planning workflow for short film

# 1 15-04-2012 , 03:11 AM

Registered User
Join Date: Jun 2008

Location: Los Angeles, CA

Posts: 229

I've got down the basics for modeling, texturing, rigging, and animating. The director has an amazing eye for shots and expressiveness; I will simply be his hands for the most part.

What I don't know and can't seem to find is the best (word used loosely here, there are always many possible solutions to a problem) method for setting the camera(s?) and what-not. Since I don't know what ?s to ask I'll give examples.

In real life, if two characters are talking, you'd stage the scene multiple times, and depending on how you like to do it, shoot the whole scene from each angle or shoot segments from each angle. Both are of course legitimate methods. So in a simple scene, you might have a 2-shot angle and two close-up angles. If you only have one camera, you shoot the scene a minimum (assuming perfect performance) of 3 times.

In 3d animation, do you create 3 cameras, position them and animate them according to storyboards, and then render the whole scene out three times (or depending on hardware, scene complexity, and directoral style, render only portions), do you create a new scenefile for each shot, or something that I haven't mentioned? i.e. what are the more common workflows for getting footage to the editor?

Thanks!

--
Peter Srinivasan
Producer

Quote

# 2 15-04-2012 , 03:26 AM

EduSciVis-er
Join Date: Dec 2005

Location: Toronto

Posts: 3,374

Create as many cameras as you want in the same scene, then just render out say frames 1-500 in one camera, 501-1000 in the next one etc. Or render out the whole animation in all three cameras (which takes longer) but then you have ultimate freedom in cutting it together in your video editor.

-stwert

Website - Blog - YouTube - Vimeo - Facebook - Twitter

-->How to Effectively Ask Questions on Forums<--

Quote

# 3 15-04-2012 , 03:31 AM

Registered User
Join Date: Jun 2008

Location: Los Angeles, CA

Posts: 229

I suspected that's how it's usually done, but you never know. There are always little things that make huge differences. For instance, "cheating" on set in film means to move something to a new place after shooting it from one angle to accomodate a new angle, and doesn't sound like much of an option when rendering from within one scene, so for cheating, if necessary, it may be useful to duplicate the scene and build a "cheated" version? But then if you decide you don't like some bit of animation in the non-cheated, you would have to update the cheated one as well. Any hints on that note?

--
Peter Srinivasan
Producer

Quote

# 4 15-04-2012 , 04:06 AM

Technical Director
Join Date: Feb 2004

Posts: 2,988

What usually happens in production, is you create a scene per shot, then reference in the objects, rigs, etc. This means that the scene stays clean, and you can still work on assets after the fact. They'll just update in all of the animation files automatically.

It's also very common to render "handles" to the animation. Usually about 10 frames at the beginning of the shot, and 10 at the end. This way, the editor has a bit more room to slide the shots around. However, the edit is usually pretty close to final, even before animation gets to it.

Imagination is more important than knowledge.

Quote

# 5 15-04-2012 , 04:18 AM

Registered User
Join Date: Jun 2008

Location: Los Angeles, CA

Posts: 229

interesting. I'll have to do some research into referencing. I assume if I keyframe a referenced control and then change something in the asset file (as long as I don't delete the control or something drastic like that) the keys will stay the same in the scene file? I think I like the method of building a scene per shot, since I can do things like optimize be removing parts of the scene if I need to... last project I did my computer couldn't handle rendering two large fur_feedback nodes at one time so I had to split it up and composite. Handles make sense, of course. I always render handles when I'm doing after effects work for someone anyway, so that seems natural. Thanks all!

--
Peter Srinivasan
Producer

Quote

# 6 15-04-2012 , 09:24 AM

Technical Director
Join Date: Feb 2004

Posts: 2,988

interesting. I'll have to do some research into referencing. I assume if I keyframe a referenced control and then change something in the asset file (as long as I don't delete the control or something drastic like that) the keys will stay the same in the scene file?

Originally Posted by petersrin

That's correct. Referencing also imposes some rules on what you can, and can't do to referenced objects in the scene, such as change the UV coords; which is helpful as it keeps all assets the same across shots. If you need to make a shot-specific change, you can unreference the object, or create a new asset file with the changes in them.

Note: I use the term asset here as a distinct object, not Maya's built in asset workflow.

Imagination is more important than knowledge.

Quote

# 7 15-04-2012 , 10:19 AM

Lead Modeler - Framestore
Join Date: Feb 2003

Location: UK

Posts: 6,287

Interesting. Ive been looking at animating cameras and stuff recently as Im doing a personal short for myself and also helping my daughter with a VFX project too, its still all voodoo to me in some respects (being a modeler) but have had good advice from the Lead Animator at work.

Theres some good ways to do camera cuts, you can use multiple cameras, or you can key the camera to cut to save on cameras....I actually found a fairly useful 'tute' on youtube. The guys voice just grated but I got what he was doing and it kind of simplified the process. Nice to do some grass roots stuff for a change, well for me anyway.

Heres the link: Maya: Animating a camera - YouTube

If theres one good thing thats made my animation skills a little better in the last week, that would be turn off Auto Key every time you set a frame, the lead at work and the guy on this video says it too, it keeps everything in a nice order and the scene is cleaner and the animation is easier to handle. And dont forget your ease in and ease out, to keep things smooth when stopping the camera, select the curves and apply a flat tangent to them....

Edit: As Next Design says as well, use referencing, it can save heaps of time and keep the scenes to a minimum too -better for you in the long run. Maya's referencing has taken a critical beating over the years as ithas always been a bit clunky and non-responsive, but as of 2011 and 2012 its alot better by far, we used it on LockOut for everything and it was really stable, depsite fears of using it. Also its nice to note, that Maya 2012 has the replace object command in there now, something that Lightwave has had for years, you can replace objects in the scene at their local position either as pure GEO or an instance or either a reference object...and it actually works well too, so more for you to have a think about....hope that helps

cheers
Jay

Quote

# 8 15-04-2012 , 04:27 PM

Registered User
Join Date: Jun 2008

Location: Los Angeles, CA

Posts: 229

Hmmm, @Jay doing it all in one scene seems like it'd be in some ways a nice luxury, but I'd imagine adding any kind of compositing beyond passes might be difficult. Definitely autokey off though. I know I've run into trouble w that before. Also, when manipulating the rendercam from a lookthrough viewport, undoable camera is a must for me... for those times when instinct takes over and I swivel the rendercam to look at the other side of a mesh user added image

--
Peter Srinivasan
Producer

Quote

# 9 15-04-2012 , 04:34 PM

Subscriber
Join Date: Oct 2011

Posts: 139

usually you can cheat.

Each scene has a set camera angle (same as in storyboard/leica) and you may only animate whatever is in the field guide. ..e.g. for a talking scene you can only animate the upper body of the character and don't have to worry about feet placement or anything like that, you can still cheat shifting weight and moving the body up and down etc.

I would always animate scenes relatively to the cameras instead of animating in the default view and then setting cameras around it, because you get a much better end result that way and you see in real time how it will look on screen.

Quote

# 10 15-04-2012 , 05:17 PM

The Nurb Herd
Join Date: Oct 2007

Location: London

Posts: 2,381

am i missing something? There seem to be a lot of people lately who are very confident about doing their own full cg shorts. I going to have to be the pessimist here, but i don't see these succeeding any time soon.
I just spent 2 months working 15hour days to do a full cg 30sec commercial. there were 4-5 animators, 3 lighters, 4 compositors, 2 FX TDs and we have a large render farm.

i would say just do one camera per scene (baked and locked) and geo cache your anim. it'll make things much cleaner/lighter when rendering.

Quote

# 11 15-04-2012 , 05:20 PM

Lead Modeler - Framestore
Join Date: Feb 2003

Location: UK

Posts: 6,287

Yep Pessimist.....LOL....Seriously though, it depends on the subject matter for a short, style and so on, for myself all I'm doing is a few fly by shots of a P51 mustang, pretty simple really, I certainly dont expect to be creating the next pixar short.....

A friend of mine works freelance full time from home creating cg adverts and various intros for games etc, he manages fine, he works hard, but he does it an reaps the rewards too. Plenty of people do shorts, some good some not so....Ive seen plenty of good ones in my time.

Yep its all about cheating believe me, also comping doesnt matter, its how you choose to render it out. If you just had one camera for the whole scene, you render it per shot basis, so from 1 - 50 frames of the first shot, then 51 to 200 for your second shot, keep it all separate...easy

Horses for courses though

Jay

Quote

# 12 15-04-2012 , 05:26 PM

Registered User
Join Date: Jun 2008

Location: Los Angeles, CA

Posts: 229

oh man. too many options! I'll just have to choose one. This piece is kind of like the culmination of all my dabbling years. So excited user added image

I'll probably start a wip once I have something to show. Thanks all!

--
Peter Srinivasan
Producer

Quote

# 13 15-04-2012 , 05:31 PM

Registered User
Join Date: Jun 2008

Location: Los Angeles, CA

Posts: 229

am i missing something? There seem to be a lot of people lately who are very confident about doing their own full cg shorts. I going to have to be the pessimist here, but i don't see these succeeding any time soon.
I just spent 2 months working 15hour days to do a full cg 30sec commercial. there were 4-5 animators, 3 lighters, 4 compositors, 2 FX TDs and we have a large render farm.

Originally Posted by honestdom

I am confident about this one thing: I have to make this video, so I will user added image

I know that there will be a myriad of problems that would've been avoided if I was in a larger production environment, just like the issues that cropped up (and even showed up) when I did APRIL FOOLS: SEGA Bass Fishing of the Dead - YouTube for April Fools, which took about a month to complete those few CG shots. But it is a great way to learn, as I don't have the creds to work IN a production pipeline. Next best thing is to screw yourself over by forcing an impossible project.

Edit: That youtube embed ended up WAY bigger than I expected :< in fact, I didn't even want an embed. sorry

--
Peter Srinivasan
Producer

Quote

planning workflow for short film

Posting Rules Forum Rules

Similar Threads

Topics

Free Courses

Full Courses

VFX News