[Keyser_Soze]

Ajax wrote:

I am creating a series of Style frames within photoshop. I have placed each set of style frames within a series of photoshop bins or folders. Can combustion recognize these folders or do I have to take each layer out of the folders and then import into combustion? After Effects recognizes the folders and creates precomps based on that, wondering if combustion does the same. Cheers, 😉

Combustion will import your different folders but won’t group or nest them. They will be split up into layers in the comp (if you want them to, they can also be merged). Not sure how that will affect transfer modes, transparancies and layer effects … probably in a negative way.

[Keyser_Soze]

_joro wrote:

yeah, i’ve seen this – that’s why i’m asking

the problem is that i can’t make the mattes

i tried difference keyer, lumakeyer, ultrakeyer, but not of them produces the mattes shown in the movie

pls tell me what tool to use and how to combine them in pairs for creating the new RGB plates

10x

This is how i think it works… if you read the text at AviatorVFX:
“In short, natural skin tone was achieved by filming two black and white strips of film (with a red and green filter on the lens) … By digitally re-filtering the layers using a version of a primary color matte.” … If you picture a density matte of the blue areas of the film only (grey in the blue area and white in the non blue)”

What you will need to digitally emulate is those red, green and blue filters to get the “density matte” of the RGB channels. That result you will later combine. I’m not exactly sure how the actual analog colour filter works but I guess you could simulate it through a keyer by keying your original footage on pure red, green and blue respectively. You will probably need to go back and tweak your key to get a nice result. Then multiply the color filter mattes that you get with your original r/g/b channels as described on the VFX site.
Maybe that’s what was in Deepray’s flow but I cannot say because I’m not using DF. So you might have the answer already. Good luck anyways.

[Keyser_Soze]

hyrlvlrec wrote:

a sequence is def a workaround and thats what i usually do….. but i’m just curious……. is quicktime export absolutely useless in the discreet products? useless as in they will not work for anything more than PLAYING in a QT player???

To make it short… YES 😀

[Keyser_Soze]

hyrlvlrec wrote:

well, the file im attempting to create (30sec) is not 2gb….. at 40mb a sec, thats only 1200mb total

and speaking of using the system’s time to create a qt….. for a 30sec spot it is def way faster to do it this way than it is to go out to tape and then back into another machine from tape…… always a pain

btw, is the irix 2gb qtfile limit just for creation of qt’s?? because i successfuly imported a 45gb uncompressed qt file that came from a final cut system the other day just fine

As a workaround why not export a tiff-sequence and an audiofile? Then you can put it all together in shake, combustion, afx or whatever and encode to your preferred format That way you will have much more control over compression-settings and such.

[Keyser_Soze]

jamie_t wrote:

Hi,

Had a question regarding tech specs for the linux flavor of smoke 6.5. i understand it can playback 2k files in the timeline in realtime but can it import/export 2k files to a network resource? is the color resolution for component still stuck with 12 bit processing or can the system import, create and export in 16bit?

Thanks in advance,

Jamie

You can import/export 2k files from disk to Smoke/Flint on linux but …
You can’t work in more than 8 bits. And if you want realtime playback you’ll have to use proxies.

[Keyser_Soze]

foetz wrote:

i remember the days some cgi facilities thought they have to switch to linux cause
it’s trendy and cheap.
the trouble began… 😆 😆

this was ‘only’ 3d desktop stuff. uncompressed video/film must be a nightmare.

We all know that your heart stays with muscle machines with hair on the chest. 😉

[Keyser_Soze]

johnmont wrote:

that’s a good call…I’ll start a forum area about the art. we’ll see what kind of response. If its good, we’ll leave it. If not, I’ll shuffle the messages around to an appropriate forum.

I’d love to get that kind of discussion started.

A really good idea. Start that forum right away.

[Keyser_Soze]

Jaycee wrote:

These, for me, are what turn into a murky web of confusion….. there has to a process i’m missing?

thanks!

I don’t think you are missing anything really. It’s just different ways of working. I think you have to be just as accurate naming your files and sort them in folders on your pc or mac as you have to be on any discreet system. I actually find the desktop and reel system on FFI really fast and neat. It may be basic but it does what you need, and fast.
Just name your clips and reels in a sensible way, it doesn’t have to be as “cryptic” as in the guide I mentioned before. Just name them what they are and maybe what you have processed them with and what version (just as you would have on your pc). Then sort your clips in reels. One for your telecine source or source material, one for your conform, offlines, master building, different reels for comps, 3d material, one reel for graphics and packshots and one for final masters, that was just a few examples.
Sometimes reels with many clips can look quite dense. To make them easier to navigate you can sort your clips like this – just put a single yellow or any other coloured frame between files that “belong together”. That way your reels will get a cleaner look.
Name your setups the same way you name your clips and reels thus it will be easier for somebody to access your work. It is also easier if you keep all of your setups to batch then you know (and anyone else) where to find things. Talk to your friends at work and agree on a standard way of naming things. You will find that with just a little effort it will be much easier to take over somebodys work or open something you did yourself a long time ago.

[Keyser_Soze]

LK wrote:

There’s a tip in the tutorial section of this site.
You can use a frame with thin horizontal lines as a mask, then a little bit of blur and cc make the rest. It’s quite an easy and speedy way if you don’t need super-realistic results.
L.

The guide:
http://www.fxguide.com/fxtips-32.html

[Keyser_Soze]

Jaycee wrote:

its really quite dizzying and I was hoping some sympathetic soul could describe what works for them.

Thanks! ❗

Maybe this guide is a bit out of date now with clip history and everything but it still makes sense naming your clips a standard way. This is how “Urs Franzen” does it, take it away Urs!:
http://www.fxguide.com/fxtips-140.html

[Keyser_Soze]

RalpB wrote:

What would be the quickest and easiest way to give a picture a television type look, with lines and such…without any type of plug-ins ?

It depends a bit on what you mean with TV-look and how “bad” you want it to look. In addition to what was described above you might want to add some “snow”. Generate a clip with bl/w noise, blur that and apply with logic ops. You could also try to achieve a “ghosting effect” by offseting your image and experiment with logic-ops, don’t know if it will work.
If you want it too look really gritty you could try and output it to VHS and bounce it a few generations and bring it back again. If you want tape distortions as well you can just rip out the tape and step on it and play that. It will be very analog and look actually better than you average tv-plug… but the result will be a bit unpredictable.

[Keyser_Soze]

Polax wrote:

Well, my question concerns about flame/flint – vertice tracking in WARP EFFECT.
What is made for?
If you can track each vertice of your mech (For example, 4 point on your source image), I assume you can use it for warping…but how?

As i’ve said before “Cunningham” type of warping is probably not very practical with vertice tracking, at least not as far as I know.
One area where I know we used the warper and vertice tracking were when we had a shot that was very hard to stabilize the conventional way. It was a shot that should be stable but had a little camera movement. That wouldnt have been a big problem… but in addtition to that there were wobbeling from a bad telecine job. So the image had motion from both camera and the bad/broken scanner. Just untrackable in the stabilizer. There would still be motion from either the camera or the telecine.
So we took it to the warper created a mesh with a few vertices positioned strategically. That grid was then copied to destination. Track your source mesh and voila you have the “stabilized image”. It wasn’t perfect but it gave us a result that we could live with.
As always good luck with your warping. Cheers.

[Keyser_Soze]

johnmont wrote:

Keyser_Soze wrote:

To be clear, the revisionfx plug in is the same (or close to it) on after effects, combustion, and shake. But it is not the same as the new Distort node in flame9 or inferno6….that was done by discreet.

Sorry, I was a bit unclear on that. What I wanted to say is that the basic design is pretty much the same. So you don’t have to go spend all of your money on Flame/Inferno/Flint if you want to start warping “cunningham style”.

[Keyser_Soze]

Polax wrote:

Do you really think this technic impossible….?

What kind of software and version are you on? I think the new distort node will make life easier for you. With better tracking capabilities and control it will be more than enough for you. And you won’t need four point tracking or whatever to get a decent track. And you will have no keyframes at all on your spline if you are lucky. By the way you will find the same distort node as a plug in for after effects, combustion or shake. –> http://www.revisionfx.com/rflx.htm

[Keyser_Soze]

Polax wrote:

Do you really think this technic impossible….?

Well it seems a bit backwards to me. Maybe you will get a better track but only if the head isn’t moving or turning, cause if it is you will end up with a bad result anyway and you will have to adjust it manually. And as a result you will have a “million” keyframes to adjust so it won’t be possible to get an even result unless you find a way to adjust the mesh at every keyframe. It might be possible through translating all keyframes in the animation menu. But I don’t think that it will be an easy thing to do. And as soon as you need to do different adjustments in time because of headmovements you’re lost. Maybe I’m wrong? It is a way that I haven’t explored so it might be possible… it looks hard to me though.

[Author]

Posts