[linux-audio-dev] [OT] Info required: FLAC on W*****S

Fons Adriaensen fons.adriaensen at skynet.be
Thu May 20 11:50:30 UTC 2004

On Thu, May 20, 2004 at 08:00:31AM +0200, Jens M Andreasen wrote:

> I went to ambisonic and read the FAQ. I do not agree with them when they
> say that 2-channel stereo is only good for imaging between the speakers.

I do not see such a statement in the FAQ. 

> It is possible by using phase differences (and the assumption that
> people are not living in sound-dead laboratories) to project sounds
> outside of the stereo-field.

It is indeed possible at low frequencies and using 2 speakers, to create
a sound field that has a gradient that points outside the line connecting
the two speakers. When you do this with only two speakers, the resulting
image will be very unstable as there is a strong standing wave component.
Ambisonics uses similar techniques (in fact using all speakers for any
particular direction) to correctly reconstruct the sound field.

What the FAQ does explain is that the old 'quadrophonic' systems and
the new 5.1 and 7.1 setups use pairs of speakers to position sounds
between them, and that this simply does not work except in the front

> I did not find anything either to support that 4-channel is superior to
> 7-channel ... Maybe I was not looking hard enough ??

Again, the FAQ does not state that '4 channels is superior to 7'. It does
say (see above) that the ITU n.1 systems are based on false assumptions,
while the Ambisonic way of doing things is based on solid knowledge about
human hearing mechanisms for its principles and on solid mathematics for
its implementation.

Do not confuse '4 channels' with '4 speakers'. The 4 channels of an 
Ambisonic B-format recording do *not* correspond to speaker signals
at all. They are decoded in function of the actual number and layout
of the speakers. In fact you can decode them for reproduction on
a 5.1 or 7.1 layout if you want.

It's interesting to note why the ITU n.1 layouts are what they are.
They where created originally for sound reproduction for movies.

- The Centre channel was necessary in order to have reasonable positioning
of the dialogue, which normally happens on-screen, In particular in the
front rows the audience would otherwise see two front speakers with a
very wide angle between them, and this simply doesn't work. In fact 
you do not need a separate *channel*, a separate *speaker* fed with the
right signal will do as well (or even better). This was not realised
at the time, and anyway there are practical advantages to having the
dialogue on a separate channel (it allows for multi-lingual distribution).
For realistic reproduction of music, the centre channel (again I mean
*channel* and not *speaker*) gives a completely artificial emphasis
to front sounds, and should never be used.

- The separate VLF channel. In movies this is used for effects only.
Why is this a separate channel ? Not because there is no directional
information at low frequencies, but just beacuse of the technology
limitations at the time these systems were developed. With analog
magnetic recording (the norm at the time), you can not combine high-
intensity VLF sound with the normal signals without introcuding 
lots of distortion. Also, as long as this is used only to reinforce
visual effects and realism doesn't matter, it's cheaper to have
a few big subwoofers than to require extreme LF performance for all
speakers. But without the visual cues which distract your attention,
this simply doesn't work.

So why, you may ask, has this become the commercially dominant
technology ? The answer is very simple : for the same reasons that
Windows is the dominant OS for computers. 

The big multimedia companies know very well that there is a lot
more money to be made from 'home entertainment' than from high
quality sound reproduction. And since the movie industry, the
music industry and the equipment manufacturers are nowadays
one and same, the result is easy to predict.



More information about the Linux-audio-dev mailing list