On 2010-12-23, [email protected] wrote:
Presumably, the 3D/AA has embraced "object-oriented audio" in order to a) abstract from speaker layouts b) reduce number of audio "channels" to 6 or 8 (i.e. fit into 5.1/7.1 distribution media like Blu-ray) and c) to make production more streamlined.
Object oriented audio usually denotes mono sources which are synthetically panned right in the end, at the receiver side. This is all fine and good -- most synthetic ambisonic material is born this way as well, and with irregular arrays, this can in fact be the best way to go. I mean, a) it decouples the panning algorithm from the transmission one, so that ambisonic, WFS, whatever holophonic formalism can be used to render the stuff (sometimes the low cost render can be better using some other formalism than ambisonic), and b) the material remains unmixed right to the end, so that it encourages reuse if the object based intermediary is also exposed to reusers. (Cf. MPEG4 Structured Audio.)
What pains *me*, and what I told 3DAA directly already, is that this does not address captured physical soundfields at all. In that space conversions (between e.g. WFS and HOA) are extremely painful if at all worked out, HOA seems like the only technology with real height capability as of now, and so I wonder whether 3DAA can actually deliver real 3D there without incorporating some old ambisonic tech. In a very scalable form as well, because net bandwidth is still a real issue for multichannel online playback, even factoring in compression.
They've yet to answer my email. So I think it could do some good if the list's big guns could also weigh in on this, individually, wrt 3DAA. Before the latter settle on something too limiting and impracticable for e.g. cinematic soundscape work. Which *is* one of the biggest sellers after all, and something the Alliance recognizes even in their publicity material...
If the quality of the resulting 3D surround sounds terrible -- yes, QUAD comes to mind -- then what a waste of time it will have been.
They seem to be standardizing a transmission format. An object based one has its limitations (e.g. usually just mono sources), but then it can also stimulate innovation in how to render the stuff -- or a rerecognition of how well ambisonic panning actually does. It doesn't set the eventual sound, it sets a stage for precisely the kind of decoder innovation MAG envisioned. Only on digital age steroids. So I don't think we have a problem in there, at all.
As I already said, I see the problem in settling for WFS *only* (I have a really bad whiff of it and the patents underlying it, Iosono/Dolby, here). This could be very bad for free/open content, online.
I also see a nasty preservation problem: WFS pickups are pretty much never dense enough to capture higher harmonics, nor well-regularized enough to deal gracefully with oblique angles of arrival, nor do they especially attenuate spatial aliasing at HF within even the prominent lobe like coincident pickups do. So a hundred years from now, what do the librarians and audio preservationists have to work with wrt captured soundfields? A totally irregular representation which isn't any too easy to make right even with the highest end DSP and psychoacoustical modelling. I don't like that picture at all. (The synthetic stuff will be pristine as ever, provided that they really parametrized it right and openly published the parametrization. Kudos for that. But it's not enough per se.)
Is there an Ambisonic and/or Ambi-derived system that sounds really good that should be considered?
I actually think synthetic sources should be kept separate and rendered only at the receiving end. Because that effectively leads to "infinite order decoding". The best of them all, by a far shot, and conducive to a race in better and better decoders. Just what we want.
As for today's capture of live performance/sound(effects) in a 3D space, even Plain Old Ambisonic fares better than WFS or any other micing technique. Having heard Eero Aro's demonstration of POA 4-speaker pantophony, and Ville Pulkki's demonstrations of both unaided and DirAC aided pantophonic/periphonic work, I can say with certainty that most people hearing a simple SoundField recording via a state of the art decoder would be impressed, if not even fooled into thinking it was the real thing, sitting at the sweet spot and blind-folded. Plus I think we could probably do so well with even the basic B-format and some periphonic DirAC derivative that I would be fooled as well, eventhough I'm a pretty picky listener.
Really, even the old, analog, pantophonic implementation of ambisonic via BHJ transmission is frighteningly good on only four speakers. You *do* startle at a lateral sound, you *do* turn you head, the source *does* stay stable, and when you turn, it *is* there. Firmly in between no speakers. Turning backwards you do hear the phasiness, but it doesn't annoy you; it's rather like a poor man's hall echo or something. That then using precisely two channels -- and now we're talking about four, ain't we? Full B-format? (Personally I then won't vouch for out-of-sweetspot reproduction. It sounds massively spatially distorted to my ear.)
My guess is that MAG would be on that Working Group, if he was still alive.
Should be. Probably wouldn't.
Who should be there in his stead?
As far as classical ambisonic goes, I'd say Geoffrey Barton. But as far as the current folks dealing with modern and open/free incarnations of the technology go, my vote would be for Angelo Farina. Maybe Filippo Fazi, if he could forget the physics for a change and be willing/prepared to explain the tech in layman terms. I mean let's face it, people in these committees don't really understand physical acoustics or the slighter nuances of spatial audio; at least yet.
I'm pretty sure who we'd ideally want to represent the ambisonic community is someone with a) noted doctoral/lectorial or professorial credentials (so that others make note of se's comments), b) an engineering background (to keep it real, simple and to the point), c) political and interpersonal background/experience (for influence within a working group/committee), d) a tight backing crowd to cover the load of the factual bombardment that will be coming (this list could be the technical one, but where to get the political, financial, etc. ones?), and e) above all a nice guy who's always present (to facilitate and gain social appreciation within the crowd).
Unfortunately such well-supported technical-missionary-diplomats are few and far between. ;)
-- Sampo Syreeni, aka decoy - [email protected], http://decoy.iki.fi/front +358-50-5756111, 025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2 _______________________________________________ Sursound mailing list [email protected] https://mail.music.vt.edu/mailman/listinfo/sursound
