Hey folks,
to keep the fps flowing here are three tips.
- bypass every inactive node (Purple)
- use two Switch to bypass a big group of nodes. The first switch to dumps a empty texture in the sequential nodes. The second bypass the texture to the rest of you patch. (Cyan)
- you do not always need the full resolution/colour depth, to render pretty stuff (blur and sample texture) (Pink)
Have fun!
Three ways of patch optimalisation
Three ways of patch optimalisation
- Attachments
-
Optimalisation.wire
- (220.21 KiB) Downloaded 831 times
https://get-juicebar.com/creator/tweelicht #tweelicht_visuals
Re: Three ways of patch optimalisation
Hi everyone,
i made another patch to speed up render times. This one lets you downsample in 16 discrete steps.
In the patch are a few comments to get you going.
(bug?)
The step size that handles the resizing steps is now 1.175, and it also works at 1.125, other steps need to be found.
Other steps give clipped video when using effects after them, this stops when changing the bit depth, but returns when changing the size-quality.
I haven not test this patch at strange resolutions.
Downsampling also affects al the relative value, so small values at 1920px are big at 240px.
Have Fun!
i made another patch to speed up render times. This one lets you downsample in 16 discrete steps.
In the patch are a few comments to get you going.
(bug?)
The step size that handles the resizing steps is now 1.175, and it also works at 1.125, other steps need to be found.
Other steps give clipped video when using effects after them, this stops when changing the bit depth, but returns when changing the size-quality.
I haven not test this patch at strange resolutions.
Downsampling also affects al the relative value, so small values at 1920px are big at 240px.
Have Fun!
- Attachments
-
Switch.wire
- (344.01 KiB) Downloaded 427 times
https://get-juicebar.com/creator/tweelicht #tweelicht_visuals
Re: Three ways of patch optimalisation
Don't forget to use attribute/event flow whenever possible!
Re: Three ways of patch optimalisation
@Markymark, i will investigate this. There are some big draw back of attribute and event time.
But its a interesting puzzle to figure out when to use them
.
Do you know why the Patch Info outputs in signal time?
It would be very helpful to get the resolution of the patch in attribute timing, so i can generate gradients in a square and then rotate them.
But its a interesting puzzle to figure out when to use them

Do you know why the Patch Info outputs in signal time?
It would be very helpful to get the resolution of the patch in attribute timing, so i can generate gradients in a square and then rotate them.
https://get-juicebar.com/creator/tweelicht #tweelicht_visuals
Re: Three ways of patch optimalisation
We would also really like the Patch Info node to be at other flows.
There is work being done on this, but since this stuff relies of Arena <> Wire communication we have to be very careful.
All flows have their benefits and disadvantages.
I tend to go over a finished patch with a optimisation lens on.
Any "constants" values get set to attribute flow.
There is work being done on this, but since this stuff relies of Arena <> Wire communication we have to be very careful.
All flows have their benefits and disadvantages.
I tend to go over a finished patch with a optimisation lens on.
Any "constants" values get set to attribute flow.
Re: Three ways of patch optimalisation
I work the same way, at first make shit pretty/functional.
Then optimise/consolidate/bypass/test.
I can build a patch in a few hour, then spend a day testing and cleaning it up
Then optimise/consolidate/bypass/test.
I can build a patch in a few hour, then spend a day testing and cleaning it up

https://get-juicebar.com/creator/tweelicht #tweelicht_visuals