Mac studio finally in my hands. Testing surprise

Flame Hardware Questions
, , , ,
21

I was testing it by creating a coloured frame at 4096x2160 then putting that through a gmask or empty action (the bug doesn’t happen if you connect the blur directly to a coloured frame or through a colour corrector). Then when you add the defocus blur at 50 or 80 you get immediate slow downs.

So using a resize node to pad the image slightly larger seems to do the trick.

I’ve tested on Linux and it doesn’t have this issue.

Would be good to know if this bug is happening in 2026 flame still if anyone can test.

22

I just tested again in 2026 and the bug (if its a bug or a mathematical problem).

My Tests:
UHD noise into blur defocus 20, 100f = 24 seconds
UHD noise into blur defocus 21, 100f = 7 seconds

UHD noise into blur defocus 30, 100f = 44 seconds
UHD noise into blur defocus 31, 100f = 15 seconds

UHD noise into blur defocus 50, 100f = 12 seconds
UHD noise into blur defocus 51, 100f = 12 seconds

UHD noise into blur defocus 60, 100f = 13 seconds
UHD noise into blur defocus 61, 100f = 50 seconds

UHD noise into blur defocus 70, 100f = 11 seconds
UHD noise into blur defocus 71, 100f = 18 seconds

UHD noise into blur defocus 80, 100f = 21 seconds
UHD noise into blur defocus 81, 100f = 10 seconds

UHD noise into blur defocus 90, 100f = 13 seconds
UHD noise into blur defocus 91, 100f = 21 seconds

UHD noise into blur defocus 100, 100f = 12 seconds
UHD noise into blur defocus 101, 100f = 14 seconds

So what I’m taking away from these tests is that if I have a comp with lots of blur nodes and the renders feel slow, maybe tweak the numbers by 1.

3 Likes
23

Yes it seems as though each resolution will have an issue with some value.

I just checked on 1920x1080 and dragged the defocus amount up until it started to pause, which for me was around 45.

For another comparison Sapphire RackDefocus works fine apparently at any value, of course that’s another cost.

1 Like
24

Here’s a video demonstration I sent to support.