On Wed, Jan 26, 2011 at 05:33:05AM +0000, Augustine Leudar wrote: > First, don't try and send HTML to this list, as you can see it will > be removed. > > Sorry - I dont know what you mean - as far as Iknow I haven't sent any > html to this list (at least not intentionally) - I assume you don't mean > links as there was no link in my original message and > there are also many links in the messages on this list.
No, links are perfectly OK. But your messages (as they arrive to readers of the list) end like this: An HTML attachment was scrubbed... URL: <https://mail.music.vt.edu/mailman/private/sursound/attachments/20110126/b08d7757/attachment.html> which means your mail program is sending an HTML copy of your text as well. Most mailing lists have a 'no HTML' policy, and some, like this one, will actually remove it. > I only got so far but as I understand it it uses sound > pressure levels and phase differences to plot x,y,z spherical coordinates > which are then reconstituted in the decoding - out of curiosity Why the need > for the w coordinate - cant the sound pressure level be gleaned from the x,y > and z ? If they were just coordinates then you could find w from x,y,z. But they are something different. Image the sound field created by some source, e.g. a sine wave reproduced by a speaker. At each point in space there two physical quantities involved: * The pressure is going up and down at the frequency of the signal. This the W signal, which will be produced by an omnidirectional microphone. * To make the pressure variations possible at all, some air must move. Image a tiny volume of air around some point. It will be moving back and forth. The velocity of this movement is a sine wave, and it will (in normal cases) be in phase with the pressure signal. But it also has a *direction* - it is a vector in 3-D space. So it can be split as the sum of three components along some chosen x,y,z axes. These components are the X,Y,Z signals. A figure-of-8 microphone produces a signal equal to the projection of the velocity vector on its (the microphone's) axis. Three of them are required to have the full vector. For a single source at sufficient distance, the magnitude of the velocity is proportional to the pressure, they are in phase, and the direction of the velocity points towards the source. Pressure and velocity are closely related in that case, but even then you can't find W from X,Y,Z alone. The reason is that if you put the source in the exactly opposite direction, and invert its phase, then it will produce exactly the same X,Y,Z - the inversion of the signal and the direction cancel each other. So you need W (which will be inverted) as a phase reference. In the more general case, pressure and velocity are independent. If you have the same signal reproduced by different sources (or just a single source and a reflection on a wall) you get standing waves, and there will be points were the pressures cancel but velocity does not and vice versa. Now pressure and velocity are the only two physical quantities that define a sound field. The basic idea of Ambisonics is that if you can somehow reconstruct them at the listener's position, the the listener must hear the same thing as for the real source. First order Ambisonics (the version using only W,X,Y,Z) can do this in a small area, where 'small' is expressed in wavelengths. So this area is large for low frequencies and gets smaller as frequency goes up. Above a few hundred Hz it has become so small that it is even too small for single listener. So for mid and high frequencies, Ambisonics will use a different strategy which is based on psychoacoustics. For installations where the listening area has to be much larger, the 'psychoacoustic' strategy is normally used at all frequencies. Higher order Ambisonics (using more signals) can reconstruct the sound field in a larger area. It still needs the psychoacoustic approach for all but the LF range, but this improves as well as order goes up. At third order and above it becomes very close to VBAP in fact, with only the speakers nearest to a source direction contributing in any significant way. The difference with VBAP is that Ambisonics will not treat the directions corresponding to the speakers as 'special'. This helps to conceal the speakers being the source of the reproduced sound. Ciao, -- FA There are three of them, and Alleline. _______________________________________________ Sursound mailing list Sursound@music.vt.edu https://mail.music.vt.edu/mailman/listinfo/sursound
