Hi Richard, thanks for your interest and investigations, a more thorough examination of DBAP would be very welcome. It has been used by a number of composers and sound artists internationally, and the reports I get back have been positive, but apart from the work at Queen Mary University of London, no systematic examination has been done so far.
> I've just done a quick analysis of DBAP as documented in the ICMC2009 paper. > Two comments: > * Equation (6) seems to be missing a square-root (not a big deal). Thanks, this was pointed out by John Reiss as well, and I have to add an erata to the pdf for it when I get the time. > * The outputs are very unfamiliar! I've attached (hopefully) a > map-of-the-world plot of the output of one speaker in a cube layout - which > has a curious set of holes. The plots are a long way from what's produced by > VBAP, Ambisonics etc (e.g. there doesn't seem to be any concern with energy > vectors etc) so perceptual studies definitely seem like the next step... The attachment was unfortunately scrubbed from the list and I didn't manage to download it. Do you have a possibility of either uploading it somewhere and post the link, or mail it offlist so that I could upload it and post the link? Generally I wouldn't use DBAP in situations where VBAP and ambisonics works well; where the audience is confined to a small area at the middle (sweet spot) surrounded by a ring/sphere/cube of speakers. DBAP was designed for situations that defy the idea of the sweet spot and where it is not possible or desirable to position loudspeakers in a surrounding fashion. Another limitation of DBAP that is important to be aware of is that it starts breaks down if the source is moved outside the convex hull defined by the speakers. With a source at infinite distance, the levels will be the same for all speakers (because the relative difference in distance becomes infinite small) and instead perceived localization will be dominated by the proximity effect. So if you discover problems in a ring/sphere/cube layout, I won't be surprised. I have tested with a ring of 8 speakers a few times, and then moving sources around along the ring. My experience has been that you do get an idea of the location of the source, but it's less defined than e.g. 3rd order ambisonics or VBAP. Analysis, perceptual studies and evaluation tests would all be very interesting to get done. If so it would be interesting to do that not only for standard rigs were he other techniques excel, but also off centre and in non-ring/sphere layouts, as this is what DBAP was made for. Unfortunately this is beyond my capacity, I don't have the institutional backing required for resources and ensuring proper scientific rigor. To get an idea of what kind of loudspeaker setups I tend to work on in my sound installations, you could e.g. view the documentation of a work I did at a fortress near Bergen in 2008. For this I used 32 speakers distributed along a line all the way down the corridor. I consider this my longest composition to date, lasting for close to 200 meters... ;-) For this I used the 1D version of DBAP. The documentation is available here. http://trondlossius.no/works/18-imploding-spaces-2008 More bizarre speaker layouts can be seen here: http://trondlossius.no/works/30-white-out-2005 I guess this might serve to illustrate that sound artists sometimes might have needs in terms of spatialisation techniques that might differ widely from the ones that the audio engineering society mainly seems to be focusing on, and that seems to be geared towards concerts, cinema and home cinema. Best, Trond _______________________________________________ Sursound mailing list [email protected] https://mail.music.vt.edu/mailman/listinfo/sursound
