On 2017-09-21, Martin . wrote:
What processes do sound designers here use to design realistic ambisonic spaces?
Unfortunately I'm no sound designer, or even practitioner, but I'll still but in with a bit of basic theory. I hope you don't mind... ;)
Do you record it all with tetrahedral microphones?
I'm reasonably certain that is the exception. Obviously, because ambisonic is precisely the only end-to-end system in existence which enables systematic capture of soundscapes, people who do such acoustic work will be over-represented amongst the system's afficionados. But it still remains the fact that true soundfield mics are expensive, as are the musicians playing to them. As such the easiest, least expensive and so the most common way to exercise the machinery remains doing it in-studio.
Are there good reverb solutions to achieve this with mono recordings - or to enhance ambisonic recordings? Do you use conventional DAWs or also game engines like UE4 and Unity?
It depends: do you just want to author audio, or do you want it to track within a game? Ambisonic uniquely goes for both, equally, so again game people will be over-represented among the doers. But quite certainly if you just want to produce a static soundscape, using a game engine would be overkill.
As for reverb in production, especially including distance from a source and its movement over time, that side of the picture hasn't really been set in stone anywhere.
Early work was content to just pan the direct sound, and maybe put in some Schroeder kinda artificial reverb into the W channel. As for distance, panning between W and XYZ was used as a quick and dirty fix.
What you really want to do instead is much more complicated. First, you'll want to maximally decorrelate the reverb over the channels, so that you'll end up with proper envelopment. For a synthetically panned source that calls for four mutually decorrelated reverb lines, running from the source to WXYZ (and more, if going to an higher order). You'd want to add early slap echoes from walls and other obstacles, just as you do in stereo reverb modelling, only now panned to come from all around. If your sound source is moving, you'd want to model Doppler effects with a delay line, and perhaps even go as far as to model Doppler in the strongest early echoes as well. Then if you *really* wanted to go hyperrealistic, you'd want to model near field effects at least for any close sources and/or echoes.
No current software does all of that, be it DAW or game like. As is usually the case, we bump into the fact that even plain old ambisonic, despite its age, remains hitech which is under development. Its full potential has yet to be unleashed, so that if you aim at the limit, you'll still often have to roll your own.
-- Sampo Syreeni, aka decoy - de...@iki.fi, http://decoy.iki.fi/front +358-40-3255353, 025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2 _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://mail.music.vt.edu/mailman/listinfo/sursound - unsubscribe here, edit account or options, view archives and so on.
