[Bill Lake]

Todd- Can you point me to some documentation on this? I’d like to understand the technology on this better.

My goal is often to render to final file as quickly as possible, once the comp, or a set of comps is completed. I find that if I generate a series of comps in AE (without rendering) and then drag them all to the AME queue, that I complete the entire render phase more quickly than if I first rendered to an uncompressed file in AE and then re-encode with AME. The AME final render’s always are a higher quality (2 pass encoding) and much smaller file size. Now, my comps are not very complicated, so maybe that’s the reason. Most of mine are 2d motion graphics, character animation compositions, logo openers (3d and effects) and such.

Where can I find out exactly how this works, software-wise.

-BL

[Bill Lake]

I thought I read (somewhere) that when you render a comp from AME, it called AE in the background, and AE rendered the frames, and then AME did final encoding. If AE is called, then why wouldn’t AE use it’s typical resourcing (i.e. multi-thread) to render?
-Bill

[Bill Lake]

Ae previews footage fastest when you have fast I/o (buss and disks) so put source media on raid zero. Never use caviar disks – you need fast disks with fast seek/read times.
Ae previews of effects requires fast CPU, large/balanced memory (2g per core min)
Never render from ae. Use AME.
Ame uses cuda to dramatically increase render speed. It also has better encoders. For this you must use an nvidia you. The more cores the better. This really only impacts renders.
Disk setup is critical: read source media from raid zero. Render to separate disk raid zero, apps/ O’s on dedicated drive, as well. Then read up on where to place page file, caches, etc. Hsrm Millard is an expert on this.