http://forums.cgsociety.org CGTalk, the CGSociety's official forum for digital artists 120 en-us http://forums.cgsociety.org/showthread.php?t=1108405 Posted By: Geuse
Post Time: 05-22-2013 at 09:35 PM
Text:

Hi everybody!
Lately, I've been interested in pursuing and creating effects for film and learning the trade of compositing.
Among other things I've been wondering about the techniques involved in set extension and I basically just have some questions on that.

To start off with I'm going to take two examples to see if I understand this correctly.

Ex.1
http://www.youtube.com/watch?v=47CMH8Aqxxo
In this clip the whole environment is basically pre-rendered and composited together with a live action set(in some parts). Evidently some parts are full 3D. I just assume this can be quite tricky to make it come together seamlessly concerning color matching, light mathcing etc. Or do they in fact mask out the live set at the end and replace it so what you see is all pre-rendered?


Ex.2
https://vimeo.com/21666290
This seems to be done just using camera mapping and doing that you don't need to prerender or match anything more than what can be mocked up in photoshop. Also the integration is alot easier to achieve. This is "somewhat" more cost effective since you usually just have a more or less fixed camera, but I assume this isn't always the case?

So I know I'm jibbering alot in my posts when I don't have a really clear question, but hopefully you can follow along and perhaps give some inputs to fill in my blanks on the subject.

My main concern is how to make the integration look great and to see if my thought process is on the right path concerning the difference in the way you create set extensions.

Thanks in advance.]]> Wed, 22 May 2013 16:35:46 +0000 http://forums.cgsociety.org/showthread.php?t=1108225 Posted By: fruityvince
Post Time: 05-21-2013 at 10:27 PM
Text:

hello there !

Ok, here is my problem ! I'm currently working on a short film, in CG, and for some shots, i need to track some objects. I can use the nuke tracker, of course, but it looks quite old school and not really efficient to me, for some specifical situations.
So i'm wondering if there is a way to get back some exact informations from my 3d soft (maya, in my situation). I'm really unfamiliar with all vfx stuff (including tracking^^), but i guess that if i get my world position's mesh and my camera from maya in a text file, i should be able to use it in nuke, right ?
So does anyone could help me with that, please ? Do you have an idea on how to do my tracking easier than with nuke tracker ?
(i assume that scripting in maya is required, of course, but that's not a problem^^ i can't expect doing what i want to do without python =)]]> Tue, 21 May 2013 17:27:54 +0000 http://forums.cgsociety.org/showthread.php?t=1108067 Posted By: kosmarnik
Post Time: 05-20-2013 at 10:48 PM
Text:

Hi,
I have thousands of tiff16 files I'd like to painlessly convert to openEXR en masse (windows).
Google is mum on the issue.
I do know that ImageMagick could support openEXR but needs to be compiled that way and no precompiled windows binaries in sight.
Any ideas?]]> Mon, 20 May 2013 17:48:34 +0000 http://forums.cgsociety.org/showthread.php?t=1107999 Posted By: Toms993
Post Time: 05-20-2013 at 12:01 PM
Text:

I have a simple rectangle shape wich is moved with hands and no camera motion. i would like to track only the object motion, but i cant find any good tutorials. Only ones i have found is with old UI. Also tried to have a look in documentation, but it wasnt helpful enough to get how to do that.

Thanks in advance,
Tom]]> Mon, 20 May 2013 07:01:45 +0000 http://forums.cgsociety.org/showthread.php?t=1107938 Posted By: maty241
Post Time: 05-20-2013 at 01:41 AM
Text:

I can't seem to properly motion track a very simple clip. I've done many successful motion tracks before, but this one seems to just not work. It's a very simple pan shot with a flat surface.


[http://puu.sh/2X4l2.jpg ]

After setting up the scene geometry, no matter what I do, and how many X Y Z reference I add, it always thinks the road is a vertical plane, and not horizontal.

I think maybe it's because the trackers that are very close to the camera move at the same speed as those farther.

I tried to solve the same clip with Syntheyes, The Foundry Camera Tracker (AE Plugin), and lastly boujou. All make the same mistake. Is the problem just my clip ?

Anyways, any help on how to improve the solve would be great! Thanks in advance.]]> Sun, 19 May 2013 20:41:12 +0000 http://forums.cgsociety.org/showthread.php?t=1107181 Posted By: rmartori
Post Time: 05-15-2013 at 03:55 PM
Text:

I wasnt really sure where to put this but two of the options might involve compositing so i figure here and if not I'd be pointed to the right forum. Sorry this is kind of long.

I am working on a mock up to show how augmented reality and 3d printing could be used in museums to display artifacts that aren't normally displayed. But I am running into a bit of a mental block in the how best to display the mock up.
The original plan was I took the laser scanned 3d model of an artifact, changed it's texture to looks like it was printed in a 3d plastic printer then made a model hand holding a model phone and had them move over the model to "scan" it, playing a movie of the camera lining up and starting the AR sequence as a texture on the face of the phone. It's....okay, though I haven't gotten to the interaction portion of this yet and having the 3d laser scanned model being "manipulated" on the screen I have hit a mental block as to if this is the best way, my 3d modeled hands are mediocre(not bad but not great) and I am not sure if i can do lifelike enough animations yet(I was trained more in making logos sadly, self taught all else).
So I am trying to come up with other ideas as to how and was wondering if anyone had any ideas

basically I am trying to show something similar to aurasma scan with a smartphone the plastic model it brings up a fairly photorealisitc 3d scanned model that you can rotate zoom etc and on top of that using after effects I added three buttons one to bring in a lightsource(an omni light) that you can move around the model to show more detail, a map button showing where it was found, and an information button bringing up information about the artifact.

So I am looking for advice should i stick with this current plan of doing all of this inside 3dsmax and after effects. or should I try and bring a live action element into this and try and use composting or some other trick to show an actual phone displaying this? I honestly haven't done composting before but I would be more then capable of learning. The two tricks I thought of are either A) maybe taking some blue tape putting it over my cellphone screen and finding a way to sort of over lay the video on it. Or B) just running it as a video on my phone, the downside with this one is that if i am to show the interaction of the model how can i touch the cellphone without pausing the video hah
Sorry that I don't have any pics, the digital artifact is protected information, I might try and make a mock up of the mock up though since the artifact is basically a cylinder.

Hopefully you can give me some experienced advice on what the best way of demonstrating an augmented reality mock up(yknow without actually programming one) would be. Thank you.]]> Wed, 15 May 2013 10:55:50 +0000 http://forums.cgsociety.org/showthread.php?t=1107138 Posted By: MBahnen
Post Time: 05-15-2013 at 09:56 AM
Text:

Hi there,
I have a serious issue that i need to fix as soon as possible: any help is much appreciated!
I have a Cinema 4D R14 multipass rendered image (PSD, layers) which has a linear color profile. Opening this in Photoshop CS6 and taking the shortest route to convert to AdobeRGB will result in a colorshift and incorrect gamma. How can i convert the mulitpass image correctly without any change in color??? ( i know...i should have thought of this before rendering...my mistake. Re-rendering is no option due to the rendertime of this 80x50cm 350 ppi image with GI and AO).
Anyway...i have this linear image...need to convert it to adobe rgb with layers intact...as soon as possible.

Thanks you very much.

cheers,

Marcel.]]> Wed, 15 May 2013 04:56:33 +0000 http://forums.cgsociety.org/showthread.php?t=1107073 Posted By: Jdunay
Post Time: 05-14-2013 at 04:00 PM
Text:

Hey Guys,

I'm trying to do a bit of compositing in aftereffects and i've run into a slight snag with my velocity pass.

I'm using maya and vray and i have added in the velocity render layer and when I render a single frame i can view my velocity pass and every looks as expected. So i did a batch render on a bunch of frames, all saved as multi channel .exr.

When i get it into aftereffects however i use the extractor filter to try and pull out my velocity pass I am given velocity.x "".y "".z and then i assign these to my rgb all i get is black frames.

anyone have any advice?]]> Tue, 14 May 2013 11:00:29 +0000 http://forums.cgsociety.org/showthread.php?t=1107023 Posted By: Frun2
Post Time: 05-14-2013 at 08:47 AM
Text:

Hello.

I´m trying to create a cube map to vertical cross in 3DMAX. In vray is really easy, in camera type i only need to put Box. The problem is i need to do the same in 3dstudio Engine.

Do you know any way for do it? or some plugis or Scritps?

The panorama export and the method create 6 planes with Mat reflect and refract, not serve me because I need to take out a render Element.

Thanks.

http://www.marmoset.co/wp-content/uploads/figure1.jpg]]> Tue, 14 May 2013 03:47:32 +0000 http://forums.cgsociety.org/showthread.php?t=1106398 Posted By: mrdannhrabar
Post Time: 05-10-2013 at 12:03 AM
Text:

hi everyone, i am quite new to these forums, and i am not sure if this is the right place to post a WIP for feedback, since this is the only VFX part of the forums that i found.

I am a year 1 vfx student, and i wanted to post one of my assignments on the forums for some feedback on my composition skills (mainly my keying skills and on edges of the alpha and matte and any artifacting), its a short clip that was provided (about 2 seconds) for us to work on.

http://vimeo.com/65704345

thank you]]> Thu, 09 May 2013 19:03:55 +0000 http://forums.cgsociety.org/showthread.php?t=1106360 Posted By: wallacemarino
Post Time: 05-09-2013 at 07:21 PM
Text:

Hi guys, I'm new here.

I've been knocking my head against this all day and have been getting nowhere. Our regular post guy is tied up on another project and it has fallen to me to key some shots for a project with a very tight turnaround.

I'm working in AE CS6 with the native keying plugins.

I have several GS scenes featuring a blonde actress, all shot on a crop sensor DSLR and at least a stop under-exposed. The issue is that there doesn't seem to be sufficient definition between the screen and the hair tones to produce a clean key no matter what I try, especially as her hair is a little messy with clumps and strands sticking out from the main body and I can't roto out the offending areas either because her hair falls in front of her shoulders so I would end up with random strands appearing from nowhere on her body.



I've tried pushing the exposure, saturation, contrast etc to try and achieve more definition and every combination of tools I can think of, but it just ain't working and I have to choke the matte down to toonish levels before it even begins to address the noise.

Can any of you people offer any suggestions? I've got a day or so to turnaround 9 of these shots so a reshoot is out of the question, any advice you can offer would be much appreciated.

Cheers]]> Thu, 09 May 2013 14:21:40 +0000 http://forums.cgsociety.org/showthread.php?t=1105085 Posted By: welou
Post Time: 05-01-2013 at 07:48 AM
Text:

Hello,

I use nuke for compositing, and use open exr as a format.
I was wondering what format should I export from nuke to avid media composer 5?
I could keep going with open exr but I've heard avid doesn't read it and that it was longer to render in EXR from nuke. I'm in 1080p.
I've been looking here and on google to see what are the common workflows but I didn't find much.

Thank you very much!]]> Wed, 01 May 2013 02:48:02 +0000 http://forums.cgsociety.org/showthread.php?t=1104814 Posted By: DoctorBeat
Post Time: 04-29-2013 at 02:24 PM
Text:

hello fellas!

I have a non-technical question that should be easy to answer by someone who has some experience in compositing:
I have this sequence, 1300 frames long, with some objects (crane, and lights) that are distant from the background (high disparity).
I've already rendered all the beauty frames, and used contribution maps to separate the background, crane, and lights, into separate layers.

Since I'll be compositing in nuke, and adding motion blur and depth of field in post, these two specific passes, should I render with all objects in the scene(thus outputing the whole scene into one EXR with the two passes), or should a render in separate layers for the engine crane, the lights, and the background (so each layer would have the two passes for the objects contained in it)?

What problems should I expect when compositing, if I opt for the first or the second method?

Thanks in advance

]]> Mon, 29 Apr 2013 09:24:49 +0000 http://forums.cgsociety.org/showthread.php?t=1104633 Posted By: Byteman3D
Post Time: 04-27-2013 at 09:31 AM
Text:

I've been rendering scenes to PNG for sometime because it is small in size and supports transparency. I opened one of these PNG files in Photoshop and background areas were transparent (checkered) as expected but I couldn't, in any ways, see the rendered sky background.

1. Where is the setting in Photoshop that displays the background area?

2. If I display it and make corrections to the background, how can I save the file again with transparency?]]> Sat, 27 Apr 2013 04:31:10 +0000 http://forums.cgsociety.org/showthread.php?t=1104235 Posted By: sten2020
Post Time: 04-24-2013 at 11:15 AM
Text:

Hi, I am trying to add a 3D CG Prosthetic on A Guy's Forehead (a Metal Horn Sticking out). How do I go about doing it using MochaPro, C4D and AE?? Please Help! Thanks! ]]> Wed, 24 Apr 2013 06:15:42 +0000