First VFM02 Environ Production Modeling


The VFM02 virtual set has officially passed from pre-production to production, as tonight I’ve made my first piece of complex geometry (ever) from a simple cube to begin replacing the graybox blocking. The previous tutorials really helped. When I hot-keyed the Extrude command, oh boy did that change my world.

To take it any further at this point, now… well… now it’s time to unwrap UV texture coordinates. Yay. Yayyyyyyy.

I tried the ‘Automatic UV Mapping’ command and it created a total mess. Some vertices went off into infinity, apparently. The command might as well have been called ‘LOL JK FU’. But then I deleted all of the UVs and did it again– with a fresh slate, it worked out really nicely. Surprised. Broke everything up by six sides and unwrapped those nicely. And I stumbled on that totally by accident– I could’ve spent hours trying to do Planar Mapping and pulling verts manually. Ugh. Next, I re-arranged the shells and changed their proportions based on what would be most visible and important, tested this with checkers, etc…

Later tonight, I’m going to burn through this tutorial from Digital Tutors, UV Mapping Workflows in Maya, and hope UV Unwrapping and Mapping can finally not feel like such a dreaded process.


I had gone through the sections about UV mapping on for Maya. While valuable for a fresh beginner like me, they ultimately weren’t nearly thorough enough on this. They never got into unwrapping complex geometry. Instead, there’s an example of using planar mapping to throw a flat texture onto the flat side of a flat mesh– a fish! C’mon… They do not adequately prepare you for, say, unwrapping a triple-A quality human character mesh, let alone this simple tech box. There were some better clues in Lynda’s Game Environment Modeling tutorial though. I’ll check out the 2015 Maya Essential Training sometime, even though I’m running 2012.

So I’ll give the UV tutorial a go later tonight. I just finished the Gnomon Game Environment Modeling tutorial, where Nate from Sony Santa Monica talks about how he’s modeling to prepare for texturing and how the modeling process is never really done until you finish texturing and even then it’s not really done… but the tutorial ends right when it should go on to the texturing part. And Gnomon doesn’t offer a part II texturing tutorial, but they do have several on the subject. It just would’ve been nice to have the continuity of process and see it through. I’ve yet to find a single or multi-part grouped tutorial that truly goes from start to finish at a fully professional level.

EDIT: Looks like Nate did do a counterpart actually, but unfortunately it’s all in ZBrush instead of Mudbox: Environment Modeling and Sculpting for Games. Still, would have liked to see the first tutorial followed through to the end, but this helps! Also just signed up for a Gnomon subscription!


Anyway, so I’ve got the UVs sorted and now I might bring it into Mudbox to make sure the polygons and UVs all play nice. If so, then I’ll do a quick texture pass between Mudbox / Photoshop / and experimenting with making a techy elements normals-baked kit in Maya. And just refine from there. Before I do this, I may also go back and refine the model as well. I’ll update this article if I make any further significant progress tonight.

02/26 — 3:30AM UPDATE: Refined the model, deleted unnecessary faces we never see, redid the UVs. Then brought it into Photoshop and sketched over it as a design pass. Will do the photographic texturing build in Photoshop tomorrow, and I think I can probably get away with doing a bump map in Photoshop manually instead of doing normals in Mudbox– but I’ll try both!


02/26 — 20:12 UPDATE:
Current Photoshop texturing and design progress:


02/28 — 21:15 UPDATE:
Had some time today to work on texture. Experimenting with bump maps– first tried simply turning off all the graphics from the base textures and creating a quick desaturated and levels-tweaked bump map. Not too sure about it really, so this probably isn’t going to be as easy. Might have to do a whole new bump map texture from scratch for it to really make sense. Want to also bring this into Mudbox next and see what can be done with normal mapping there. Also, this was rendered in Maya with Mentalray and getting VRay is on my list… This material is a standard Maya Blinn so I’ll look into experimenting with other materials later as well.


03/01 — 21:15 UPDATE:
Manual bump mapping. Much better, more like modeling by grayscale graphic design. Not as pronounced as I would’ve liked. Maybe I can get some more texture into it on a separate layer next. I tried bringing this into Mudbox, but upon upscaling its polys so I could sculpt it and bake a normal map, the UVs didn’t perfectly upscale as I had hoped and distorted the base texture. I have a few ideas for some things I can try to fix this next.


03/02 — 21:15 UPDATE:
Successfully upscaled the geo of the model in Mudbox without distorting UVs and did a bit of sculpting on the mesh. Then figured out how to export this to a normal map for the low-poly mesh target, and applied this in Maya. It appears I had to choose between using the Bump Map and the Normal Map. I couldn’t apply both? There’s probably a method to do that in the Hypershade.


When I tested this in Maya, the results didn’t quite transfer as effectively as they appeared in the Mudbox sculpt. So I’ll have to experiment with this more. Currently, I’m fairly happy with using the Bump Map instead, as I was able to get very precise cuts (plating seams and vents, etc…) and extrusions (bolts, etc…) in a very straightforward manner directly over the 2D texture map in my PSD. That method was also very fast. I’d probably entirely rely on Mudbox for more organic sculpting needs instead of hard surface stuff. The bump map just didn’t deform negative values as much as I’d like, and increasing the bump value in Maya introduced distortions.

Lots to keep experimenting with as I learn this stuff. For environment art, I’m starting to see how you can go a little more high-poly than I was expecting (especially these days) and not have to do all of the detail in bump maps or normal maps, especially if the goal is to make a movie and not a game with these assets. I’d like to figure out the game method, though, as it’s most optimal and seems like a best practice for getting an asset done efficiently and effectively in both arenas. Plus, I’m looking into actually using a game engine like Unreal Engine 4 to render scenes in realtime to save enormous amounts of time and bypass the need to buy and build a renderfarm– will have to test if this will hold up in 4K.

I’ll keep digging and continue messing around with it. But as far as a pipeline, I’ve never gone this far, so that’s cool.


I think at this point I can probably move on and build out the rest of the set with low poly objects and bump maps. Especially for a movie, where I know the camera never gets in too close on any of this and details are lost, even at 4K. It’s important that they’re there, but I think they don’t have to be as in-depth as a model you could walk right up to if this were a game.

Alright– moving on!
IMDb / LinkedIn

← Blog Home

WIPs: VFM02 Environ: Concept 2


Had about two hours available tonight to quickly photobash together a totally different concept for VFM02’s test environment. This is a great test project to get on with learning how to make production environment art in Maya / Mudbox or ZBrush (which I have but am hoping I can stick with Mudbox as the UI seems a lot more logical to me right now), so I wanted to move fast with the concept stage at this point.

As mentioned in the last post, this Virtual Filmmaking 02 test scene is set in an aerostation in the upper atmosphere of Earth instead of the typical space station. So to try something totally different than the last environ concept (which was very dark), I thought I’d open the space up with windows and go bright. Strong colors. Blues, oranges, yellows. I wanted to see the atmosphere and the exterior of the station.

I ended up selecting, morphing, and combining elements from about a dozen or so different references (also giving me ideas for who to reach out to for future collaborations). I’ll let this settle and come back to it– If I’m still happy with it and decide to go with it, then it will be onto the more intricate process of designing out this stuff as the direct original assets.

I like this so much better. It reminds me a lot of how refreshing Mirror’s Edge’s art direction was over the usual drab cyberpunk fair– here bright, airy, and also nicely striking.

Alright, timeboxed and out. Later!
IMDb / LinkedIn

← Blog Home

WIPs: VFM02 Environ: Concept 1

For this snowy Saturday, I’m gonna do a bit of VFM02 Environ concept art.

To kick this off, here’s a recap of the conceptual WIPs (works-in-progress) I’ve done so far. I’m not really satisfied with it yet, so today I’m going to start over and try some new things. But here’s what I had:

After Effects build of First Styleframe from 2015.01.31:


Here it was rebuilt in After Effects from the photobashed styleframe concept below. This is still just a styleframe (though from a full video version at this point), as the actual shot would have everything modeled out as original assets.

For years, I have been studying my favorite cinematographers the Cronenweths (father / son who did: Blade Runner / Fight Club, etc…). I have this inverse method, where I love shadow and prefer to ‘work with shadow’ more than light (even though technically they’re two halves of the same whole), which is very apt for a future noir look. But it clearly needs refined here. For this scene, the characters are in what I call an ‘aerostation’ floating amongst the clouds, instead of the usual space-station, and it’s sunset with a possible storm brewing and maybe some light rain going outside (the reverse shot on Hale shows this through a window).

So, I was also referencing Blade Runner frames, particularly when Rachael and Deckard meet in Tyrell’s office to get that color noir sunset look– but the cinematography on film in the 80’s is a lot less sharp than, say, that in Elysium, where a lot of detail is packed into every frame. I think I’m going to pop in the Elysium Blu-ray and do some reference capture next. I had also reframed this shot after matching the Maya virtual camera’s film back to a standard 35mm film camera.

At this point, I want to start over on it, and play around fresh with concepting the scene. That’s what I’ll start on today.

First styleframe concept from 2015.01.31:


This is the first photobashed styleframe concept done in Photoshop. After letting this rest for awhile, I can now really see how this first pass was pretty fast and loose– it’s kind of a mess, isn’t it. Less future noir and just kind of muddy. You can compare it with the After Effects build above and see how I really tried to clean it up though.

Overall, I don’t think the giant brains in vats really work here, especially out of focus– you can’t really even tell what they are. I really love that idea and find I keep coming back to it, but it’s not really working here (or I haven’t made it work).

Base Plate: Graybox Composite from 2015.01.27:


The base plate I started with to concept over.

Any feedback is appreciated. Hit me up at @XanderDavisLive.
IMDb / LinkedIn

← Blog Home

XD&A Greenscreen Stage and VFM02



In December, after the Cinecomic Experiment unexpectedly proved virtual filmmaking was viable for me, I had this spare unfinished room in my house built-up enough to put a greenscreen stage in it, featuring the new Panasonic GH4 4K DLSM camera, a professional light-absorbing greenscreen backdrop, photography lights, and a 40″ HDTV monitor. At first, I got a small 25″ inch monitor (in the above pic), but we actually missed that a take was out of focus during a wardrobe test, so I upgraded that (below) to be sure.

The implications of this were profound for me. I wanted to see if I could shoot a professional-grade short film virtually from within my studio space on a micro-budget, or even no budget at all, while being unlimited in where the film could take place. Environments could potentially be done all virtually, with only actors captured in principle photography. Essentially, what was only possible through ILM pushing the envelope with millions of dollars in 1997 (despite that film’s script being not so hot) was now within reach in my home with prosumer gear and software (a big thank-you to George Lucas and ILM for figuring out how to make this more affordable and accessible for future filmmakers now almost two-decades later). My mind was reeling with the possibilities.

So, the December Cinecomic Experiment culminating in the test shot of me on the bridge of the Prometheus was retroactively deemed Virtual Filmmaking Test 01 (VFM01). From there, a new two-page script and storyboards were made, and in mid-January (after a texture-hunt vacation around Chicago), principle photography on Virtual Filmmaking Test 02 (VFM02) began. This would eventually be known as simply ‘Prototype 2’.


For VFM02, I essentially used a scene from Prometheus (where David poisons Holloway) as a model and an emulation target, directorially. The idea being if I could re-create virtually an actual dramatic scene by Ridley Scott (in a new context with a new original environment and characters), this would be a proof-of-concept that a virtually-composed film production could be doable for me. I was also tracking my time put into this to begin developing metrics for how long it would take to create a minute and a half dramatic scene with about twelve angles over many cuts. That could then be extrapolated to estimate how long it might take to produce a short film or even a feature through these virtual filmmaking techniques.


For this I would play two characters, an older scientist and a ‘fresher’ experimental clone. I knew I’d need to really differentiate my looks for each to create such a contrast that you might forget the two characters are played by the same person when watching. We would need makeup, another new territory for me. But I wanted to learn everything involved in this process as much as possible. After some research on old age makeup and prosthetics, the wonderful Liz Irish was put to work as my makeup artist for me and both characters. Cortni Fleagle was the hair stylist– the scientist unkempt and the clone tightly-trimmed.


Between the various wardrobe and makeup tests, I first shot all of Hale’s angles, then got my hair cut and could shave (finally). Then I did the Clone’s shots. Both characters over twelve angles or less each took a five hour shooting session– I think that was largely due to the fact that I was ‘acting in the blind’, since I was both the actor and director. After I’d do a ten-minute run through take, I’d stop, sit at the monitor, watch playback, make mental notes to direct myself, and do it again. Plus, this was my first time in front of a camera acting in a long while. Logistically, the shoot was over a span of two weeks in spare time, but when it’s all cut together, it’s magically a single moment. I love stuff like that, the filmmaking magic that creates impossible results in a single frame.

I went the Fincher-50 route and did about fifty takes for each individual line. I believe in this a lot, especially with digital not costing you anything to do that. Since we’ve gone to all the trouble to get to these moments of shooting, might as well be absolutely sure we ‘get it’ and try many different things. This becomes a joy then in editing, because you can custom-tailor a seamless performance between each cut, line-by-line at times even, to get the perfect total performance. This is even more critical when your actors are acting to focus targets on a stand– you need the flexibility to gel the performances together for cohesion. With that many takes, it’s almost impossible to have a bad cut– that is, if your actors can give you those choices.


I actually don’t really want to be in my films, but it’s sort of hard to not utilize myself– I’m here, I’m always available, I’d like to think I look alright, I have acting experience, I can exhaust myself with fifty takes per angle no problem, and I’m free for the production. I’d much rather be behind the camera focusing purely on directing, but at least until I can prove out my filmmaking abilities to attract solid actors for these productions, I’m gonna have to be the lab monkey. I guess it’s good anyway, because it will force me to get warmed up with acting again (it’s been awhile), study the craft, and know what it’s like to be on that side of the camera– all good experience to make me a better director and even screenwriter. For example, I bought all of the Stanislavski books, and am learning that side of it. It’s actually my goal to know everyone’s job on set better than they do if I’m going to direct; I believe that’s required. So begins that journey…

After I did the storyboards and before I did the principle photography, I constructed the virtual set at a ‘graybox’ fidelity. This is common practice in game development as well. Once all the acting was shot, with every take cut up and selects made, I then was able to finally comp them into After Effects together. Once all of that was done per angle, I was able to export out a ‘graybox cut’, so you could -finally- watch the whole thing cohesively in one go. It took about three or four weeks to get there in spare time.


That’s basically where I am now. I’m building up the environment art more fully, first in concept art. This is definitely something I would normally do before shooting anything, but I was satisfied with having just the graybox environment and a solid idea for what it will ultimately look like to begin shooting for the sake of a test. I’m going through training right now at (Get 10 days of free unlimited access to on Maya, Mudbox, and After Effects courses (two playlists totaling in 250 courses of stuff to get through to really up my game). So it’ll be awhile before these shots are completed.

But the idea is once I get VFM02 done and under my belt, the next go at this will be way faster. The skills will be there, I will have done this before, it will even hopefully get mundane. Then it’ll all be less about how to do it, and more about what to do. And what to do better. I am just starting out on this stuff. But for my first go at 4K digital filmmaking, a virtual cinematic pipeline, and learning film VFX, it’s very encouraging so far.

And then, suddenly, VFM02 got a bit derailed, because I got the camera ‘out of the lab’ to do straight-up practical cinematography tests. And the results of what became ‘Prototype 3’ would again surprise me. Can’t wait to share that with you too…

Until then, back to client work! (I don’t sleep much!)
IMDb / LinkedIn

← Blog Home

December Cinecomics Experiment

Cinecomic Test - Alien

In December last year, I was so inspired after completing (on Hard) Alien: Isolation (my personal 10/10 Game of the Year), that I wanted to do some kind of project like that. Between client work, I thought perhaps I could try once again to do a comic reformatted into a more cinematic style, a ‘cinecomic’ as I call it.

As I was developing the project’s original IP, I thought I’d also do a styleframe test using Alien as a general framework to play around with. The comic wasn’t intended to ever be Alien specifically; it just provided a shorthand for me to focus on building up the image and experiment with its look. This could show myself what’s possible, which could contribute to how I would shape the IP and the comic’s eventual treatment and screenplay. Sort of reverse-engineer how to approach the project.

Well, for this test, I ended up pushing it all the way with photography and CG– and it only took me a full day. Doing this hand-drawn would’ve probably taken so much longer for just one frame. In previous cinecomic projects of mine, like Radical Sleep, I like to rotoscope draw over actor poses and CG block-in environments anyway to give the comic a more cinematography style look, but this cut out that entire step. So, I realized I might actually be able to pull off a cinecomic photographically.

As this shot started from nothing (no plan really at all) into apparently a scene from the Alien universe, it evolved iteratively along the way.


Yes that is me three times standing-in for three characters, with facial replacement of Sigourney Weaver from 1979 Alien! Usually for illustrative cinecomics, I’ll draw over myself as each individual character with their own unique features, but in the photocomic route, I realized I could stand-in and compose the entire thing and then do a reshoot, involving other people in the end to not bother anyone before.


I am just now getting the hang of Maya to the point that I can use it pretty freely, but I have a long way to go to get it at the level of quality I want. Still, this was a happy surprise to myself that I could even do this. Since then, I’ve been doubling-down on training over Maya and Mudbox. This virtual set was also originally intended to just remain ‘graybox’, as I was planning on drawing over the reference shots from a single environment for consistency. After realizing I wouldn’t need to draw at all, using Maya to construct fully-realized virtual sets would later become my new standard.


I then would create textures and comp them directly into the shot (including a shout-out to Seegson!) :


Throughout my progress, I’d switch gears and draw over the frame to do notes to myself. Some of this stuff I was never able to fix (glasses are highly problematic in both still tests and greenscreen tests), but a lot of that had to do with this being a first test and not really planning anything, as I would for an actual shoot.

Cinecomic Test Notes

For a still-image based cinecomic, I like to shoot reference video and pull the perfect frame out to roto with hand-drawn art (the original intent of this test), instead of try to awkwardly pose for the perfect still photograph. You can get better motion and even performance out of this. However, even at 2K (1080p), my Canon Rebel T3i DLSR (which we shot Think Tank! on four years ago) was very fuzzy and sub-par for this– I wanted to compose my shots in at least 4K but more like 8K ‘Scope since you can really stop and stare at a comic panel for hours.

But the most obvious limiter from this test here was definitely my camera resolution. I actually ended up blurring everything else down to match it, particularly the environment art. The hands of the corpse especially look very Photoshoppy from the poor video quality. You win some, you lose some. As long as you learn something…

During this time, I also picked up some awesome books about Alien / Aliens and printed out some scripts. I was just all about it, immersing myself in this stuff as I developed the project’s original IP. It was really interesting reading the behind-the-scenes stuff from 1979 Alien from Alien: The Archive. I’d also highly recommend The Art of Alien: Isolation. The behind the scenes of the ultimate Alien Series Blu-Ray and Prometheus are fantastic as well.


After this, I did a separate greenscreen HD video live-action test (a little GIF clip below) that worked out even more effectively and efficiently to my total surprise. I used some previous motion graphics work to test out holograms as well. I wanted to see if the greenscreen could match an actual photo-real environment, so I took a capture off the Prometheus Blu-ray, Photoshopped out Fassbender on the bridge as a clean plate, and added myself in there as three different test characters for depth placement, all in a few hours.

(preview GIF)

From that test, I realized again… I might as well push even further from ‘cinematic comic’ instead all the way into filmmaking and VFX! Ha! I still like the idea of doing my own hand-drawn anime-style cinecomic one day, but this proved so much more efficient a process for me– the effort is more sustainable actually in filmmaking for even one guy. Which is totally ironic because I always thought this would be immensely harder. One of those times where things just all come together and make new sense. This is a perfect example of why I test first to reverse-engineer an approach. This totally changed everything for me. But this first Alien cinecomic test frame was how I unexpectedly got there.

One way to fix the video quality issue, which is what I did next, was to get the Panasonic DMC-GH4 4K DLSM, a camera designed specifically with indie filmmakers in mind. Not that long ago, a camera like this at 4K was only viable for indies through the RED One. Now I can create full fidelity video compositions in After Effects in full Cinema 4K resolution (4096 wide).

While I had always wanted to be a filmmaker my whole life, this cinecomic test was accidentally my first step into actually doing it (more on that later…) in a very roundabout way. A large part of these final results here is how I made myself move quickly through this– ultimately this pipeline has to be sustainable for me on my own right now, to be viable at all. But this can be better. I know there’s a gap I have to close yet; the key is to keep going. After these tests and getting the new 4K camera, even bigger tests were about to begin…
IMDb / LinkedIn

← Blog Home