On Wed, Mar 07, 2012 at 09:59:20AM +0000, Dave Malham wrote: > Do you actually need to match the encoding hypothesis, except for > documentation purposes? You could just build a matching decoder on the > spot, couldn't you?
Yes. But IIRC, the original question was to identify the encoding used on certain historical recordings from a discrete set of possibilities. In that case one wouild try and find the best match. > Actually, one thing that occurs to me is that maybe > I've missed something but what do you match the statistics to do do this > - surely you would need to know the original positions of the sources to > do anything meaningful? If they were entirely random, yes. The assumption is that they are not, e.g. that most direct sounds would be near horizontal. If that assumption is valid, nothing more is needed. The locus of the horizontal sounds on the Scheiber sphere would reveal the encoding, just as the relative frequencies of the characters in any language allow to solve any simple substitution cipher - assuming the message is in that language and not random data. Ciao, -- FA Vor uns liegt ein weites Tal, die Sonne scheint - ein Glitzerstrahl. _______________________________________________ Sursound mailing list [email protected] https://mail.music.vt.edu/mailman/listinfo/sursound
