Ultraviolet

[SoLiTuDe]

i'm no pro, but i'd say 50% of the movie was not up to movie quality standards anyway... i'd expect to see that movie on the sci fi channel, not in the theater... not just because of the effects, but the entire movie sucked... I'd say the the short answer is no.

[dellis]

Haven't seen it yet...but i'd pay the price of admission just to see Mila on a 20 foot screen.

[jeremyhardin]

I saw this movie last weekend, and I have a few questions. I know this board is full of VFX pros so my question is this, were the effects in that movie acceptable to the studio movie standard? There were some shots that I know I could have done myself on my 1 grand pc with combustion only and it would look just as good or better. I am not trying to insult the people that worked on the movie, I am sure there were many reasons it looked the way it did, too many changes, to little people, too quick of a deadline, I dont know. Thanks in advance for your response.

I'd say no personally, but due to the nature of some of the things they did (almost multi-plane anime-esque camera moves), I assume that some of that was a stylistic decision. In which case I can only say that I didn't care for a lot of style.

I liked the look of the opening motorcycle chase (and multi-plane stuff), but the median/smart-blurred faces in closeups were KILLING me. By the end of the movie I wanted to never use or see those effects again.

It's kind of like the behind the scenes DVD of Queen of the Damned. Did you ever see that? They talk about the vampire-speed-trail effect as the product of all this work. They tote it as an achievement of VFX. And I hated it. I thought it was cheezy.

[Bonedaddy]

Hi. I worked on UV -- the opening shot and the first half of the motorcycle chase, doing the particle simulations and technical direction. Check my CGPortfolio for specifics. I only worked on a small portion of the movie (~30 shots); the rest was done by Menfond, a Hong Kong based visual effects company. I believe this may have been their first movie, although I do not know for certain.

Let me just say I was disappointed with the end product. I worked my butt off on this movie, and I don't think a lot of it shows. We had a lot of difficulties with the show, the main one being that the look had to be really clean and antiseptic -- meaning we couldn't put any dirt or grime on anything. We tried to compensate for that by putting a lot of little details in, but some high-contrast post-process was applied after I handed in my work that destroyed a lot of that detail. So that was frustrating.

For before/after comparisons, you can look here (http://www.complexdream.com/tmp_images/cgtalk/uv/). I took frames from the trailer and did an a/b with what I had done.

I am not sure why they put the post-process filter on. I heard, and this is strictly rumor that I got off another website, that there were some troubles with the colorspace (they shot on HD), and because most everyone was dressed in black, they either had to smooth over peoples' skin and drop everything off into black, or bring up the black details and make peoples' faces hideous. I guess that was the problem? I don't know for certain.

I wish I had had more time to polish the work. I wanted to add dust to the bullet hits, and have it affected by the motorcycle, swirling in its wake -- I even had a rig to do this -- but didn't have time, and as I understood it, the powers-that-be didn't really want it. I also wasn't satisfied with the shaders or the models for the scene in general, but that wasn't my department, and apparently people higher-up than me liked them and signed off on them before I even came on board.

I don't think my work was brilliant, but it was the first show I was an artist on, and I worked really hard on it, for around a year and a half. You win some, you lose some.

[jeremyhardin]

Bonedaddy, good work. Like I said, a lot of it comes down to what 'they' want, so I mean nothing negatively toward your or other artists work. I know how it is in this biz. :D ;) We do what we can, and they keep what they want. Sometimes we agree, and sometimes we don't.

What did you do the particles in, if I may ask?

[Bonedaddy]

Bonedaddy, good work. Like I said, a lot of it comes down to what 'they' want, so I mean nothing negatively toward your or other artists work. I know how it is in this biz. :D ;) We do what we can, and they keep what they want. Sometimes we agree, and sometimes we don't.

What did you do the particles in, if I may ask?


I did it in Maya, with some funky PRman implementations. It was actually a pretty complicated rig, although it may not look it. I did 30 different individual debris particle instancer simulations in Maya -- 10 for glass, 10 for marble, and 10 for metal sparks. I then had a perl/PRman script that would basically take a proxy object (a large cube) and render it as whichever particle simulation I wanted, at rendertime. So I never had to do any particle sims within the scene, I could write the .ribs very quickly, and I could even jitter the position/rotation/scale of each individual particle in the system on a per-proxy basis through the perl script.

I then linked this up to a script where I would place "squib" locators around the scene, which had two custom attributes -- a frame that the squib would go off at, and what sort of hit it was, whether it was marble, glass, or metal. What the rig would do, given those locators, was:

Aim the gunners' gun
Fire a bullet far enough ahead of time to hit the squib locator at the right time
Fire a tracer round (with smoke trail) every 5th round
Emit a shell casing
Turn on a "muzzle flash" object
Turn on a "muzzle flash" light
Emit a "sprite" at the hit that would be some instanced geometry of a small marble dent, or some broken glass, or whatever
Turn on the proxy object

So all I had to do was mess with locators and timing. The director was very particular about that. The tough part was actually getting the renders through. On one of the shots, where you could see every bit of debris throughout the entire shot, I had to divide the shot into part 1 and part 2, and render both on 4x4 tile solutions. There were just too many particles, all instanced geometry, all motion blurring and rotating. I think that shot had upwards of 20 hour frames for our main pass.

[JoshKirk]

That was actually really good stuff. The parts that werent so good were the effects on her hair in one of the first shots, and the fire at the end wasnt very believable blowing out the glass window.

[jeremyhardin]

I did it in Maya, with some funky PRman implementations. It was actually a pretty complicated rig, although it may not look it. I did 30 different individual debris particle instancer simulations in Maya -- 10 for glass, 10 for marble, and 10 for metal sparks. I then had a perl/PRman script that would basically take a proxy object (a large cube) and render it as whichever particle simulation I wanted, at rendertime. So I never had to do any particle sims within the scene, I could write the .ribs very quickly, and I could even jitter the position/rotation/scale of each individual particle in the system on a per-proxy basis through the perl script.

I then linked this up to a script where I would place "squib" locators around the scene, which had two custom attributes -- a frame that the squib would go off at, and what sort of hit it was, whether it was marble, glass, or metal. What the rig would do, given those locators, was:

Aim the gunners' gun
Fire a bullet far enough ahead of time to hit the squib locator at the right time
Fire a tracer round (with smoke trail) every 5th round
Emit a shell casing
Turn on a "muzzle flash" object
Turn on a "muzzle flash" light
Emit a "sprite" at the hit that would be some instanced geometry of a small marble dent, or some broken glass, or whatever
Turn on the proxy object

So all I had to do was mess with locators and timing. The director was very particular about that. The tough part was actually getting the renders through. On one of the shots, where you could see every bit of debris throughout the entire shot, I had to divide the shot into part 1 and part 2, and render both on 4x4 tile solutions. There were just too many particles, all instanced geometry, all motion blurring and rotating. I think that shot had upwards of 20 hour frames for our main pass.

Nice one! Sounds like a b*tch to set up and a joy to work with once it was done. :thumbsup:

[ParamountCell]

I did it in Maya, with some funky PRman implementations. It was actually a pretty complicated rig, although it may not look it. I did 30 different individual debris particle instancer simulations in Maya -- 10 for glass, 10 for marble, and 10 for metal sparks. I then had a perl/PRman script that would basically take a proxy object (a large cube) and render it as whichever particle simulation I wanted, at rendertime. So I never had to do any particle sims within the scene, I could write the .ribs very quickly, and I could even jitter the position/rotation/scale of each individual particle in the system on a per-proxy basis through the perl script.

I then linked this up to a script where I would place "squib" locators around the scene, which had two custom attributes -- a frame that the squib would go off at, and what sort of hit it was, whether it was marble, glass, or metal. What the rig would do, given those locators, was:


Aim the gunners' gun
Fire a bullet far enough ahead of time to hit the squib locator at the right time
Fire a tracer round (with smoke trail) every 5th round
Emit a shell casing
Turn on a "muzzle flash" object
Turn on a "muzzle flash" light
Emit a "sprite" at the hit that would be some instanced geometry of a small marble dent, or some broken glass, or whatever
Turn on the proxy object
So all I had to do was mess with locators and timing. The director was very particular about that. The tough part was actually getting the renders through. On one of the shots, where you could see every bit of debris throughout the entire shot, I had to divide the shot into part 1 and part 2, and render both on 4x4 tile solutions. There were just too many particles, all instanced geometry, all motion blurring and rotating. I think that shot had upwards of 20 hour frames for our main pass.



hey good work bonedaddy dude what is the process for getting a job like this?

[Bonedaddy]

hey good work bonedaddy dude what is the process for getting a job like this?

Doing something terrible in a past life.

I started out as a render wrangler, was brought on as a junior TD, because I had fairly technical leanings (shell scripting, PHP, HTML, Javascript, MEL), and they thought they could groom me for the position. It worked out pretty well. I started out just modeling simple buildings (actually, the windows of the buildings) for the first shot, with the balls flying over the city, and moved on to more technical stuff.

[SoLiTuDe]

Bonedaddy -- great work man... I really don't want to put the movie down. I know the hard work that goes into it, but yeah I'd say the parts towards the end were just crap.... a lot of the other fx were good... the bullet stuff was great... but the fire... I'm an afterburn junkie, so it guess i'm biased, and just nitpicking... but even then I'd say as an overall movie it was just a crap story with a crap cheese all over it :thumbsup: But at the same time... if somebody wanted to pay me to work on the movie I would've said hell yes in a heartbeat! :D

--20 hours per frame... eewwww but i've done that before.

[Bonedaddy]

No, don't feel like y'all have to hold back on the movie just because I worked on it. I thought it was a terrible, terrible movie -- one of the worst I've seen in recent memory. The effects work throughout was shoddy, and even the work I did is nowhere near the level of polish I expect in a major movie. It could have, and should have, looked a lot better. Such was the nature of the beast, but it doesn't excuse the movie from sucking harder than a dehydrated hooker.

[Aneks]

Bonedaddy we have all worked on poxy films, I just kkep telling myself... at least its a film.. if you can get a few good shots out of it then consider it a blessing.

Just a quick question on those stills you posted. Are the actual final or just slap comps ? because that girl on the motorbike ..... Dude !

[Bonedaddy]

Honestly, I don't know if they were final. I was put on another show after my fx duties wrapped up on it; those were current as of last check. I think they're pretty close to final. However, don't take the quality of those comps, or the rest of the work, as indicative of the work all of us were capable of -- it was a tough show, for many reasons.

[Aneks]

Didin't mean to sound critical. Just having an sunday-afternoon-at-work Grump session ...

Seriously its a good shot on a feature anyone would be happy with that !

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.