Forum Replies Created
-
Posts
-
Octane MXE can not race with QuadroFX cards, a Linux PC will be a lot faster both on wireframe and texture fill performance. If you plan to use a 3D software. So for modeling definetely go PC! Also the CPU’s of the Octane is I guess 600Mhz at maximum nowadays, and those processors should cost a fortune, when compared to P4 processors! Altough they have very high performance at that frequency, compared to P4. But still a 3Ghz P4 is a hell faster on many things compared to R14k-600Mhz
For editing, Octane is a very different platform. The softwares are very expensive. And also the video hardwares. A DIVO board should cost $3k at least, 2nd hand. I have seen as much as $6500 for the DIVO board, on a dealer website.
If you are looking for a cheaper workstation, definetely stay on Linux, if it meets your needs. Also smoke on linux will be appearing nowadays, just for SD resolution. NO HD! But I heard the prices will be $60k for smoke on linux, for turnkey systems.
hope these info helps,
$68k for smoke_tezro seems to be a good deal. I mean really good. Tezro is HD capable. Only the hardware of tezro would be close to this price. It is a very good deal. Where did you get this price?
I didn’t have any art education before starting to compositing. Like some of the compositors, I was a 3D animator in the old times. So I had lots of ideas about compositing with computers, when I started to do that job. While I was working for years with art directors, I had a feeling of what is good composition, and what is the correct color to use. Because thousands of times, I saw art directors creating a composition from scratch. I saw what mistakes they see.
Ofcourse you might say, what if you have learned all of these at school? Ok, it could be. But I was a technical guy, who learned about art later. Here comes the big question.
Is it easier to teach an artist about computers, or is it easier to teach a computer guru about art?.. Most of the people say that, it is easier to teach an artist to use computers, then to teach a computer guru about art!
However this might be true, I think that people working in a team should come from different roots. Would you think a team just of very technical people, or just the reverse, all team is composed of artists. I think, a good team needs both of the person types.
If you ask me, there is not one ideal compositor model for all of the industry. But there is one important thing, to be a good compositor, you need lots of artistic skills. You also need lots of technical skills as well.
If you need to do compositing for the broadcast design, you would need lots of sense for typography, colors to use… But if you are compositing the layers in a movie like Matrix, I think it becomes more a technical job.
At last, there is not one correct path only, that a compositor should follow in education.
What is the correct path to become a good director?
As far as I know, AiffExtract will export the amplitude, in the given frequence range. It is not a simple VuMeter.
Let’s say you want to animate, according to the bass drum in the music. Then you give the frequency range of the bassdrum. This way you can easily animate a 3-way speaker 😀
Visit http://www.speedsix.com
They have a spark called AiffExtract. That can export the data. You can analyse any frequency band. It is exactly what you are looking for. 😀
Especially I liked the Z-keying technology. On video productions, it may help a lot. I am now looking to 3DV website for more information. If combined also with the traditional techniques, the Z-key is a good information.
It is also helpful in modeling. You might have heard discreet* 2.5D thing. It is a displacement function by a luminance map. The Z-video is also ideal for that kind of jobs.
I think the technology is evolving. And we go somewhere, where the 3D and 2D jobs come together a lot. Like 3D tracking, etc.. 2D post processes became also part of 3D jobs. So why not do these post processes, even render the 3D scenes in a compositing environment?
Flame can run on 1 processor already. Look at the flame* manual!
“flame -p 1”
will just run on 1 processor. Also the same for inferno! You can run inferno on 1 processor also. This way, you can use your onyx as a renderfarm if you need. This is ofcourse for the good old days, where an R10k processor was the fastest CPU around, while PC’s had only P200MMX!!!
Regarding the cost of dual CPU’s. Does a flame suite owner care, if the CPU costs $1000 or $5000. If you ask me, I wouldn’t even trade 30% of performance increase for $5k! Ofcourse for short terms, if you don’t have spare parts, you can make discreet software work on 1 CPU. But not for price!
“inferno on O2”: I heard that too. But this is completely reverse engineering, and completely illegal. Theoretically, you could steal a Ferrari Enzo too 😯
Flint on the Octane works already on 1CPU only, which is limited softly by discreet. And it can support stones. I haven’t seen any disclaimer in FFI manuals telling that; working on 1 CPU might cause a stone filesystem problem. However discreet might answer your questions better.
Is smoke6 released? Or are you a beta tester? I thought discreet was just announcing smoke6, not releasing it yet.
By the way, regarding the film look. I remember some posts, which told about blurring hi-light areas, and layering the blurred version over the original footage, but just blur the hi-light areas. But this trick was for SD res. I never used that kind of thing with HD.
I would first prepare a matte animation for the blood spread from top view. Then I would try to make color corrections, reflections, highlights using that matte, while matching the perspective of the matte, to your shot.
But I think it will take quite a long time to composite.
Another method would be, shooting a white paint to spread on a black background. Just shoot it, if the PAL res is enough for you, even a DVcam will work. Because you will use monochrome image. This will even look better then particles. And much more possible to achieve… Because to create blood like particle movement, your particle should be flow based, and I don’t think it will be very easy with combustion particles.
One more thing. If you have shot/prepared a matte for blood movement. I would want the matte to displace my surface in 3D. Just like the displacement function in flame/action surfaces. That way, shading, and adding reflections would be easier.
As John said, you problem might be low polygon geometry in action.
But if it is the rendering of action, or text module. Then you are talking about antialiasing problem. If you are in action, try to increase the softness value. This will help, but your text will be blurred.
If you really want thes best antialiasing, you have to go the most old fashion way. Render it at big resolution, and scale it down with sampling.
For example, if you need a 720pixel wide text, I would prepare a 1440pixel wide text, then blur it Gaussian, I think 1 pixel will be enough. Then scale it down to 720 pixels. And always visually check, then you don’t blur the text too much.
mseymour7 wrote:Discreet had a technology showing of a flame rendering using Mental Ray.I think that MR technology, within flame*, is a big power. 3D and 2D are getting close. Let’s see where it will go.
inferno* renders via OpenGL, all the layer transformations, image blending, etc. Blur’s, keys, etc are rendered with CPU.
But to disable “texture” feature in inferno. Just go, and toggle “setup-texture” button. This way, transformations, and blending of images will be rendered with CPU only.
Hope this helps
In import image menu, you have to select the option “RGB+Matte” option. It is a blue colored box.
If you use v7 of flame, there are 9-10 blending options. But if those blending modes don’t suffice, why don’t you look to logic-ops?