Fons and Jörn, If i can stretch your time a little further:
The material to be projected in the installation with some degree of "precision" is voice. Will also be projecting more atmospheric sounds that don't require accurate placement. I've uploaded a plan of the installation (which also functions as a user interface in Pd) here: http://reverberant.com/tmp/plan.jpg The red cubes are the speaker-positions. The blue cube is the listener seated at a dressing table. The listener will view in a false-mirror to his/her front, various characters entering and walking through the room - speaking as they go. It's these voices that need to positioned in a reasonably convincing way behind and to the sides of the listener. Fons - when you say a "corner case" you're saying that 6 speakers aren't enough? Just how many do you think is necessary (i've been trying to keep costs down - the other two channels of this 8 channel sound card are being used for other things). Cheers and thanks for your attention, iain On Thu, 2012-02-16 at 22:14 +0100, Jörn Nettingsmeier wrote: > On 02/16/2012 09:59 PM, Iain Mott wrote: > > On Thu, 2012-02-16 at 20:34 +0000, Fons Adriaensen wrote: > >> On Thu, Feb 16, 2012 at 06:22:15PM -0200, Iain Mott wrote: > >> > >>> To clarify, for a 6-speaker 2d array - ambdec can decode at maximum, the > >>> following inputs (3rd order horizontal, 0 order height): W X Y U V P Q ? > >> > >> Yes, but it doesn't make sense doing so. With six speakers you can't > >> reproduce 3rd order correctly - 2nd is the limit. > >> > >> Ciao, > > > > that's a shame - i've only got two speakers to monitor with presently - > > but strange, 3rd order seemed to be a sharper image than with the 2nd > > order - wishful thinking perhaps. > > well, you can use 3rd order hypercardioids to drive six speakers, and > indeed channel separation will improve, which results in less phasiness > and more pleasant diffuse-field sound. > but the soundfield will no longer be homogeneous - a phantom source > between two speakers will be softer than one on a speaker. the sound > field will be pretty garbled. > > > Would you recommend then for 6 speakers, 2nd order ambisonic for the > > simulated room reflections (and jconvolver tail reverb) and > > vector-panning (vbap) for the direct? > > you mentioned moving sources - for those, i would advise against vbap, > because the timbre and apparent source width will shift a lot as the > source moves. otoh, if you have static sources, vbap for direct sound > could outperform 2nd order ambi. > personally, i wouldn't bother with such a hybrid system, but there may > be good reasons to do it. > > one i could think of is when you have signal sets which are sensitive to > crosstalk, say an a/b stereo set. it will always sound nicer when routed > discretely to a pair of speakers than as two second-order ambisonic > phantom sources panned 60° apart.... > > best, > > > jörn > > _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://mail.music.vt.edu/mailman/listinfo/sursound
