On 05/01/2011 01:15 PM, Richard Dobson wrote: > The issue for me is no so much the encoding (though asking content > providers, a.k.a. composers, to supply even a 9-channel file is IMO > pushing it), but the decoding, where the number of speakers required > seems to have its own version of Moore's Law. If encoding in 3rd > order means you can get a[n even] better decode to 5.1, well and > good; easy enough to understand why game developers would do that, to > get the best possible experience over the one truly existing and > established surround standard. > > For outreach purposes (promoting periphonic as well as horizontal > surround, promoting composers working with space) people need to talk > up the "simple" affordable layouts and delivery formats rather more > than has so far been done. The vast majority of works posted to > Ambisonia have been plain 1st-order; a few IIRC are second order. So > managing with the smallest possible channel counts at both encode and > decode stages remains IMO an important strategic as well as an > engineering objective.
if you were actually working with composers and producers rather than reminiscing about the grand old days of POA, you would find that the size of listening area and stability of POA simply doesn't cut it, ever, anywhere. check out the reality, the actual demands that film scorers, composers and producers have, and deal with it. POA is nice as a home listening format, and i agree it has a very good price/performance ratio. but the only way POA listeners will ever find actual content out there is from "backscatter" of HOA productions. in surround terms, HOA is like these little mono kitchen radios that are all the rage now: they sound surprisingly good, given their size, but they are not exactly turning heads out there or nudging the industry in new directions. POA is a lowest common denominator (and a very nice one at that), but not something with a future in film scoring and music production. > The danger with the arguments that, say, third-order is actually not > good enough is that commercial developers will just not touch > Ambisonics at all, since it is a territory that is forever changing > and remarkably lacking in consensus. It has taken long enough for 5.1 > to reach lower price-point DAWS. When even 7.1 is exotic, nobody is > going to make a DAW with a 16 channel bus only to be told a year > later that "we need more". it's really a bit frustrating to have to educate ambisonic elders about the up-and-downwards compatible nature of SH reconstruction. :( the point is: if you want to produce in tenth order and have the means to do it, your customer can still enjoy it on his/her 2nd order rig. vice versa, if tenth-order reproduction is the cinema gold standard of 2030, older productions in any of the x.1 formats and any lower ambisonic orders would still play just fine. the bus width restriction is actually a bogus argument, which is only true in practice because that's how avid and friends milk their customers. heck, going from a 16ch bus width limit to 32ch should not ever include more than a recompile on decently written software. > Have any listening tests actually been carried out to establish what > "typical" users consider to be sufficiently good localization? since we have no ambisonic content, the real question is: what do actual producers consider sufficiently good localisation? guess what: first order it is not :) > The > higher orders are sold as offering the most precise localisation; but > it seems to be more of an assumption than a proven fact that > localization (as distinct from "separation") at that level is > actually desirable. how can localisation and separation be distinct? if two sources are, say, 20° apart, it's very hard to separate them when you're sitting in the precise sweetspot of a FOA system, and totally impossible outside. unless of course they are of sufficiently different timbre, but that's a non-argument, because then i could also separate them on a mono rig. > At the end of the day, the problem is that HOA is not one standard > but a multitude of them - each combination and size of order, and > size and shape of speaker array, constitutes a separate "standard". that is utter nonsense. the most important selling point of ambisonics is precisely that it decouples the transmission format from the speaker layout. > So the final question is: if you had to choose just ~one~ HOA > standard for general production and delivery, to embed in the modern > equivalent of the AD7 (or in some future generation of Logic Pro), > what would it be? Or is that question simply unacceptable in > principle? there doesn't have to be. if i want to take my productions from third to fifth order, all i'd need to do is chance my panners so that they produce harmonics up to fifth order, and click "export". all that DAW manufacturers need to learn is that marketing and sales should have no say in how wide the busses can be - if you leave that to the techies, the answer is "arbitrarily wide". next very simple lesson is that panners should be plugins, and they should be the only entity that defines the bus semantics. then you need an n-way sum compressor with an intelligent enough (i.e. not very) sidechain to be able to perform a max() operation, and presto: arbitrary order ambisonics. -- Jörn Nettingsmeier Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487 Meister für Veranstaltungstechnik (Bühne/Studio) Tonmeister (VDT) http://stackingdwarves.net _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://mail.music.vt.edu/mailman/listinfo/sursound
