Skip to main content
Participant
January 11, 2022
Question

Workable GPU for RAM previewing Trapcode Tao + Shine + alpha layer iterations?

  • January 11, 2022
  • 1 reply
  • 169 views
Okay, so I've been using After Effects with the Trapcode suite for several years now, and I need RAM previewing to be faster. The videos I make generally involve animating visuals to music in Trapcode Tao, applying Trapcode Shine and other video effects, and layering precomps on top of precomps. This, as you can imagine, is rather resource-intensive, but my computer should be able to handle it - I have an overclocked Ryzen 5950x with a 420mm AIO, 2TB Sabrent Rocket 4.0 NVMe SSD (not the OS SSD), 128 GB of 3600 Mhz RAM, and an Nvidia GTX 1660 Super.
 
I understand that AE runs primarily on the CPU, and for previewing video everything works just fine - but the Trapcode suite runs on the GPU. I need to know if upgrading to a workstation GPU (such as an Nvidia A4000/6000) or a gaming GPU (such as the RTX 3080) would increase the speed at which the RAM preview renders with Trapcode Tao - so that it loads 60 FPS or faster.
 
I've done a whole lot of optimization, but when I've got a precomp that's 8 precomp's deep, work slows to a crawl and making even slight changes becomes a hassle. Previewing 5 seconds of content takes 3-5 minutes. I have to render the whole thing in segments because of how heavy it gets, and while that has sped up my production time significantly, at the moment it takes 25 hours to produce a single minute of content. The video I'm working on now has taken me 60 already. I need this to be faster if I wanna continue making these. Any help/advice y'all can offer would be greatly appreciated!
This topic has been closed for replies.

1 reply

Mylenium
Legend
January 11, 2022

Long and short: No. That's not how TAO or MIR or Element 3D for that matter work and even if they did, there's a whole slew of AE around it that gets in the way. Point in case: At the end of the day, all getting a faster card would do is accelerate the actual render calls, but in no way would it speed up anything like e.g. TAO requesting the position of a motion path when extruding along it or similar stuff.

 

Mylenium