[Mark Suszko]

Oh how I used to dream of some day working for Charlex, back then! Are they still in business?

[Mark Suszko]

My take on this is, it’s a multi-step process, but doesn’t necessarily require roto. A lot of time color grading, but not really “roto” in the Ralph Bakshi sense.

If the talent wears chroma blue and chroma green, that can help, but what I see here are two main effects; the final one is a posterization or solarization effect with the bit depth turned way down. But before I would apply that, the raw video needs to be simplified, that’s where the keying comes in. It reduces greatly the need for roto if you have a keyable background, and clothing that is a different color than the background yet also keyable. And some of that, you can change around by first doing some passes thru color re-timing, to reduce the overall number of color shades and enhance the contrast. Then you posterize the final comp. Tri-tone effect might also help. Bret FX, Motion VFX, and a few others have plug-ins to get there faster, but you certainly could make these FX from “scratch” as well.

[Mark Suszko]

Maybe dust on the hot bulb – switch to an LED or flow. Otherwise, it’s the rotator motor starting to go. Amazon sells some great little LED based DJ lights, very compact and affordable, and they can do the same smooth “color organ” changes as the wheel.

The light I’m talking about is bright enough that I could in theory take it on locations and use it to put up static or moving kookaloris patterns on walls behind an interview.

[Mark Suszko]

I can hear Vince Guaraldi playing now.

[Mark Suszko]

I use GIMP at home now, and it does the job, but I am strongly considering Affinity instead, as the GIMP interface and workflow, while competent, has a few idiosyncrasies that I personally find hold me back.

[Mark Suszko]

Does it happen if you swap the order of layers and mask the other side of the twin shot?

[Mark Suszko]

Spill suppression uses a virtual version of the lighting technique we use in real life, that is, to kill spill off a blue screen, you back-light with what they call a “bastard yellow” or amber light, because yellow is across the color wheel from blue. When keying off green screen, the color on the opposite side of the color wheel is magenta. This is why your tint looks “pink”: it’s a weak magenta tint being used in the spill suppression software.

Don’t even start messing with your keyer, until the raw footage is properly adjusted on your scopes for exposure as well as color saturation. A lot of beginners miss this step and wonder why the shot is such a struggle. Good fundamentals here make the keying go much easier, later in the chain. That lets you generate the key, and then you can go on and color-correct the composited output all you want.

When I’m serious about a greenscreen shot, I bring my MacBeth chart or at least a physical color bar chart to the set and shoot samples with it in the frame, in the actual lighting, then in post I color-correct the raw footage until the chart looks right on the scopes, only then do I attempt keying and and after that, keying is usually instant success, no need to make adjustments. Camera-generated bars are not going to be any help with this; you need to have the chart imaged thru the actual lens.

The pink cast is a sign the spill compensation is cranked up too high; try adjusting the other spill-related parameters, or re-sample the green at a spot closer to the hair, to try and fine-tune the range that gets cut out. I often have to get into the manual control interface and mess with the end points on the controls when this happens to me.

You may want to play with doing a separate luma key and adding that as a traveling matte.

Another thing folks forget is that you can invoke multiple instances of a keyer, and mask it to work in just one area.

The Drop Shadow tool defaults to black… but, most don’t realize you CAN change it to another color, like chroma green, or anti-green magenta, and even use it as a kind of “wrap” light generator along a keyed edge, with variable opacity and softness. Might want to play with that.

Another way to save a bad key is to deliberately mess with the color balance and saturation, exaggerating it to the point the shot is useless visually, but has enough contrasts for the keyer to make a clean matte. Then build that matte with black/white and port that matte to a separate layer as a “hold-out” matte, restore the source footage to something human-looking, composite everything.

Some time back, I tamed a troublesome green screen shot by first keying it over a solid, saturated blue, essentially making it a blue-screen shot, and losing all the garbage mattes.

I kinda miss using blue screens – I think when the key is iffy on blue, it’s a little less obvious.

I predict we’re not far off from the day video cameras using light-field technology record the z-depth of every item in a shot, and on that happy day, we can generate clean “chroma” keys just by selecting where in the depth of field range an image segment is supposed to be visible or not. Call it “Z-keying”. And I want the royalties on that idea/name. 🙂

[Mark Suszko]

Whiteboard animation, unless it has some new gimmick, has gone out of fashion. I wouldn’t pin a business model on it.

[Mark Suszko]

Don’t see a link to the video, so I can’t say regarding cost.

But I can say, I’m not a fan of canned generic animations in general. I believe the most effective animations are custom-built to the specific needs of each client – “bespoke”, if you will. Not to say you can’t save a buck by using some pre-generated elements, templates, and assets, as long as those too are customized.

A boilerplate or template animation risks not being specific enough to a client to really be effective, or something about it may not be the best fit with their own aesthetic, brand, or message.

If you only offer something that’s mostly boilerplate, two things happen: if it’s cheap enough, too many people may end up using it right away, and it becomes something meme-worthy, ripe for parody, and for criticism that your client is unimaginative or imitative. Does your client want any of that potential distraction?

Second, if you limit yourself to making templated media, everybody burns out on your ‘”look”, and then nobody buys it. The next thing you make will then have to be radically different in every way, or you’ll sort of get “type cast” or pigeon-holed as the source for one particular thing. Why limit yourself?

[Mark Suszko]

Well, it’s not impossible to make something like this in AfterEffects or Motion, using 2-D pieces made in a program like Photoshop or whatever, manipulating them in 3-d space to build up a projected image. Projecting it would probably use something like Stumpfl or Wings or whatever the hot current multiscreen display software is. That gets around the task of modeling it in a 3-d application first.

The hard part from my perspective would be getting the composition accurate in proportions for matching up the projection with any physical real world backdrop. To do that, I think you need very accurate scans of the projection site, from which to get the measurements for the modeling.

If I was just starting out and thinking of working in this niche, besides apprenticing or interning with someone already doing it, I’d practice by working on much smaller projects, things that would require only one or two projectors, and a target object or surface that was small and manageable. A car or van, maybe, and what’s good about that choice is, you can access pre-made 3-d models of many vehicles already for free or low cost, so a lot of that work is done and you can concentrate on the more fun parts…. Or perhaps work with a white foam core model of a stage play or ballet scene or an architectural model that was no bigger than 4×8 feet. Requires less space for the installation too, and makes moving it easier.

Once you’ve mastered working on this smaller scale, scaling up to things like the sample you showed would be relatively simple. Anyway, that’s what I think I’d do.