R/E/P > Klaus Heyne's Mic Lab

When to reverse polarity

<< < (2/4) > >>

Scott Helmke (Scodiddly):
Thinking in terms of instrument mechanics, what would you expect to get from a microphone?  

For the most part regular sustaining sounds wouldn't matter, although with bowed instruments you do see a sawtooth wave with the polarity depending on which way the bow is moving.  For most of the other instruments I guess you'd see the initial attack showing up as important.  But... say if you've got a mic above a drum and the drummer hits the drum, would that initial hit be a positive or negative spike at the microphone diaphragm?  How about a mic under the drum?

David Satz:
Bill, hooray for your skepticism regarding this question (how people can claim to distinguish correct from incorrect absolute polarity without a live reference), but it does have a real-world answer. A small but significant fraction of the listening population can in fact correctly sense the absolute polarity of an audio signal on percussive sounds such as bass drum without a live reference for comparison--more or less as most people can generally tell whether something is green or red without holding up a printed color sample for comparison (but some people cannot), and a different fraction of the population can identify the pitch of a tone without an external reference. It isn't necessarily a matter of "golden earhood" as such, however. Some people who can identify correct signal polarity are not what we would call highly skilled listeners, just as some musicians who have "absolute pitch" aren't very good singers or instrumentalists; they are simply people who can hear that particular distinction clearly.

I don't know whether any experiments have been done to find out whether people can train themselves to hear absolute polarity or not--that would be interesting. All I know is that I don't seem to hear it. So I try to be careful to record in the correct polarity for the sake of those who can. Most professional audio equipment does preserve absolute polarity--but it's still worth testing, or else one's recordings will (literally) suck when they shouldn't.

Bill Yacey:
I imagine for some low frequency percussive sounds phase might be distiguishable. with higher frequencies the phase changes many times over a very small  distance between the mic and sound source. What if the violinist starts with an up bow or down bow, or the guitarist starts picking the string upwards or downwards? I am not sure that anyone could determine a difference in this case...

David Satz:
Bill, it's the attack transients of certain instruments, where the peak-to-average ratio is very high, that seem to be the main issue for people who can hear absolute polarity. They're really a form of noise--not smooth, continuous sine waves. But even some continuous waveforms produced by musical instruments are markedly asymmetrical as well, such as a trumpet with a "Harmon" mute.

This sometimes creates other technical problems as well. A microphone, preamp, recorder, or processor may be perfectly well able to handle a given sound pressure level or voltage on a continuous basis, but if the waveforms (for that same SPL) are severely asymmetrical, then the peaks in the positive-going direction or in the negative-going direction can exceed expectations. This in turn causes DC offset, and may cause asymmetrical clipping or other distortion on half of each cycle.

In turn that creates a situation in which phase distortion can sometimes be "heard" indirectly--not because the ear is particularly sensitive to phase distortion as such (mostly it isn't), but simply because any equipment can being pushed into audible overload on peaks. And the amplitude of any signal peak depends not only on the amplitude of its signal components, but also on their relative phase. If you want to consider slew rate limiting, that can fit in here as well.

Bill Yacey:
It sure does David, I appreciate your insight. I didn't consider assymetrical waveforms on my previous post; but I am still skeptical about this being an audible artifact when concerned with frequencies above 200 Hz. It would be interesting to conduct some blind listening tests to see if some people can actually determine this above low end frequncies.
Bill Yacey


[0] Message Index

[#] Next page

[*] Previous page

Go to full version