Tuesday, December 8, 2009

The golden $1600 new computer ratio


I have this theory. The best home workstation should cost no more than $1600 to build, with or without keeping parts from your old box. I work on some of the most expensive computers that make business sense to purchase, and run around $4k-$6k. I get a new computer at work about every 12-18 months.

It breaks down like this, I can spend $1600 for a great mid level rig or $4,000 on high end one. The $1600 computer is designed to last me 18 months and then I get a new one. The $4,000 rig will have to last 3-4 years at least to be worth it money wise. The second $1600 computer will be as fast or faster than the high end workstation from 18 months before.
I can tell you that a hand built (by you @ newegg) will not be THAT much slower than a $4,000 computer to begin with. More importantly, what is slower is ok to be slower. If you're doing particle/fluid simulations, or 3d rendering, then the slowness will cost you in time, which is just inconvienient not costing you money or frags. Small freelance gigs will be manageable with the newish proc and won't be significantly slower.

The difference will be ram size and proc count, not ram or proc speed as much, esp if you get into even minor overclocking. The sad truth is hardware depreciates to "nothing" so fast whatever you spend on a computer is money being slowly thrown away. Keep it at midlevel and keep it fresh, and you'll be fast and always up to date with the most bang for your buck.

I'll end on this story. My friend got a $4,000 workstation at the same time I got my $1600 one. His was 8 core (2 quad @ $1500 each) His computer is fast and so is mine. One night his one proc burned out due to a cpu fan lock up. He didn't have a warentee and didn't buy a whole system from a company like Dell, totally his responsibility. He couldn't afford another $1500 proc and the other didn't work with just one. He had to buy 2 new procs at a lower speed in order to get back up and running. Another reason to buy a computer where any fail component is easily replaced.

Monday, December 7, 2009

fluid rendering: handware/software mix

Many Most people who've worked with fluids, have noticed that it looks better in openGL than the render... in some ways. I'm refering to smoke, not more shader based fx like fire. Smoke will have some details that seem to get lost in software render, but software render also has the benefit of smoothness of gradients and propor lighting. Those 2 key aspects keep up software rendering. My friend at work was talking about trying to hardware render his fluids and said it was almost good enough. I played around and came up with this.
hardware render for the added detail and blur/add it into comp over top of the software render. I haven't done this on production level stuff but on my initial tests the results are promising.
Here's one test frame, I'll try to get the render online.


Maya Fluid Software Hardware Mix from destruct007 on Vimeo.



Friday, December 4, 2009

Fracture alpha/beta Testing.

I'm alpha/beta testing 
http://fracture-fx.com/
very cool shatter tool with an RBD solver akin to physX but working with it on 64bit linux maya2009. Some of you may know that there is no physX solver for that combination... which is probably the most common combination in the industry right now. Annnywhoo. See if you can get in on the alpha/beta it's cool.

Since BlastCode has been waiting for library files to be updated that don't seem like they are going to be, it's been dead in the water. Fracture is a differnt method of shattering, more like rayfire on 3dsMax. I'm currenltly working to build this into my destruction pipeline... so far so good.

The key to a good plug-in or even software is the motivation and passion of the developers. Work hard on it and push it to be great. Take mudbox for example. Was making leaps and bounds then got bought, and what? ZBrush blew it out of the water. Passion, drive, motivation, skillz... these are the things that make tools great and me want to use. I think fracture will make some waves, and I was skeptical about it for a long time.

Maya could do a great nRBDsolver and shatter tool but I don't see that happening in a timely way and like I said before, I think fracture will push harder and make greater strides for a number of reasons. Time will tell.

Wednesday, November 4, 2009

Motorola DROID: Stealth commercial

30 frames of fx, hah. 
http://www.youtube.com/watch?v=o9fXYQjwR0w

I did the ground explosion at :30 ubber quick, helped out with a few hours after my normal hours to help them finish (still at Asylum just not on my normal project)

Friday, September 25, 2009

The Game Plan

EDIT/UPDATE: Taking a term off from teaching and going to push hard on upping my skills. Of course in Maya effects, but also in other areas... here's the plan. Right now, 3dsmax, Fume, Krakatoa, Fracture, and Maya are my main focuses.

Nuke :: working knowledge - advanced
    currently : none
    plan : Gnomon Nuke DVD's or just have some guys at work show me the ropes.
General Compositing, going through "The Art and Science of Digital Compositing" <- great book

3dMax :: general good working knowlage
    currently : Got 3dsmax, doing some basic tutorials to learn the interface
    plan : Assortment of Gnomon DVD's (not sure yet which ones)
Fume :: advanced
    currently : got it, not working.
    plan : mess around with it, have my friend Ian Farnsworth  show me a thing or two
Krakatoa :: working
    currently : none
    plan : online tutorials


Maya
advanced lighting; rendering in Mental Ray
    currently : good
    plan : Gnomon DVD's and read "Introduction to Computer Graphics" by James D. Foley

Fracture (Maya Plug-in): advanced
    current : some tutorials
    plan : use it at work (got it) and get really good with it.

Houdini :: not actively 
    current : decent can get around and do what I need to, but I'm slower right now in H than Maya

The overall plan is to get 3dsmax, Fume, and Krakatoa into Asylum so I can use it everyday on shots that would benefit from it. Thats the best way to learn it, but have to start by showing some cool examples of what I can do with it first.

Comments enabled!

Finally figured out why comments weren't enabled.

Friday, September 4, 2009

Halo ODST

It's not much (what I did) but in 2.5 days I made 2 of the alien ships explode with smoke trails. The easiest one to point out is the crash right before the fade to the future of that Marine or whatever master chief kind of guy he is. Anyhoo, cool trailer/short thingy!

http://xbox360.ign.com/dor/objects/852871/bungie-project-2/videos/haloodst_liveaction_trl_090409.html