Full program multi-threaded support
After Effects would benefit greatly from being able to actually utilize all cores when rendering, rather than having to rely on a third party solution, like the excellent program Render Garden by Mekajiki.
(This post was co-authored by Tim Kurkoski (After Effects Product Owner), Andrew Cheyne (After Effects Engineering Manager), and Victoria Nece (After Effects Product Manager).)
This thread has been sitting here for quite a while without a response. To start, we want to apologize for that. We haven’t been ignoring you or your feedback – this is just a particularly complex topic. That said, it’s time for us to check in with you, clarify a few things, and give you an idea of what we’re looking at for the future of After Effects, especially when it comes to performance.
Before jumping into the specific request here (multi-threading AE), we should talk about how the AE team looks at performance in general. There are three areas:
- Rendering Performance: How fast can AE get pixels onto your screen?
- Interactive Performance: How fast does the UI respond to your actions?
- Workflow Performance: How fast can you complete a task in AE?
All three are important. All three impact how quickly you can make creative decisions and get your work done. This request and this discussion are focused on the first area, rendering, so we will focus on that, however we don’t want to lose sight of the bigger picture.
What have as the AE Team been doing? When we set out to tackle rendering performance, we looked at the potential improvement offered by different technical paths. And we set a high bar: we didn’t want to just have an incremental speed increase. We wanted a major leap forward in rendering power.
The obvious technology that would enable us to achieve that goal was leveraging the GPU. Because GPU processing power has leapfrogged the CPU — and is explicitly designed to handle this type of processing — the decision was made to invest in getting AE’s core rendering pipeline running on the GPU.
This is not a small task, and we’re not done yet. The most visible result of this effort has been the porting of individual effects to the GPU – more than 40 so far. Less visible is the work we’ve done getting the rest of the After Effects rendering pipeline on the GPU, such as layer transforms and motion blur or debayering RED raw R3D footage.
We still have further to go, as the real power of the GPU is unlocked when you don’t have to send a frame back and forth between CPU and GPU for different stages of rendering. As more links of the GPU chain come online, you can expect further rendering performance gains.
We do know what you’re thinking at this point. You want to know what we’re doing about multi-threading.
We all recognize that After Effects would benefit from additional general-purpose multi-threaded rendering. And while we can’t get into specifics here or make any promises about our future roadmap (that’s all confidential when you’re a publicly traded company), we are actively working on multi-threaded CPU rendering.
Ultimately, when it comes to achieving the best rendering speeds possible for AE, we know we’ll need a combination of GPU and CPU processing that maxes out all the resources on your machine.
We recently partnered with the folks at School of Motion and they did a fantastic video on how to get a more optimized system for After Effects: https://www.schoolofmotion.com/tutorials/after-effects-computer
Please keep posting and letting us know what you think. We’ll continue to watch this thread (and all the others on UserVoice). And we appreciate your patience with our response to this post.
Peter Labrow commented
Everyone calm down. If Adobe's track record on this thread is anything to go by, you're in for a wait of several years. If they cared, it would be fixed already.
“Please keep posting and letting us know what you think”
Dear developers from Adobe. We are all tired of waiting. There is no real alternative for AE, you know that and you take advantage from it. AE exists because it has no real competition not because it is great tool. If you listened to the people you could make it great. Years are passing and AE feels the same like 10 years ago. Without plugins it would be dead. 3d tools are joke. Have a look at Blender development where its developers listen to the people, talk to them, make promises, publish roadmaps and improve the app. They are doing this with much less budget, super stars and useless shows. People who pay you want a whole new "SUPER CAR" not just a one single shiny piece of it every half a year.
Patrick Proier commented
Don't try to translate the 30 year old core of AE 1:1 into the new GPU/ and Multi CPU world of 2019!.
Go with your partners from C4D back to the drawing boards and design the old core totally new.
After Effects leaks not only at GPU and Multicore performance it needs also a total new and simple 3D Engine. A kind of native (and better) Elements 3D with basic features to import and manipulate simple 3D data. (geometry, projection mapping, depth data, lights, particle, pointclouds, and camera/tracking data)
Also, a solid and modern color management (native ACES support) and new workflows to deal with 3D passes like in OpenExr files is needed.
the old core of AE is simple not made for Motion Graphic artist and VFX artists of the feature!
Saar Oz commented
“Please keep posting and letting us know what you think”
Dudes, this is what we have been asking you to do for years
Hi if Render Garden can make a program for using all cpu’s at render, it could done this Adobe. If you cannot do this properly stick your head with render garden.
@Victoria, Tim and Andrew
Thanks for giving us an update on this. It comes appreciated, and I am sure that making these sort of updates to such a long standing product must be rather challenging.
I appreciate there are many product features and use cases to focus on, but when using AE from a UX perspective, the bitmap and 3D effects are of limited use. If it would be possible to look into improving the performance of shape layers and expressions, productivity would benefit significantly.
There is so much potential when it comes to expressions in After Effects, especially when working with master properties. With better performance, AE could be much more valuable as a tool for designing and building software and interactive experiences.
THANK YOU for letting us know that our concerns have been heard and are being addressed. I really appreciate the details about critical paths, options, strategies etc. :)
Thanks for communicating with us Victoria, Tim and Andrew.
The performance steps you've taken and the need to better leverage the GPU were necessary and welcome.
The problem is that these three areas of performance cannot be weighted equally.
You've invested an incredible amount of time into focused performance features that have little, incremental ROI. This comes at the cost of immediate overall performance needs, in which you've capped user ability to increase performance through hardware investment. As the last five years has shown, your road-map fundamentally ignores the requirements & workflow of your primary market and remains overly optimistic on the timeline in which GPU's can reliably supersede CPU's in a production environment.
We understand a road-map can be construed as a forward looking statement to investors, which becomes problematic if you don't deliver on promises.
This has in part however, resulted in a development culture at Adobe, that appears to lack a sense of urgency and ignores precedent, while undervaluing the constant deadline pressures faced by many of its customers. Despite Adobe's promising developments in visual AI, this is a symptomatic issue that should concern shareholders & investors.
I point all this out, because your response suggests that full multi-threaded CPU rendering solutions are still likely years out. Your roadmap seems casually focused on the long-term, when interim solutions for this most critical of deficiencies are needed.
Two third parties and their multi-frame render queue plug-ins have admirably delivered where your team hasn't; but as these plug-ins can't be applied to previews, they're inherently limited. Can you not address that?
I think like most established companies that have a near monopoly, high user base or veteran users, they steadily begin to ignore them, hoping name recognition will keep them on. We left AVID specifically because they were NOT innovating while still demanding proprietary hardware and expensive upgrades. It took some time to migrate off, since we had many ongoing projects, but we said good bye.
I've been(or team members) to AdobeMax for many years and each time I hear them blather on about their AI doing this or that, or their speakers get so excited about marketplaces to sell templates, or they show video of people editing on an Ipad in a park, I look to my left and right and all our eyes are rolling in unison.
I know talking tech isn't **** but that truly is Adobe's BIGGEST issue. Premiere just crashes for absolutely no reason. AE doesn't take advantage of our workstations, because they need to make it compatible with a bunch of old outdated laptops? I've gotten straighter answers out of the evangelists.
AE unfortunately doesn't have a real competitor. When it does, watch people jump ship. Adobe won't get nervous until it's too late.
Peter Labrow commented
I didn't want to leave an ungrateful combative response to Victoria's post, but I'm sorry, I'm finding it impossible. The majority of the post essentially is PR fluff telling us to be grateful. It starts by saying that Adobe hasn't been ignoring us... on a post with over 800 upvotes started.... back at the dawn of time. Victoria. I know it's rude to come back at you like this but please read the mood in the room. This is a critical pro issue. Updates to XD arrive so frequently training companies must be laughing all the way to the bank. Meanwhile lots of the suite has time spent on ... non-standard Sesame Street dialogue boxes rather than actually integrate with the OS. A 'major feature' of the last release was 'GUIDES', honestly. (Yes, content-aware fill... that works sometimes.) And don't get me started on the disparity of UI, shortcuts and terminology between 'integrated' products. Waste. Of. Time. The argument for going to subscription was it would fund development and innovation and I don't see enough of that to have warranted paying for an upgrade so the need for subscription is obvious. Meanwhile products like Resolve don't just gain features they smoke Adobe's performance and look more tempting by the day. If it wasn't for the fact that there isn't really anything quite like After Effects, I suspect many wouldn't be using it. And what kind of company is happy with unhappy customers?
Paul B commented
1. Rendering Performance: How fast can AE get pixels onto your screen?
2. Interactive Performance: How fast does the UI respond to your actions?
3. Workflow Performance: How fast can you complete a task in AE?
All three are significantly slower than they were 4 years ago. YANG YANG
That was quite the measured response, NON-ANSWER, everyone expected. Kudos. Currently still being forced to use BG Renderer MAX in order for AE to render at acceptable speeds. We still have to use a hack to get the GPU to work with RayTracing on our Quadro P4000 cards, because AE "doesn't recognize" the GPU no matter what driver is installed. Yeah OK.
How about quit putting so much effect into your apps working on an Iphone and more time into them working properly on what REAL professionals use. Because we all need to edit in a park on an IPad. Sure=) Thanks for your continued persistence and patience on this matter.
Liam Clisham commented
This isn’t an update. This is the same stuff you’ve been spouting for years now on MDA, Twitter and at NAB. I can import an AE file into PR and have it render faster and preview in real time. Why has that team done something you can’t? Why has Fusion/Resolve tackled this? Padded, corporate responses are no longer acceptable. Just because AE has no real competition doesn’t mean you get to treat the users like idiots and not fix or make progress. No one cares about new features or “40 effects on GPU” when they can’t even RAM preview ten frames on a 128 gig of RAM system. Calvary and other competition is coming and AE is going to be the continue to be the joke of the industry.
I have to agree with the general sentiment here. I might have sympathised with this response 4 years ago, but its unacceptable to remove a performance feature from 2014, and do nothing about it. I dont see why the focus has been on other performance enhancements when we are all hemorrhaging time with the most important/significant one (cpu). Its like going to fix a leaking tap when you have a burst pipe. FIX THE PIPE!
Sounds like it's finally going in the right direction. AMD's core offensive is forcing you to improve After Effects multi-core support. We use AWS Deadline 10 to render our projects with After Effects. Here we let each computer run up to 16 instances of After Effects in parallel on one machine. Why? We use an enormous number of expressions. The GPU is of no use there. Here often only 1 single core is used. Professional projects often require many expressions. Here we have up to 16x more render power with an AMD 2990WX compared to rendering a preview in the After Effects GUI while working. Furthermore we often deactivate the GPU renderer and switch to CPU. Why? With 8K RED recordings and many effects the GPU memory is often not sufficient anymore. Even on a 2080Ti After Effects crashes. So from our point of view the focus on GPU is currently only useful for smaller projects. Especially large projects with large footage suffer enormously from your decision since CC2014 not to offer multi-core support anymore for preview rendering. A short number for comparison. With 30 computers and AWS Deadline we can render 300x faster than a preview can be generated while working on the project. This is where After Effects gives away its potential and this is where we waste time on projects. With 15 years of experience as a VFX supervisor, I can only say that I'm wasting about 60% of my time waiting in After Effects (because of unused power). Here must be improvement. Many professional colleagues, especially freelancers, are already switching in some areas to software from Blackmagic and co. Make sure that your software remains attractive for real professionals. Thanks a lot!
The one thing slower than After Effects : Adobe's response time.
That was exactly the kind of feedback that will win you fans...very loyal fans!!!
what a crock of ****, victoria. switching to davinci now! thanks for solidifying my reason to leave. :)
Glad to finally hear back on this one! One thing I can offer is that while GPU rendering is fast, it is unreliable (as I'm sure you have run into). I personally work at a studio with quite a large CPU farm that cost us a ton of money to build several years ago. Due to so many AE plug-ins switching to the GPU I cannot currently render most of my comps on the farm and have to fall back to workstation rendering, which is just... heartbreaking. That said some companies have done a very, very good job of creating CPU cross-compatibility. The Foundry is an obvious one, and Chaos group has done very well with thin in VRay. So I'd just like to throw my vote in there, while GPU does speed things up please do not drop attention on CPU rendering as well (hence multithreading).
@ian. Sure like to know why 3 Mac pros I’ve worked on render the same project in 10min in the morning, but it takes an hour in the afternoon.