Integrating 3D models with photography
Interested in integrating your 3D work with the real world? This might help
# 1 17-12-2009 , 06:06 PM
ben hobden's Avatar
Subscriber
Join Date: Jul 2008
Location: Chelmsford, Essex, UK
Posts: 1,162

positioning of additional cameras.

I thought this seemed the most suitable place for this question.


I have had quite an ambitious idea, which im sure has been done many times before but im not sure quite how to put it into practice. What I want to do is create three more render cams in my scene (dino_chick(wip thread)) so that what I currently have only makes one quarter of the final image.

What im not sure of is how to place my camera so that it lines up with the position of my existing render cam. say i have my camera set up for the bottom left quarter of the scene. I thought I could just duplicate the camera and translate it up in Y and fine tune it by sight. But as I translate the duplicate cam up, its angle of view rotates with it. I tried changing the local and world pivot values in the cams attribute editor but this didnt make any difference. Maybe it could be done with the help of locators?


Has anyone got any ideas of a way this could be done accurately.

hopefully my crude idea illustration gets accross the idea...

Attached Thumbnails
# 2 17-12-2009 , 07:22 PM
G-Man's Avatar
Subscriber
Join Date: May 2004
Location: Bowling Green. Kentucky U.S.A.
Posts: 810
wouldnt changing the focal length of the camera and the fov of the camera that you're rendering in accomplish the same thing?

You're basically just wanting the rendered area to be larger right?.. more, image without zooming in or out?...

g-man


Follow My Business
On The Web!
Or
On Facebook!
# 3 17-12-2009 , 07:52 PM
EduSciVis-er
Join Date: Dec 2005
Location: Toronto
Posts: 3,374
It sounds like he wants to create tiled cameras... like each camera captures exactly a quadrant of the final image, correct?

Not sure how to execute that... you could try creating a simple box in front of each camera that would frame the view (camera parented to the box), then use the align tool to snap the boxes to each other. That seems a clumsy solution, but might get you somewhere.

# 4 17-12-2009 , 08:41 PM
ben hobden's Avatar
Subscriber
Join Date: Jul 2008
Location: Chelmsford, Essex, UK
Posts: 1,162

Originally posted by legendofzombi
wouldnt changing the focal length of the camera and the fov of the camera that you're rendering in accomplish the same thing?

You're basically just wanting the rendered area to be larger right?.. more, image without zooming in or out?...

g-man

Yes, but the idea being that If i can break the whole image into smaller chunks it might enable me to put more in...

My efforts so far, I got this by just adjusting the film offset of each camera. Starting from the bottom left I adjusted the film offset 1 up for the top left, across 1.4 for the right hand side pictures. BUt somethings obviously amiss somewhere

Attached Thumbnails
# 5 17-12-2009 , 08:51 PM
ben hobden's Avatar
Subscriber
Join Date: Jul 2008
Location: Chelmsford, Essex, UK
Posts: 1,162

Originally posted by stwert
It sounds like he wants to create tiled cameras... like each camera captures exactly a quadrant of the final image, correct?

Not sure how to execute that... you could try creating a simple box in front of each camera that would frame the view (camera parented to the box), then use the align tool to snap the boxes to each other. That seems a clumsy solution, but might get you somewhere.


Yeah, thats exactly what im thinking.


With this way though im thinking its still gonna have problems when you line up the resolution gate?

I guess it would be easier to line things up in an orthographic view, but grouping everything in the scene to rotate it all and get a better view would surely cause problems too. hmmm

# 6 17-12-2009 , 10:20 PM
EduSciVis-er
Join Date: Dec 2005
Location: Toronto
Posts: 3,374
Is the reason for you to break up the render so you don't run out of memory? I'm sure setting up a number of passes must be an easier solution, right?

# 7 17-12-2009 , 11:28 PM
ben hobden's Avatar
Subscriber
Join Date: Jul 2008
Location: Chelmsford, Essex, UK
Posts: 1,162
basically yeah. I guess I should look into the passes way before going to a bigger scale. I'll do a google search and see if i can get some useful info

# 8 17-12-2009 , 11:47 PM
honestdom's Avatar
The Nurb Herd
Join Date: Oct 2007
Location: London
Posts: 2,381
this is crazy... in render settings you have image size, um wouldn't it just be easier to make that bigger and position your camera?

what version of maya do you use? render layers are slightly different after 2008.
you should learn about render layers if its a memory issue.

# 9 17-12-2009 , 11:59 PM
ben hobden's Avatar
Subscriber
Join Date: Jul 2008
Location: Chelmsford, Essex, UK
Posts: 1,162
yeah, that would be easier, but if i have loads of objects in the scene, then mayas just gonna crash when i render. so i thought if i break the render up, if it was possible, then it would be a quarter of the work but four times over.

Im going to try making the image bigger and try hiding some stuff on certain passes and other bits on others. All the information i find to do with compositing layers though is about specular pass, occlusion etc. ive never done anything like this so i dont really know how to go about it


i do use 2008


Last edited by ben hobden; 18-12-2009 at 12:07 AM.
# 10 18-12-2009 , 05:33 PM
honestdom's Avatar
The Nurb Herd
Join Date: Oct 2007
Location: London
Posts: 2,381
render passes are great but you will have to have an understanding of compositing to get the most of them.

# 11 18-12-2009 , 06:16 PM
G-Man's Avatar
Subscriber
Join Date: May 2004
Location: Bowling Green. Kentucky U.S.A.
Posts: 810
....I've personally never had Maya crash on render due to poly count, or object number....
If Maya is going to crash for those reasons( In My experience ) It will happen before you even try to render, while your attempting to model.
Now, high poly and object count, will make a render slower then molasses on Christmas in Antarctica..

G-man


Follow My Business
On The Web!
Or
On Facebook!
# 12 18-12-2009 , 06:18 PM
EduSciVis-er
Join Date: Dec 2005
Location: Toronto
Posts: 3,374

Originally posted by legendofzombi
Now, high poly and object count, will make a render slower then molasses on Christmas in Antarctica..

Christmas in Antarctica would be the warmest part of the year, right Chirone? :p

.... but point noted.

# 13 18-12-2009 , 10:44 PM
ben hobden's Avatar
Subscriber
Join Date: Jul 2008
Location: Chelmsford, Essex, UK
Posts: 1,162

Originally posted by hammer.horror
render passes are great but you will have to have an understanding of compositing to get the most of them.


Just to be straight on this. do you mean render passes as in spec/shadow/reflection/ambient etc, or passes as in everything rendered at once just at , well, actually i guess a combination of both would work wouldnt it.


as far as compositing. I only use photoshop. No toxic or any of those other plug "channel 'x' into channel'y' " type ones. Im happy playing about with blending modes and opacity levels and whatnot. My main idea I guess is to just try and get a bunch of renders together that i can clone tool out of one and into another, but I doubt thats what you mean by compositing skills!

I watched a digital tutors render layers tut last night,, but it was maya 2009, using tokens, and contribution maps (which looked like just what i might need), and render 'passes' of seperate render 'layers'. But as ive got 2008 it wasnt really that helpful.

Posting Rules Forum Rules
You may not post new threads | You may not post replies | You may not post attachments | You may not edit your posts | BB code is On | Smilies are On | [IMG] code is On | HTML code is Off

Similar Threads