June 21, 2017 at 11:33 pm #59282
I love and use syphon frequently but do not know much about calculating the computational “overhead” of where and when to use it.
I have been using it in Quartz Composer but more recently in Processing.
My question is this: how can I calculate how computationally taxing a certain syphon use case will be?
For instance. Is it more taxing with color? What about if I use more than 1 syphon instance in a program? How about resolution? Frame rate?
How can I begin to calculate these taxes without taking the time to build out the program and see for myself.
Is there any rough math? References? Maybe general rules to stay away from (i.e. large resolutions, etc).
This way I could avoid pitfalls before coding and structure the program best before developing.
Thank you so much for any tips or leads!June 29, 2017 at 7:07 am #59285
Syphon shouldn’t be adding significant overhead – the parts of the process to do with generating and affecting your content are likely to be the drain on resources. Only start to worry about it if you are starting to hit limits (eg dropped frames). In that case, try each part of your pipeline independently to identify the part that is struggling. If you still see problems without Syphon being involved, then Syphon isn’t the problem.
There is no simple maths you can do. Syphon treats all content the same – the only significant hit Syphon will cause is when you resize the Syphon frames, in which case the underlying surface is rebuilt. Otherwise the only limitation is, as with any graphics work, that the more pixels you are working with the more work needs to be done.June 29, 2017 at 3:24 pm #59286
Well that sounds like good news.
I was hoping that it would be lightweight enough. Most of my image pipelines stay at the same resolution all the way through such as 640 x 480.
I have found some trouble using Processing and Syphon. I have found that sometimes using Syphon in the code, as illustrated in the examples – using a PGraphics object – the sketch runs significantly slower.
However, when I use something like “Syphoner” and removing the Syphon code in Processing it runs much faster.
This prompted me to investigate and sparked the initial post here.
I will prepare a bit of sketch code to show this.
You must be logged in to reply to this topic.