Hello everyone, We at Derivative are doing some 2026 planning, and we would like to know what emerging or long-standing devices, protocols and other tools that TouchDesigner should interface with in the future.
You may be aware of the Interoperability page of our wiki which is the central hub of what we currently connect with. Of those items are some that need an upgrade to newer manufacturer devices, updated drivers and API versions. We’d like to hear about what you need there as well.
So this is a general question to you all, what would you like us to hook up with next? What are we missing, what’s emerging?
Here’s the first vote for .gLTF. This is also a superfast format to load pointclouds from as it uses binary data buffers so the pointcloud data buffer can be directly copied from disk to GPU memory.
Taking a quick look at that page, “TUIO and TUIO 2” is mentioned but when you go to the TUIO In DAT page it says “The TUIO 2.0 protocol is not yet supported.” So that is a gap, or the docs need to be tweaked.
I own Hokuyo but broadening LIDAR support (SlamTec?) would lower barrier to entry to others.
Direct Display Out felt revolutionary but almost bricked the card I was using it on, so very finicky. Those difficulties are likely upstream of the Derivative team but it has so much potential, I’m curious if Vulkan exposes other APIs for this?
With the departure of Oculus tooling, what will the Steam Frame look like? Will there even be demand for this, worth keeping an eye on.
Overall quality of life for DATs we already have, like async support, or toolkits for javascript to dynamically tie web state to Touchdesigner params. This can be DIY’d but having a standard would accelerate development!
Very very curious about MCP tooling for Touchdesigner, if only to experiment with. Once again, first-party standards can help consistency with training/tooling, etc.
Package Manager (integration) and a first party package repository/index run by derivative (with support for Patreon/similiar). This is nothing that should be handled by a third party.
built-in async support.
async webserverDAT
nodeJS signaling server
hot reloading with the new project format
some form of remote engine/gang grouping/remote editing
I have some rough ideas on how but probably tricky to get CANopen nodes
Lab Streaming Layer (LSL) seems to be of interest, nice door to research/science
not sure if @plusplusone actually means the same thing here but some sort of “remote COMP” would be very interesting. A COMP in my main patch that is actually a .toe running on a different machine on the same network, with all ins and outs etc.
Sorry, the wiki for TUIO was out of date - yes the TUIO In DAT supports TUIO 2.0, let us know of any issues or questions about it. TUIO In DAT and TUIO.
OK we hear you loud and clear about glTF in and out, and its new splat support. Thanks for the push.
(aside - the File In POP supports .spz spalt files and the POPs examples have an example of its use with a splat shader. Also Alembic .abc files out and in are on deck for the next Official release with 2 new Alembic POPs.)
We will look at broader and inexpensive LIDAR alternatives - SlamTec seems in that range.
Re OMT, why does NDI not serve you well enough?
Not sure what you mean about the Oculus and Steam Frame, cem?
yes TD Remote Tools has been quite interesting too.
We are poking around at async webserverDAT - not trivial actually but not giving up.
As for OMT support, i currently heavily rely on NDI setup on live shows, IMAG and multicam setups. Just thought OMT can deserve a little attention and being open source i’d like to see it grow in the industry and see some software and hardware developers taking part of the journey.
Well although it is quite new but there is a lot to expect and test in the future. But features are quite on par with NDI (not all available for now though)
I think the main advantage here is the opensource aspect of it, making it available to anyone (devs, hardware solutions, pros ans anybody who wants to add their features…).
It is based on VMX codec which is also opensource and pretty well tested for years now. Though on paper it is not limited to this codec and could handle other opensource ones in the future.
Could be compatible with SMPTE 2110 in some ways (like NDI can be a codec used in 2110)
Bitrates are higher than NDI so possibly better quality and less latency. While sometimes lower bit rates can be useful when dealing with lots of NDI feeds, i could use much higher ones on some projects. No NDI HX equivalent (for now) but you have 3 bitrates to chose from : high, medium and low. Also supports Alpha.
Also can handle 32 audio channels (NDI limited to 16 if i recall)
All freely available for everyone to implement without any additional fee unlike NDI SDK which has limited features if used without fees.
I started looking at NDI alternatives when i started to see more and more compression artefacts when using it lately, which i never had since i used it from version 3. Seems like you can keep control over the OMT quality while NDI is doing its thing in the background for you (maybe i am wrong).
…also lighter codec and easier on CPU (if we are to believe the developers)
My note on Steam Frame is not a current request, but I was trying to cue you to what might be next for TDVR support. Its not even available yet but looks like the best platform for PC to VR streaming since the Vive.
Now that we have POPs it would be pretty sweet to have NanoVDB support for volumes - which used to be nvidia-only, but 5 years ago it was officially adopted by the Academy Software Foundation into the OpenVDB spec, so now it is also gpu-accelerated on non-nvidia hardware! (it supports Vulkan and others).
One goal would be to load vdb files into a POP, but another more advanced goal is, it would be awesome to be able to use such a sparse grid in a GLSL POP - which could open the way to to all sorts of larger realtime custom fluid/smoke/volume simulations developed by the community.
Updated nvidia flex op that works with rtx 50xx cards would be great.
Dont know if this fits, but toe as json export would be a welcome addition. Make it easier to work with version control on teams and might make LLM workflows in TD easier. I am familiar with the td-expand & collapse, but something more human readable and that does not require multiple steps would be nice.
It might be a minor issue, but the default integrated MIDI devices are somewhat outdated. BFC and BCR, for example, could perhaps be replaced with more modern ones like the Akai APC40MK2 or Launch Control, which are more up-to-date and standard.
Sorry for the bit of OT but before more stuff is added, I’d actually love to completely hide integrations. I don’t use 90% of them in most projects. There’s a 99,5% chance that I’ll never work with a blacktrax system in my life. Nvidia and DirectX are wasting space for 5 years on my M1 Mac.
I made a post to the wishlist about RIVE. It looks like you could build a native c++ node in TD based on their open source runtime. It has 2 way communication so you could trigger animations on command. This would be a great gap filler for doing more creative UI/UX and animations in TD projects. I would prefer not to have to do it thru a webrenderTop to avoid the extra overhead and lack of easy 2-way communication.
