Blender Institute Podcast #012 - Training

This is for Hjalti. When animating, isn't annoying to have "Jump to Keyframe" and "Change Frame" operations into the Undo queue? When animating it's usual to jump between keyframes several times to see how the poses work together. Eventually you'll want to undo the last change in the character and then you have to undo all the jumps between keyframes to, finally, undo that change.

It would be great to have the option of avoiding that to be registered in the undo list.

Can you please beg for me to Doctor Sergey to implement this? Pleeeease?

As I try to learn animation the one thing that seems to be missing is an explanation of how audio fits into the process. We have nine old men with 12 principles of animation, and that covers the visuals. Phonemes, music and sound effects support the story as well, but how, and at what stage, should they be brought in?

Thanks for these Podcasts. I'm enjoying them very much and learning so much.

I was wondering if you could talk more about the workstation setup and process you use to create the tutorials. What flavor of linux are you running, what screen capture software do you use, what does the workflow look like, etc...

Thank you.

That was a hell of excuse about comments on soundcloud..
Mostly Copy/Paste from there:

I think it would be worth it to do final compositing also though, there are things you could improve just by postproduction. The result is what matters in this case and not the speed, it's Blenders marketing material, so the better the result, the more professionals will be interested.

Postproduction motion blur and depth of field is much, much better choice, it gives flexibility and in many cases eliminates the need to rerender which can be expensive if you are not sponsored by render.st. In every day work speed is crucial. And talking about that and workflow with other softwares, can you please fix vector blur pass to the industry standard? To use RSMB for example. Yes, there are nodes to convert, but that is tricky and adds more complexity to the process. The same with normals pass, they are different in all other applications. And I don't want that because others do it, it would just be better for professionals who will never use just Blender and this makes their lives difficult. I remember someone talked about this in Blender conference a while back.

Another thing.. why is it not possible to render only a pass without rendering final result? For example if a mistake is made and you need to rerender ID pass or AO pass or any other pass, I don't need the final result rerendered. The only way is to change node setup by replacing comp output to be one of those passes and then renaming output file, hence more mistakes are made in the process and more time spent. Changing node setup can be very tricky..

Another timesaver would be to automatically name outputs by it's pass. For example if main comp output is "Final", then pass outputs would be named "Final_AO", "Final_Obj_ID", etc. And if render layer is set as output, filename should be the same as render layers name. Automatically organising output would be super nice and of course it takes tome to develop things like this, but it's not like depthgraph or hair system, no super deep research is needed, it just takes a little bit of logic mind. :)

Can I please have my answer now, thank you very much.