Million Dollar Idea

[alexyork]

It's all about cost though, really. You could pay someone £200, for example, to optimise your scene in a day so it renders in 1hr instead of 2hrs. But if you're paying for a farm that extra hour might cost you less than £200, so it was a pretty much pointless expense. If you are on a really tight deadline this sort of thing could be a good idea. Then again, it might be cheaper in the medium>long term to invest in more render nodes, which might work out cheaper than constantly paying someone to optimise your scenes. Personally, I will spend a maximum of 30 mins optimising scenes for still renders, then just throw nodes at it. I try to aim for a bench of around 2hrs for a full 5k render at very high settings, distributed across my nodes. That way I know I can get a good full set of stills out for a job overnight, which is fine. If I paid someone to optimise those scenes I might have to pay them £200 or so to do that, or I could just throw more nodes at the job and it would have paid for itself after a few jobs.

Animations are a different matter. £200 to optimise a few scenes for render farm use could save you hundreds or even thousands if the optimisations are significant, since you're paying for CPU time.

I guess it all depends on:

1) if you really don't know how to optimise your scenes, at least to a basic level (which I argue we should all)
2) how much this service would cost
3) how long it would take (tied to above of course)

[glorybound]

I can imagine it being a difficult scene, or for newbies. After a few times, if the person shared what they found, you would learn to do it yourself. So, it wouldn't be a very sustainable business model. Some studios have lighting guys, modeling guys, post guys, and texturing guys. I am wondering if they also have optimizing guys.

[Dariusz Makowski (Dadal)]

Not really an optimizing guy... Its just a guy that does lighting and rendering that takes care of optimizations. He usually should have ultimate knowledge over the shaders, lights and render settings to make it render faster and properly. If I could do it as a freelancer and earn 400£ each week then I could see it as sustainable job. Since optimization should take only around 2-6h or so, depending on scene.

[glorybound]

I placed a lot if trust in SolidRocks, and got really lazy, which is hurting me in the end. So, from now on, I'll have to know each scene intimately, so I know what's going on.

[Neilg]

Might be worth making some kind of "things we know" sticky thread with a run down of what causes major slow downs in render times, a short explanation of why they slow things down and how to fix them (or in the cause of unavoidable slow downs, compromises that can be used for speed.).

This will be good. A big sticky at the top, with someone holding the first post who can go back and edit in the best tips posted in the thread. I found someone in our office that had never heard of turning filtering off on opacity mapped leaves - and it's not exactly something you'll discover by accident.

[glorybound]

I have seen lists, like this, on other sites on the web. I personally like the idea of scripts like, the Teague Scene Checker. A list of the great optimizing script would be good, too. The Rebus plugin does some nice troubleshooting, as well.

[Neilg]

They cant look through your scene for materials applied to fabrics and turn the reflection depth down to one though - most optimizing cant be done automatically.

[glorybound]

no, but it can tell you when something seems out of whack.

They cant look through your scene for materials applied to fabrics and turn the reflection depth down to one though - most optimizing cant be done automatically.