On Thu, Nov 07, 2013 at 11:14:10AM -0800, Eric Benjamin wrote: > Thank you; that's very informative. It would seem to me, with my > simplistic ways of thinking, that this method of detecting UHJ would > depend on there being a preponderance of data being associated with > a center image.
There is in the first example, and it surely simplifies things, but the method doesn't really depend on it. Unless the recording really contains sounds from all directions distributed evenly over the sphere, there will be some pattern in the cloud of points. And this pattern will be different for a straight virtual stereo mic and for UHJ. As long as some reasonable assumptions can be made about the distribution of the sources, it should be possible to identify the encoding. > Indeed, being able to see a group or cloud of data in the graph > depends on one direction being dominant. One or more, but not all. For example a mainly horizontal distribution (quite a common case) should still work. > But in practice it works. Does this explain why the second instance, > using the Midas Studios string quartet recording, has a different > phase offset? That's still puzzling me a bit. Even the straight stereo decode shows some dependence of the phase on direction, which for a strictly coincident mic should not be the case. Do you remember the details of the Midas microphone (I'm not even sure they were ever revealed) ? Anyway, these results are just a first attempt, there must be many ways to improve on them. I don't have much time ATM to research this further, but I will try some other recordings. Ciao, -- FA A world of exhaustive, reliable metadata would be an utopia. It's also a pipe-dream, founded on self-delusion, nerd hubris and hysterically inflated market opportunities. (Cory Doctorow) _______________________________________________ Sursound mailing list [email protected] https://mail.music.vt.edu/mailman/listinfo/sursound
