Quote:
Originally Posted by Schmidty
Theoretically in a perfect environment the difference "should" be 6dB, however since no environment is perfect, it is often closer to 3dB.
|
Ok, with the proviso that I'm no expert in audio engineering (in fact, I'm no audio engineer at all!): I'm going to bite here because although this is an old post, it's also the first post I found when googling 'reaper pan law' so others will probably be referring to this post, too.
Let me start by saying, if reality differs from your theory by 3db, then maybe it's time for a new theory :-)
Now, if you play the same sound through two loudspeakers (instead of one) then you are essentially outputting twice the audio power into the room than you were before, since the speakers behave essentially independently. Put simply, if you put 100W into one speaker, then some (small) number of watts of sound energy will be radiated (the exact number dependent on the speaker sensitivity). If you put 100W each into two speakers (so 200W in total) then twice the total audio power is radiated (since the speakers' sensitivity hasn't changed, so each speaker will now output the same number of watts of sound as the single speaker did previously; twice the electrical power in in total, twice the sound power out in total). Twice the power equals 3db. So your sound pressure levels in the room will be 3dB louder, if you just played the signal, unchanged, through two speakers instead of one. (But see the footnote.)
Now, it's true, if you take the same stereo signal that was used in the above experiment, and convert it to mono the 'normal' way (i.e. by simply summing the two channels) then having the signal on both channels gives you twice the signal level, which equals four times the power or 6db (power goes as signal squared).
Yes, it's true, doing the experiment through stereo speakers gives a different result from converting the stereo signal to mono and playing that; that's because converting a stereo signal to mono makes it sound different from the original stereo source (but you knew that already). And one effect of that is that it makes sounds panned to the centre 3db louder than they did before conversion.
But anyway, playing a sound through both channels
normally makes the result 3db louder (unless you're doing any subsequent processing that combines the channels again). Hence a -3dB pan law (which reduces the levels by 3dB when you pan to the centre) makes sense for the normal case - it keeps the sound pressure level constant as you pan given reasonable assumptions about the playback environment.
Quote:
That combined with the fact that a signal being passivly split, loses 3dB
|
That's true for certain kinds of signals, like radio frequency signals, where when you put then through an inductive splitter half the power goes down one route, and half down the other. But inductive splitters aren't really relevent to audio work.
For normal line-level audio signals, with a low impedance output and a (relatively) high impedance input, splitting a signal has no effect on levels. If my output is 0.7 volt RMS and I plug it into an input, that input sees 0.7 volt RMS. If I split it by a Y-cable into two inputs, both inputs see 0.7 volt RMS.
Put another way, normally with audio, it's levels that you're sending, not power. They actually work quite a lot like digital signals - you're just sending information, it's just the way you encode it that differs.
Corrections welcome, but I think the above is broadly correct
Regards
roy
Footnote:
Now, if you did the above experiment in an anechoic chamber, using a sine wave as your source, and moved a microphone about, you would find something interesting. At some points in the room, the signals would contructively intererfere - they would sum - and you would get the same 6dB effect that the stereo-to-mono conversion gives you. And in
other locations, the signals would
destructively interfere (because they'd be out of phase), and they'd completely cancel out resulting in silence. So although in
total it's still true that you'd be putting out the twice the power - in some place you'd get
four times the power (+6dB) and in others you'd get silence (-infinity dB).
In practice, in the real world, interference effects are rarely significant (and then only at very high frequencies). But interference effects are the reason you can never find the lost smoke alarm or other high-pitched beeping device - the sound gets quieter and louder as you move about - without giving you any clear idea as to where the source is.
[Edited slightly to correct and clarify]