that.

I tried to think of something more dramatic, and thought up a concept for another image, with the Enterprise-G over the Founder’s Homeworld seen in Deep Space Nine, returning the renegade Changelings who’d infiltrated Starfleet to their own people. Luckily, my prep came in handy, and I already had a Jem’Hadar fighter and a Defiant ready to go for a suitable escort. I spent a bit of time making new decals for the Defiant-A (I know in-canon the second Defiant had the same markings as the first one, but Ron Moore wanted it to be the Defiant-A, I wanted it to be the Defiant-A, so I made it that way). I was pretty far along before I remembered the Defiant was a museum exhibit now, so I just went with it, not having any better idea what ship Deep Space Nine might have assigned to it in the PIC era (or if there’s even still a DS9 at all). Maybe they flew it out as a goodwill historical thing.

Enterprise-G by Marc Bell, Jem’Hadar Fighter by Chris Kuhn, Defiant mesh by Chris Kuhn, textures by Marc Bell, and Lightwave conversion by Matt Christou.

Star backgrounds in both images are NASA’s Deep Star Map (though I should’ve been used the fictionalized version without recognizable constellations for the second render). The Founder homeworld is NASA photo ISS048-E-010018, recolored in Photoshop to match the planet as seen on the show. Both images had compositing and post work done in After Effects.

Edited 2024-10-08

Recently I learned of JPEG XL, the little image format that’s going to solve all of my problems (and by “all” I mean “delivering HDR content via the web”). I’ve been post-processing my images in full HDR for a while, but I never had a good way of distributing them.

As a test, I’ve exported the full version of “Repatriation” in HDR JXL, just to see if it would work, and I’m tucking it away in this old post behind a read-more banner so I can check it on different devices and see what happens.

These are the photos that she chose to use on her website. She’s also posted some outtakes to her Instagram feed.

Perhaps it’s appropriate that my final image would be an ambitious misfire. I’ve occasionally created accidental long-exposure-style images from time to time. I’ve wondered about doing a sci-fi long exposure since I saw one in the 2002 edition of the Ships of the Line calendar (which sadly doesn’t even have a thumbnail online I can link to). My big mistake was having it be a daylight image. The ships orbiting Babylon 5 are so underexposed you can hardly see them even if you know they’re supposed to be there. There were also some technical gambles I took with lens flares and antialiasing that didn’t pay off. I think the concept is worthy of a second attempt based on what I learned, though.

So, one hundred days, with one hundred images to show for it. A lot has happened in the past three-plus months. I went from being unemployed, and wondering whether it was even worth the trouble to continue to try for a career in 3D to having a great job with a great team doing meaningful work that makes the world a better place (and, yes, draws on my experience in 3D and knowledge of visual effects. And I even get to use Lightwave at the office). The support and advice of my internet friends over at Foundation3D led me to explore new-to-me techniques, challenge myself, and experiment with my work in a way I hadn’t in about seven or eight years.

I mentioned back on day zero how I credited the drop-off in my output primarily to a university drawing class that was a prerequisite for my animation major. It was a very holistic course, with not a lot of specific instruction in terms of technique. Very much being thrown into a deep end, which didn’t help considering how, while I always enjoyed drawing as much as any kid, I also always recognized that I was solidly in Ira Glass’s gap between taste and ability, and that I wasn’t moving anywhere with it. In retrospect, I’d blame being cross-dominant, but I’m sure there are plenty of left-handed artists with the same crossed wire who are much better drawers than I am. In any case, I spent two semesters smashing my head against a wall producing pictures that weren’t very good that I wasn’t very interested in. That was around 2008, 2009, and you can certainly see the effect by the number of blog posts after that.

And then, once I started to get my mojo back in my animation courses, I always felt an obligation that everything needed to be a technical exposition, an animation I could put into my reel. Nothing could be small and fun, everything had to be ambitious, something that would get me noticed (or even, God help me, go viral). When I started this project, as much as anything, it was a declaration that I was going to do what I wanted. A reel full of spaceship beauty shots won’t get me anywhere? Well, the hell with that. I like spaceships. And if I was truly going to end up finding a new direction in my life, then I was going to close out my time as a 3D guy enjoying myself and doing what I loved. And, just like every longshot success story and parable about serenity, genuinely letting go of others’ expectations worked. As soon as I stopped trying to be employable, somebody told me I was employable. So, yay. And it’s at a place that doesn’t pretend human talent doesn’t exist, so I can actually put my day-job work on my real, so I’m no longer obligated to let my career anxiety have control over my free time. Double yay.

And I have a hundred images to look at. Triple yay. Some are real dogs, some have promise and will shine after a second pass, and some are so stone-cold gorgeous that when they cycle up on my computer’s desktop, I have to just take a moment to stare and convince myself that, yes, I made that. And I was able to reconnect with the idea that it was okay to do something easy that would only take an afternoon or two. Not everything needs to be a two minute animation or a fully rigged model. Even when I started working again and didn’t have a bunch of time to kill, doing my daily image was always an obligation, but rarely a burden. Not to say I won’t enjoy being able to just flop in bed without thinking about whether I’ve got a render cooking, of course.

The events of the last week have been disturbing. I’ll admit, fun sci-fi fan art doesn’t seem quite as worthy now as it did a couple weeks ago (and I know I’m not the only one feeling that way). I’d initially had some fairly ambitious plans for hobby projects over the coming year. I’m not sure if I’ll end up being able to devote as much of my time and energy to them as I intended to before, but I do expect I will go through with them. While I’m never going to obligate myself to post once a day again, I’ll try to have something weekly.

Finally, I want to thank all of you who have been following along. I do check my traffic stats, so I see that people have been coming through day by day. Like I said, I mostly do this stuff to make myself happy, but it’s nice that other people like it, too.

Almost to the end. A quick image for today, to give me a little extra time to do something special for tomorrow. Hopefully it’ll turn out well.