On Sat, Jan 22, 2011 at 04:38:19PM +0000, Augustine Leudar wrote: > Im sure Im missing something obvious here but humour me. With a stereo > signal I can just place two speakers in a line and have my stereo signal > send two discrete channels to each speakers, each channel representiong one > channel of my stereo microphone. The same with quadrophonic (with no > matrixing nonsense) - four mics go to four speakers placed in a square - > works fine, tried it hundreds of times, no decfoding involved. Why cant you > do the same for 3 dimensianal sounds ? Four mics surround sending discrete > channels to four spekers placed in a square and one for height information > going to a mic above your head - this should naturally represent the sound > field without any decoding , Ive done this and it has been quite effective > - so why the need for elaborate and expensive decoding ?
First, don't try and send HTML to this list, as you can see it will be removed. Four cardioid mics placed to be coincident for sounds arriving from horizontal directions (i.e. stacked above each other) and feeding a square of speakers will produce something equivalent to a first order Ambisonic system with a rather primitive decoding. Adding a mic pointing up and a corresponding speaker may produce some effects but there it ends. In general feeding speakers with mic signals will not repeat not 'naturally represent' the sound field. There is nor reason why it should. Intuition may be misleading in this case. Your question reveals that you have not even started to study and understand Ambisonics theory - the answer would be quite evident in the other case. You could as well ask a engineer why he needs complex numbers while you can do your bookkeeping without. Hoping you will eventually have a go at it, I'll provide a provisional answer. There are several good reasons why AMB uses 'encoded' signals: * It makes the recorded/transmitted content independent of the technology used to produce it and of the speaker setup used to reproduce it. It provides a 'natural' representation that will always work and capture the essential information. * The encoded form makes it easy to apply some transformations on the signal wich would otherwise be quite difficult to perform, e.g. rotation. * The encoded form is required anyway for correct reproduction as this requires some processing wich has to be performed on signals exactly this format, and and can't be done on speaker signals (unless you encode them, operate on them, and decode them again). * A 1:1 mic to speaker mapping may work in simple cases, but it does not scale to the equivalent of higher order AMB. First order AMB was the start of the art 30 years ago, today we can do much more, just because we are using an encoded format. Ciao, -- FA There are three of them, and Alleline. _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://mail.music.vt.edu/mailman/listinfo/sursound
