Dayton Audio OmniMic V2 Precision Measurement System User Manual
Page 59
59 |
P a g e
levels of background noise or weak signal levels are indistinguishable from distortion. You can find
the distortion "threshold" of a measurement due to noise by simply silencing the test signal. Only
distortion levels that appear significantly higher (at frequencies of interest) than this threshold when
the test signal is playing should be assumed to be valid.
I can't measure harmonic distortion at higher frequencies -- the curve drops off the plot
OmniMic can "see" sound only in the range of human hearing (20kHz and below), so it cannot see
harmonic energy that falls at frequencies higher than this. So 2nd harmonic cannot be measured
above 10kHz; 3rd harmonic cannot be measured above about 6.7kHz, etc.
Why do square waves played through my speakers look bad (using the Oscilloscope)?
Most loudspeakers cannot reproduce square waves well. This may be due to imperfections in
frequency response or phase response. The reproduction of a square wave that looks like the ideal
typically requires the speaker (and OmniMic) to have flat response from about 1/10th to 10 times the
square wave frequency. It will not be possible to show a square-looking waveform outside of the
range of about 100Hz to 2kHz because of this. But it may also be because of sound reflections off of
surfaces within the room. Try moving the microphone closer to the speakers.
I get erratic looking plots or messages saying that "this computer may be too slow for these
settings"
This may be because of insufficient computer speed. In general, a processor of 1GHz or higher clock
speed (for Intel processors, possibly more for others) is required for use of OmniMic.
Problems can also occur because of interruptions in the computer's operating system or hardware
devices. In particular, wireless LAN cards inside computers frequently prevent some computers from
being able to handle microphone sound continuously. Try turning off wireless or bluetooth features
on the computer and disconnect from any internet connections. The small program "DPC Latency
Checker" (downloadable from
can be very helpful in debugging such problems when they
occur.
The frequency response at very high frequencies seems to drift around over time.
This is likely to be a side effect of slightly different sample clock rates used in OmniMic and your CD
or other digital signal source. To minimize this effect, choose the "sine sweep" signal rather than the
"psuedo noise" signal for measuring frequency response for your more critical measurements. Room
positioning of loudspeakers and setting of equalizers at lower frequencies can generally be done
using the more pleasant psuedo noise signal.
The impulse response of my speaker shows ringing before the impulse happens! That's
impossible!
This happens when your speaker has strong response out to and beyond 20kHz. OmniMic can only
"see" out to about 23kHz (displays to 20kHz) because of its 48kHz sample rate. It accurately
calculates both magnitude and phase response right up to the highest audio frequencies, but as far
as it can tell, the response drops completely beyond 23kHz, and data above that is unknown. The
correct impulse response that would be equivalent to a response that is strong with smooth phase to
some frequency, and then drops away above that frequency, will have "pre-ring" in the shown
response. This doesn't mean that the speaker has output before it sees input -- it only means that the
time response corresponding a calculated abruptly ending frequency response would have such
effects. If higher frequencies were known, and could be included, they would cancel the pre-ring
(notice that the pre-ringing wave appears to have a period of approximately 1/23kHz). Or, if the