Another Car render, but this time much more fun.

Posted on by

No rant today. This is just an update on some of my little projects. After my last Buick that I just couldn’t decide on I stuck with a new car, and renderer, rather than playing around. Like I said before my fluids are still a mess, but they’re MUCH better. Maya fluids in Arnold is a no go (I’m missing something I’m sure) because they just take years!! I think next I’m going to try to get FumeFX working with my Arnold version. I hear good things about it.

So first off, here’s my playblast. I animated the car using Craft Director, and did four basic camera moves. I then pulled the playblast of each into After Effects to test out a cut and that then gave me a good idea of what frames to render rather than rendering 200 frames x 4 cameras.

 

This is the full render and I added some sound of my own. Took a little bit of hunting but I found some guy on YouTube with the same Ferrari doing a donut, so that worked out perfectly. This was comped in Nuke 8 and the grade in After Effects. I’m proud of myself, I pulled in my 3D scene through FBX with some nulls attached to the lights, and then used their 3D positions to place the flares with Optical Flares.

 

And what’s happening with me? My eyes are starting to wonder; I’d like to get my reel properly done and see what is out there internationally. I’ve got Weta and ILM on my radar but who knows. I can only hope that I won’t be too much of a small fish in a big pond again.

I found some Transformers models and I was keen on using them for something new, but that isn’t happening because they’re terrible. It’s mostly outside shells and none of the bits and pieces. So scrap that idea. Maybe I’ll do something more along the lines of VFX and make something build itself onto my arm a la Iron Man. We’ll see. I may just go crazy into the detail of the mechanism… I’ll try not get carried away again.

I’m tired… But look what I finally decided to finish!

Posted on by

So there’s 16 minutes left until the video gets converted on Vimeo… so let’s rant… or something.

What you’ve got here is a good example of lack of direction and vague final goals. I did learn a lot though. I’ve learned that you can’t practically render Maya Fluids with Arnold, that Maxwell is pretty… VERY pretty – but there must be a better choice out there, and that Mental Ray is probably the best thing for Maya Fluids.

I also reinforced my rant that I always have about scripts, storyboards, shot lists, and basic pre-prod stuff. When I started this I had a vague idea of this car running down a road and doing a wheel spin donut type of movement. What a laugh. So I took a car, set up shaders, put it in an environment (learned a lot about proxies), hit render and that took forever. I put it in Maxwell, did shaders, put it in a studio setup, rendered and it took forever. Some of the render times I had shot up to 20 hours a frame. Now that’s ridiculous.

Another thing that this taught me is that you can’t really use a single renderer for everything, not REALLY. I’ve been shifting over to Arnold and using it for my last few productions (professional things at work) and it has so far proven the best in terms of realism, speed, and ease of use. Problem with it though is when I render Maya fluids in Arnold (especially the large puffier clouds) then it takes so long that I would have waited until Christmas. I’m obviously missing something. The same fluid took at most an hour per frame in Mental Ray and looks good.

Moving on…

My little YouTube channel is ready to go. I’ve got a good microphone, tracking hat is ready, show body is all set up, now I must just add content and post my first video.

I recently got a call from a new client to do some logos for an event of theirs due in three days. Obviously wasn’t a problem, two days before the event they suddenly wanted a three minute animation with what was essentially a two page script. An animated power point if you will. Needless to say things didn’t go to plan and they scrapped the animation on the day; they did like the logo animations as far as I know. We actually had time for previews and changes for those. I don’t know what happened on their side that this got left to the last minute so I can’t blame them for the insane deadline. I’m sure I’ve said this before, but pre-prod and proper brief and planning seem to be held in a very low regard, maybe not in the country but definitely in my experience in this city. I have to start putting my foot down, and stop saying yes to everything that comes my way. If us pixel jockeys don’t set the right standard and procedure then things inevitably fall apart. I suppose you could say this is my fault then… hahaha!!

On the same note I’ve also got this director that likes to sit next to me while I work. Who does that? Whoever you let do that, that’s who, stupid, Charl! I’ve actually been ranting with this guy about the lack of planning, shot lists, etc., and he’s concocted all sorts of elaborate explanations for why he has to sit with me while I work. Lack of planning, that’s why.

If I don’t stop now I’m going to make myself angry, I’m already cranky having had such little sleep.

Fluids tests

Posted on by

https://vimeo.com/77487784;

I finally decided on what to do with this car. I’ve made the side badge into my name, used Craft Director Studio to animate the car doing a wheelspin, and then used Maya fluids to do the smoke. I originally tried using particles and planned to use Krakatoa as the renderer but I just couldn’t get the right motion I was after. So this shot is going to be the opening to my new reel. I haven’t made one in about two years now and it’s high time that I did.

This took about two days of fiddling to get right. Most of the time I would run out RAM and the computer would freak out and crash. Also if you were to cache from the wrong place with the wrong method then you’ll break the cache file as well. I didn’t want to compromise on the fluid resolution (It was only at 100) so I had to use the auto resize feature and I messed that up as well. I thought I knew this stuff!

Auto resizing doesn’t work if you’ve got your threshold at 0. Surprise dumass! So I eventually settled on a threshold value of 0.1 and a margin of about 4 voxels. Finally the thing would resize appropriately and I would NOT end up with a 1.5GB cache file per frame. I’ve got 16 gigs of RAM on this machine and it got filled up so quickly.

Right now I’ve got the render going on the farm and it seems to be coming out quite nicely. Let’s wait and see.

First Shader Test

Posted on by

I’ve started a new project. I’ve seen many people do cars; I’ve never done one myself. So here’s my attempt. This is just a shader test. It’s going to be on one of those LONG American roads in the desert somewhere. I’m going all out, smoke, tire screeches, the whole lot.

I’d recommend downloading it and playing it in a loop.