Request: Pause network wire/op rendering while dragging/zooming

I’m always blown away by how smoothly a brand new default .toe pans while mouse-dragging the network.

Would it be possible for TD to pause its rendering/updating of wires/references/op viewers/etc while in a panning/dragging/zooming state so that the act of moving around the network can become smoother?

I notice a lot of judder on a busy network while panning, and I presume that the calculations pertaining to the rendering of the wires/references/viewers are a significant contributor to that judder. All I really need to see while panning a network is positions of ops, not so much the wire flow or the real-time content of op viewers.

My personal opinion is this would cause more issues than it would solve, and make live editing networks and tweaking settings very painful when the entire system pauses every few moments.

Perhaps this would be best suited to being a toggle in preferences or something like that if implemented!

Yeah, I mean for non-live editing. I imagine that a chunk of the optimizations that Perform mode does could be used for panning/zooming state. But maybe not. Anyway, I hope they consider it since every bit of smoothness counts in such a visually dynamic programming environment as TD.

Remember that you can also deactivate the viewer of the operators that you don’t need to see all the time.

This helps TD a lot with the rendering and with navigating and moving in/out networks.

It depend where I am at with my network build, but in a live scenario I often turn all the viewers off, and keep on only the ones that I need to see / edit / tweak. If a section of my network is build and it’s solid and tested, I turn all the operators viewers off.

I find also good practice to compartmentalize my network in chunks. Instead of having a massive network with everything in it, I make containers or base COMPs, called for example MIDI_IN, OSC_IN, GENERATIVE, OSC_OUT, DMX_OUT etc…

Great suggestions for dealing with the current state of performance.

Personally I do all of the things you’ve suggested, but there are sometimes scenes that I’m building quickly that need to have a lot of exposed viewers and un-collapsed COMPs because otherwise I’d have to keep re-opening/viewing them to remember what’s what. I agree that best practice is to keep optimizing the network as you go, collapsing things, closing viewers, etc, but there’s a use case here, when you’re ‘in the zone’ where a bit of better panning/zooming performance would be nice just from a UX polish perspective.