Steam Audio Open Beta

  • No off-topic posts- Don't report more than 1 issue at once- For isolated issues or customer support visit help.vrchat.com ~ Thanks for your bug report!
ONSP ambisonics aren't translated to HRTF spatialisation correctly
Second world: two zones, both use first order AmbiX-style ambisonic clip. ONSP-spatialised one requires a specific setup which involves 2D mixing and decoding via ONSP middleware, while SA doesn't natively support AmbiX-style ambisonics, but can passably parse them under binaural HRTF mix. While not a correct way to do it, in a pinch this method can be used just as well. Unlike ONSP, SA handles spatial mixing just fine, but it doesn't handle 2D sources. Sources not upgraded to SA middleware are read as full track mixdowns instead of being processed as ambisonic clips, meaning you lose channel 3 and 4, along with directionality. Non-spatialised sources are also read as a basic mixdown. Technically speaking, losing direct first order ambisonics is not a significant issue only if we can gain access to post-process effects, as they are mixed down to ambisonic streams themselves. Effectively, having middleware that's capable of producing ambisonic streams is more straightforward than mixing phase dips yourself; 3D environment acting as a form of a DAW is much more convenient. However, this is an issue when it comes to backwards compatibility. As it stands, there doesn't seem to be a valid way to convert ONSP ambisonics to SA spatialisation, not out of the box. As such, here are some examples of what breaks: wrld_08f14312-1f6b-4284-ba18-e6443c2a6a8f: environment track is no longer spatialised properly, always appearing static (though this is true for current mainline just as well) wrld_aa458f70-972b-494b-9eaf-6f6b1483ca9b: ONSP ambix clip is broken the same way wrld_c716477a-de43-4fe1-8a87-9177cc05345e: ambisonics are preserved but appear rotated about 45° CW with significantly reduced mids (~1500-7000 Hz); no idea what ambisonics formats are used there
0
[1603] Loving Steam Audio so far... so far!
I've yet to experience any technical bugs or difficulties. The documentation for setting up steam audio is pretty straight forward and my test curves have been holding up well with the directivity setting toggled on (highly recommend by the way). If you're careful about your curves and speaker placement, the potential is absolutely there. The largest noticeable improvement has definitely come from the directivity setting which allows for realistic projection of sound volume within a sound space! I think creators will be happily surprised with the quality of just slapping on the steam audio source component and building a directivity volume that fits their source. Fidelity of sound appears to be higher as well. My ears are a little rusty but it sounds as though the sound is less compressed over steam audio versus the previous spatializer. Perhaps the processing of things is just cleaner which gives this illusion? I'm unsure and honestly a little lazy to look into the technical aspects of this. Try it out for yourself and see! Something that I've really wanted to see is the addition of audio mixing so that I can get even more in-depth with sound design. VRCHAT PLEASE ENABLE AUDIO MIXER GROUPS OR AT LEAST SOME AUDIO MIXING COMPONENTS LIKE PARAMEQ, LOWPASS, HIGHPASS, ETC. Additional items to test are: Stress testing player audio (goal: 40 active voices and 10 active sound sources Density and quantity tests (goal: 50 sources in 40x20x30 space) Check back in on my test world found here ( https://vrchat.com/home/world/wrld_61ae6613-4ff8-48cf-94c8-d79543f745d7 ). I'll be testing silly stuff over the next few weeks prior to implementation in my VR audio-visual experience venues.
0
Load More