I have conducted some tests this morning in order to try and clear up the confusion regarding measurement headroom and microphone setting. Specifically, why are some of us able to measure at higher levels than others without clipping?
The settings/configuration for the testing was:
- AVR set to stereo mode, all trims set to zero
- REW signal is being output to Left+Right+Subs
- Using UMIK-1, 12dB sensitivity, with CSL 90-degree calibration file, "Sens Factor =-6.352dB"
- UMIK-1 in Windows set to 24-bit, 48000Hx (only choice)
- RealTek HDMI Output set to 16-bit 48000Hz (my system has no 24-bit option, don't think it matters)
For the initial set-up, I opened the REW signal generator and output a Pink Noise Sub calibration tone at -12dBFS (same level as the REW measurement sweeps). Using the REW SPL meter, I adjusted the AVR master volume so that the SPL meter was registering 90dB. This is the "desired level" for my measurement sweeps. On my system, the AVR MV value of -15 achieved the desired level.
Now, for the test, I:
- Opened the Windows Input Device (UMIK-1) and varied the input level in increments of 25, from 0 to 100.
- For each value of the mic level, I ran the -12dBFS pink noise tone and observed the output level in the REW SPL meter.
- And finally, I ran a test measurement sweep, 15-20,000Hz, and observed the headroom during the sweep. The results are listed below:
- The resulting output level of the measurements is consistent above a mic setting of 25. This should alleviate concerns that some of you have expressed WRT mic input levels affecting REW measurement results. As long as the mic level is 25 or higher, there is no need to worry. The REW measurement curves were identical for each of the four mic input settings, 25, 50, 75, and 100.
- The headroom is indeed sensitive to the mic input level that is set in Windows. If your input level is currently 100, you can improve headroom by approximately 3-5dB for each reduction of 25 in the mic input sensitivity. This is good news.
Finally, I conducted one more test. With mic input sensitivity set to 100, I increased the AVR MV to -9, ran a measurement sweep at a level of 95dB, and received the clipping error message. By reducing the mic input sensitivity to 50, I was able to run the same 95dB sweep with a headroom of 6.8. Problem solved! I was even able to measure at a 100dB level and still have a headroom of 1.5 (darn, that is loud!).
I encourage everyone else to run a similar set of tests in order to establish the safe operating levels for your mic and AVR. For anyone with a UMIK-1 with 18dB sensitivity, the measurements are likely to produce different results.