Hello and thank you for reading!
I'm currently starting work on a paper comparing the performance of the Unity Audio system with the Audio system if it used the DOTS.
For this I'm planning to write my own audio engine in Unity so that the ECS has the same functionality as Game Object audio except more performant🤞🏼 . The question is** should I use the DSP Graph or create my own Burst compiled audio system?** From what it looks like to me DSP Graph has been left behind and is poorly documented. Are there any plans for this system to be re-maintained in the near future or is Unity going to continue using the FMOD Engine for now?
#DOTS Audio: DSP Graph vs Custom Tooling
digital urchin
crisp ibex
Seems like you want to get into a lot of work, as unity audio system has many features, such as mixers, volumes, pitches, loops, spatial audio, effects, etc.
DSP graph imo would save a lot of time, rather than doing it from scratch scratch. However, there is already an open source attempt at that with LatiosFramework: https://github.com/Dreaming381/Latios-Framework-Documentation/blob/main/Myri Audio%2FREADME.md
(Keep in mind, it has been in the works for a few years, and still lacks some unity builtin audio features)
I personally wrote a little ECS wrapper for FMOD and use it, easy and fast
digital urchin
Fire response! I'm still working on the theoretical framework so Latios is exactly the type of thing I was looking for. Thanks a lot.
digital urchin
Just saw the beta release of 6.3. I think it will be more valuable to use the scriptable audio pipeline. Perhaps I'll do a full comparison of all systems? Depends on how fast all the testing goes. Will post any changes or conclusions here.
digital urchin
Update on this: I ended up making a simple audio system with a single rootoutput. Its currently a bit dirty but i intend to polish it and make it open source. Its a simple ecs audio system and runs about 4x more performant than the unity system. Its not limited so you can have as many voices as cpu and memory allows.