Improve animation workflow speed for large images
Posted: 12 Feb 2018, 10:15
Hey 
I know it's a difficult thing to ask, cause every soft has it's limits regarding performances (so I'm talking about some solutions but please forgive me if i'm completely wrong about it haha). But since i'm on a production with a lot of camera movements, I often animate on shots with lots of pixels (even with a 1080p camera the canvas can often be more than 3 times this size). And when I do so, TVPaint handles perfectly when i'm drawing but not when i'm working on the movement :
- Flipping beetween images is not really smooth (even just going from one instance to another, being too slow on flipping you can't ready "see" your animation move to adjust it). It's problematic cause I'm not flipping anymore, i'm kind of animating pose to pose with my lighttable guessing what it will become while rendering. Could be nice to have some memory dump on a small amount of instances around the one you're working on to be able to flip more easily. Sometimes I animate on the spot with a moving background, makes a smaller canvas, but you can't do this with every shot.
- It can become even slower when using lighttable and out of peg (and flipping out of peg is even slower, with OOP scripts and even more with the flip panel). So I can't really feel the movement on my animation here neither. Because of this I can really feel i'm much slower while animating (I think it takes me 1.5x the time it would take if I had a smooth experience cause i'm waiting at each flip, and the result could be better because of the fast visualisation)
- Even when calculating the shot befor playback, TVPaint is unable to read all the images. I need to export the shot on my computer in order to read it (on this one I think that a hard render on the software can be managable right?)
As I said I understand that it can't be solved easily but for me it's a huge drawback right now. And having this solved could also be a huge improvement for storyboarding in tvpaint, being able to work with more margins could be great. Idea : Maybe some evolutive margins, with the canvas resized each time you stroke ou a camera comes out of the actual canvas? Would save us the difficult part of resizing the canvas and having to save the new project created.
Am I the only one bothered by this or is there other animators that feel the same? Maybe there is some alternative workflows to avoid this?
Nathan
I know it's a difficult thing to ask, cause every soft has it's limits regarding performances (so I'm talking about some solutions but please forgive me if i'm completely wrong about it haha). But since i'm on a production with a lot of camera movements, I often animate on shots with lots of pixels (even with a 1080p camera the canvas can often be more than 3 times this size). And when I do so, TVPaint handles perfectly when i'm drawing but not when i'm working on the movement :
- Flipping beetween images is not really smooth (even just going from one instance to another, being too slow on flipping you can't ready "see" your animation move to adjust it). It's problematic cause I'm not flipping anymore, i'm kind of animating pose to pose with my lighttable guessing what it will become while rendering. Could be nice to have some memory dump on a small amount of instances around the one you're working on to be able to flip more easily. Sometimes I animate on the spot with a moving background, makes a smaller canvas, but you can't do this with every shot.
- It can become even slower when using lighttable and out of peg (and flipping out of peg is even slower, with OOP scripts and even more with the flip panel). So I can't really feel the movement on my animation here neither. Because of this I can really feel i'm much slower while animating (I think it takes me 1.5x the time it would take if I had a smooth experience cause i'm waiting at each flip, and the result could be better because of the fast visualisation)
- Even when calculating the shot befor playback, TVPaint is unable to read all the images. I need to export the shot on my computer in order to read it (on this one I think that a hard render on the software can be managable right?)
As I said I understand that it can't be solved easily but for me it's a huge drawback right now. And having this solved could also be a huge improvement for storyboarding in tvpaint, being able to work with more margins could be great. Idea : Maybe some evolutive margins, with the canvas resized each time you stroke ou a camera comes out of the actual canvas? Would save us the difficult part of resizing the canvas and having to save the new project created.
Am I the only one bothered by this or is there other animators that feel the same? Maybe there is some alternative workflows to avoid this?
Nathan