On Mon, Apr 23, 2018 at 01:10:56PM -0700, Fernando Lopez-Lezcano wrote: > You would need to "calibrate" the created 8 capsule array. That is, record > impulse responses all around it in a big space or anechoic room (enough to > accurately sample the spherical harmonics you want), and then derive an A to > B converter from that. I have some preliminary code in my SpHEAR project > that tries to do that, but it is not a "push a button and you are done" > thing at all...
Indeed not. To start with, how to do measurements that can actually be trusted is something you learn only with experience. Rest assured that the first five times you do this will produce near useless results, and you will notice that only *after* most of the processing is done. That is assuming that you have sanity checks in the procedure at all. If not, the result can easily look OK but actually be totally wrong. > In my case to get good calibration data I need to rotate the microphone with > no wobble and at different orientations (or if it is not _exactly_ perfect, > try to get away with calibrating out the average delays to all capsules). There's practically no way to obtain precise positioning, unless you'd use a robot to rotate the microphone, and that comes with its own sort of potential problems (like reflections on the robot arm). The only practical solution is to make the processing software detect small position (and other) errors and compensate for them. This, plus sanity checks, makes up at least three quarters of the code I use to calibrate Tetra and Octomics. And those parts are also the most difficult to get right and verified. So unless you want to spend a few months full time to develop and test all of this, I'd stronly advise to not even try it. Returning to the original problem: it is perfectly possible to obtain higher order room IRs from first order measurements. The dominant early reflections tend to be separated in time, so they can be isolated and recreated in higher order. For the reverb tail only the statistics need to be correct. It's not a trivial thing, but it can and has been done. Ciao, -- FA _______________________________________________ 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.
