Narrow band amber LED would probably not be terrible for this application.
Or even magenta ones with different filters.
Re-creating Disney's sodium vapor process [video]
91–100 of 114 posts
Re: Re-creating Disney's sodium vapor process [video]
#92Very very cool, mindblowing actually, seems legitimately useful. There have been a few art installations over the years that use these lights: https://www.youtube.com/watch?v=hd077pa-5CI For a second I thought that the video was from this channel, which also does experimental cinematography stuff: https://www.youtube.com/watch?v=-4uzyhbDFas
Yes, extremely impressive. "Mary Poppins" rightfully received an Oscar for best visual effects. I am continuously amazed at the level of innovation done at the Disney studios back when Walt Disney was still alive. Rewatching Snow White with my daughter a few months ago, I was flabbergasted by the animation quality, even when compared to animation movies made decades later. Just look at these water scenes: https://you…
Around 1996 I worked at Walt Disney Feature Animation (Hercules, Tarzan, Mulan, Dinosaur era). I was talked with implementing the Disney multi plane 2D scanning camera in a digital environment, Maya, a 3D modeling and animation package. I was given full specs and it was straightforward. One thing was odd, the 2D coordinate system used 5000,5000 as the origin, I forget what units. No one seemed to be able to explain why. A while later the original mechanical multi plane scanning camera was put on display. Looking at it, it made sense. It used a 4 digit mechanical counter for the position of each axis with a range of [0,9999] so 5000 is approximately the center. Warner Brothers Animation used software called Animo and it had the same origin. I ended up working with an Animo developer at my next job and asked him about it. He said Warner Brothers told them to implement it that way with 5000,5000 as the origin because that’s how they do it at Disney and Disney must have a good reason.
Re: Re-creating Disney's sodium vapor process [video]
#93Very interesting stuff. I'm curious if this means anything in terms of good vs perfect technology or what tech wins out in practice. This process seems just plain better than green-screen, so why did it lose out? Is is just that green screen is good enough? Was it too hard to do this sodium vapor process with random cameras back in the day? Are sodium vapor lights bad enough in some way that practically it lost out?
Some of why it "lost out" just seems that it was a proprietary technology carefully guarded with a rather low "bus factor". Blue/Green screens and chroma-keying have been "common knowledge" for a long time, with a lot of implementations over decades, but the sodium vapor process was a more closely held secret because it was a competitive advantage. Fewer people knew how to do it. Fewer people had access to do it. Eventually those people passed on or retired and the industry "forgot" how to do it entirely. It may have been "lost" simply because not enough people were allowed to know how it worked or have access to recreating it.
I think there are a lot of analogies here to find related to the software industry, such as the relationships between closed source and open source software.
Re: Re-creating Disney's sodium vapor process [video]
#94I wondered previously how feasible would it be to manufacture digital camera sensors with modified color filter mosaic, so instead of usual Bayer RGBG you would have RGB+589nm mosaic.
Re: Re-creating Disney's sodium vapor process [video]
#95Earlier quoted context omitted.
Yes, extremely impressive. "Mary Poppins" rightfully received an Oscar for best visual effects. I am continuously amazed at the level of innovation done at the Disney studios back when Walt Disney was still alive. Rewatching Snow White with my daughter a few months ago, I was flabbergasted by the animation quality, even when compared to animation movies made decades later. Just look at these water scenes: https://you…
Around 1996 I worked at Walt Disney Feature Animation (Hercules, Tarzan, Mulan, Dinosaur era). I was talked with implementing the Disney multi plane 2D scanning camera in a digital environment, Maya, a 3D modeling and animation package. I was given full specs and it was straightforward. One thing was odd, the 2D coordinate system used 5000,5000 as the origin, I forget what units. No one seemed to be able to explain w…
Hahaha that’s awesome. Witnessing the formation of a mini-cargo cult, so-to-speak.
Re: Re-creating Disney's sodium vapor process [video]
#96Earlier quoted context omitted.
Incentive is one thing, budget is another. Keep in mind stuff gets lost all the time, even really important things like nuclear weapon components like https://en.wikipedia.org/wiki/Fogbank
Fair. But looking at that story (very interesting btw, thanks), they did recreate it. So incentive does mean a lot when $$ is also available, which I'd suspect it is in movies.
I think the US nuclear program has a much bigger budget than Disney, and if they need something, they just get it.
The movie industry had cheaper options, so they went with those.
Re: Re-creating Disney's sodium vapor process [video]
#97The original beamsplitter cube isn't some exotic custom device that would cost tens of thousands to make. It's just a dichroic beamsplitter, which is a standard optical element available from online sellers. Variants of these are commonly sued in scientific experiments, optical devices, etc... Here's one for under $300: https://www.meetoptics.com/beamsplitters/plate/longpass-dich... Just search for: "589 nm beamsplit…
Presumably, the exotic part of the beamsplitter is that the dichroic filter must be a bandpass filter with a very narrow passband around 596nm. Yours is a high-pass filter, which wouldn't work for this application.
I mean, a bit more searching did seem to turn up a few notch beamsplitters near the required frequency (not quite the same frequency though)
I am wondering if part of the reason that finding these at the frequency of sodium vapour lamps is partly due to those lights no longer being common. Or perhaps that those just aren't frequencies common to laser optics, which seems like the biggest market for these kinds of devices anyway.
Though, I see no reason why you couldn't use a laser for this process anyway, with matching notch filter.
Re: Re-creating Disney's sodium vapor process [video]
#98I’ve been doing something similar using “cheap” Chinese band notch filters. Had no idea about this Disney process. Also using this for AI training purposes. It’s amazing what you can buy on Alibaba, fully custom frequencies.
> Also using this for AI training purposes. Cool! Training for what? You can't just drop a juicy comment like this and then leave us hanging ;-)
It's explained in the video. Keying algorithms are getting better thanks to AI image techniques. However the training data they use is based on green screen technique so has the same issues at times. Using the sodium vapor method to film training data that looks correct to start with would improve these AI algorithms.
Re: Re-creating Disney's sodium vapor process [video]
#99Hmm, why don't people do this with near infrared? It would allow using full spectrum lighting, and would be easier to selectively filter. I'm guessing the answer is "they do, it's just uncommon and you haven't heard about it"
Maybe... it would get too hot? EDIT: actually, since it's outside of the visible spectrum, typical lens design correcting longitudinal chromatic aberration and doing the spherical to rectlinear projection simply might be off, resulting in soft edged and general artifacts unwanted in a clean mask
You're right, there would be lens distortion. But it should be correctable. You'd have to characterize the distortion for the lens you're using and then unwarp the infrared image.
Re: Re-creating Disney's sodium vapor process [video]
#100The key element here that wasn't explicitly mentioned in the video is that this would likely requre genlock to sync both cameras facing the beam splitter for this to work decently well on any moving subject.
I think this is the reason they're using the Z Cam (presumably E2) cameras visible at 2:48, 3:31, 4:05, & 11:05 because that's one of the most inexpensive genlockable cameras available. This would rule out most mirrorless and prosumer video cameras unless the subject isn't moving fast or a half-frame misalignment is acceptable. Heck, a non-genlocked version of this might be preferable to weird motion blur on existing chroma keyed backgrounds. Very exciting tech