Mixamo

I’d heard of it… but not really paid much attention until a couple weeks ago when I stumbled upon Fuse in the Creative Cloud app. A few hours later and I made a quick animation using one of their stock characters with a mocap clip from their library. Being inspired by Beeple, who I also only discovered a few months ago, I ended up making a couple more anims just to see how quick I could turn them around. They’re only a couple hours of work each but pretty fun to do! Now to try Mixamo’s autorigger with my own models… (!)

Inktober!

After years of watching this October trend grow, 2019 felt like the right time to give it a go! I decided to stick with the ‘official’ prompt list which forced me to venture into new, therefore challenging, areas of subject matter. As a digital artist, especially 3D, it is easy to spend less and less time using traditional media. Up until I discovered computer graphics software when I was @ 20, however, I drew all the time. I then went to art college (Art Center) to study Illustration which resulted in two intense full-time years of ‘foundation’ studies (anatomy, perspective, etc). But… I then quickly discovered 3D. And in the 20+ years since, the amount of time I’ve spent drawing is embarrassingly little. A handful of hours/year if that… because ZBrush. Maya. Etc. So… Inktober was a scary idea. Do I even really remember how to draw anything other than a skull? (I drew a lot of those in High School). The first few I was only half committed and spent little time… But… I had fun. I started to look forward to it. A week in and I was blocking out all of my evenings (most), to sit and do Inktober ‘properly’. Meaning take some time to think about the word… do some little thumbnails.. then pick one and blow it up. After a month, I very much felt like I had finished a journey and accomplished something with profound personal meaning. And that my love of drawing was rekindled. I’m rusty… and it is easy to think how much easier it would be to do this or that in 3D…. but 3D isn’t real. Pen on paper you can put on your wall and say, definitively, that you made it. And there is something satisfying about that which I’ve missed. Hopefully this renewed ambition stays with me. But work and 3D always beckon… Anyhow, here’s a quick video I made of the drawings… All 4x6” with pen and marker. Hopefully I’ll go bigger next year!

Decided to give Inktober a try for the first time... as I hadn't drawn for way too long. It was tough, specifically once I realized that this wasn't going to be a 20 minute/day exercise as I did with the first few. I began taking the time to think about the word prompt, do thumbnails, then pencil, ink and marker. Had varying amounts of time per day... some I surprised myself with and like a lot... some I really don't. But in the end - with 31 drawings in 31 days and probably @ 100 hours of work, I can say that my fear of drawing in ink is diminished. Bummed that the last one is meh... but Halloween prep beckoned... Hats off to the many others who inspired me with their awesome work this month. I didn't want to post as I went as that would add to the stress, but it is nice to do it now and check something off the ol' bucket list. Music is Mogwai. :)

Loc-Nar

After watching the awesome 1981 film Heavy Metal again and introducing it to my 15yr old son (who thankfully thought it was as cool as I do!), I was inspired to represent the Loc-Nar in a new story. Maya, Redshift, several World Machine terrains, a few greebles from KitBash3D, some textures from Megascans and then I did the unusual thing (for me) of some hardsurface modeling. I blocked out a ship and some helmets/gear in ZBrush then modeled them in Maya… I’m very far from a good modeler of such things, but was fun to use the new(ish) Modeling/UV Toolkits which have really simplified things! I also incorporated a couple of the new VDBs that we’re putting together for a new Gnomon Workshop tutorial/library on using Houdini to create smoke/fog/etc. Coming soon… :) Lastly I experimented a bunch with different tools/plugins for grading… Aurora, Topaz, Magic Bullet and also tried out Optical Flares from Video Copilot. All cool in their own ways… In the end it is mostly Photoshop, lots of AOV (Redshift render pass) tweaking with a mix of Camera RAW and Aurora. Final image is 8k, but unfortunately I can’t share that here… :/ Don’t touch the Loc-Nar… bad things happen.

Redshift and VDBs

Decided to try out rendering VDBs (volumetric containers) with Redshift as I found some online that I could download and experiment with. Eventually I need to learn Houdini to make my own smoke/fog elements as it seems far easier to do it there than with Maya Fluids… but for now, these are cool to have. Crazy how fast they render with Redshift as not that long ago rendering such things required a lot of patience. Buildings are from Kitbash3D who kindly asked if I’d like to play with their library for an upcoming demo. :) Using model libraries seems to be becoming a lot more popular lately which is a sensible progression I suppose. Since the early days of 3D everyone was happy to use purchased texture libraries… yet it was also easier to find high quality textures than models. These days with photogrammetry that has changed and it’s fun to have access to more assets. Becomes more like playing with legos as the end result is more an exercise in composition, lighting, materials, rendering and grading. Feels a bit like cheating, but I suppose a photographer isn’t criticized for not having sculpted everything that is in front of the camera. Seems like such work in 3D is an evolving hybrid… and definitely still a little harder than snapping a photo. Unless you’re Gregory Credson. :P

More Experiments

Just started using MASH in Maya 2017 this week and it is really cool! While it is marketed as being a Motion Graphics tool, it is ultimately an amazingly powerful object instancer/duplicator. This is my first test with it and basically... I don't think I'll be needing to use EITHER particle instancing OR the SPpaint3D script anymore. Which is awesome. i.e. MASH has geo paint built-in which is ultimately way more powerful than SPpaint, while its instancing tools skip the need for writing particle expressions on custom attributes, caching, runups, etc. Another big plus is that instanced objects on a surface, by default, orient themselves to the surface normals, which is why I used grass for my first test. While this was doable with particle instancing, it required a hefty and slow to calculate expression. Anyhow... going to be a big time saver while making the process much more fun!

I've also been building a library of World Machine assets that are ready to import from a Maya shelf, primarily for when I need Mountains/etc in the background. This image was the result of importing one, adding some Megascan materials (which I also have on a shelf ready to go), importing my water plane (on a shelf as well) and... quick island image!

Experiments

Been playing with Maya, World Machine, Redshift, etc as I get ready for demos in Melbourne and Sydney, Australia, next month at Gnomon Live. The goal is to give a variety of tips and workflow ideas as I create a scene which will include some things made on the fly while populating the scene with elements from my library. Hopefully I can share something useful to those who attend! Below are some images that I've added to the 3D Development section as these aren't planned images, just things made while messing around. The character models are the only things I didn't make, but downloaded from RenderPeople.... although I did swap the heads on a couple with ZBrush sculpts. :)

Journeyman

New personal work to test out Redshift with a relatively heavy scene, as well Redshift proxies. As one would expect the interactivity of IPR did slow down, but easily managed by using display layers to hide/show objects. Render time for 4k was about 90 minutes with dual Pascal TitanX cards, although I did crank settings higher than likely necessary since it was for a still. This is still super fast, especially since both 3D fog and DOF are on! I'll note that the DOF is what requires higher sampling, so if I were to do the DOF in post the render time would likely be cut in half... 45min for 4k? Pretty darn cool for a scene like this... I still haven't played with Redshift's AOVs, so this is a straight 32bit EXR render with just a little tweaking in Photoshop with the Camera Raw filter. I also set-up a bunch of Megascan materials that are ready to go for future projects, using several of their rock and leaf materials for this scene. Anyhow... has been a great experience so far switching over to Redshift and GPU rendering!

Redshift

A couple weeks ago I decided to try out GPU rendering. After doing some research and talking to a few people, I was steered towards Redshift. Several studios have been using it and it seemed worth trying out. Was simply curious. A couple years back I switched from Mental Ray to V-Ray and the idea of converting my library of assets, again, was not super appealing. But, fun to try something new nonetheless. Anyhow, after a weekend with Redshift... holy $@#*! Running on a single 980 card it was insanely fast. IPR became useful for complex scenes including fog and DOF, seemingly about 10x faster than Vray. And easier... less fuss to manage sampling as the speed allows for the use of high rates and Brute force. And a bunch of other things but to sum up. It is cool and I've 100% switched. Just spent a couple weeks converting my library. Also upgraded my machine from a single 980 to dual TitanX cards. Yay for nVidia... and Redshift!

Here's a quick test I did after converting my library. IPR constantly refreshing as I imported random assets, set up a couple lighting styles with DOF and fog... all in about an hour. Really fun...

Blog