Calibrated REW measures different levels of broadband noise

[tcoor8325]

Hi, I am trying to figure out why REW will measure the same level of a calibration tone as my SPL meter, but not measure the same level of pink noise.

I am using a GRAS Pistonphone calibrator that generates a 250 Hz tone at 114 dBZ, 105.4dBA when a GRAS microphone is inserted. My GRAS Microphone is connected to REW and I have calibrated the input.
The image titled 'REW 114dBZ Tone' shows the RTA analyzer measuring the tone when the pistonphone is mounted as shown in the image titled 'earTone'.
I get pretty much the same measurement when I insert the head of my SPL Meter into the pistonphone.
However, when I play noise in the room (4 speakers in a circle, 1 meter away, ear not attached), and measure first in REW (image 'noiseSpectrumREW') and then on the SPL Meter in the exact same position (image 'splMeter_Noise') I get very different measurements (88 dBZ vs 83 dBZ)

Perhaps something is wrong with the experiment. Maybe the SPL Meter and REW calculate energy differently. Maybe the microphones are significantly different. I am most skeptical of the REW noise measurement because it measures 88.61dBZ, but the highest point in the plotted curve is 76 dBZ.

Thanks!

 

[John Mulcahy]

Better to use a window rather than leave it at rectangular, particularly for the tone. The level above 22 kHz looks unusually high for a pink noise measurement, what's the sample rate? In the RTA screen shot the A weighted figure is also quite close to the Z and C weighted figures, which again suggests there's a lot of high frequency energy. What's happening in the plot above 20 kHz? How omnidirectional is the SPL meter mic at high frequencies?

 

[tcoor8325]

1. Good point about the window, will leave it on Hann
2. The sample rate is 192k. Were you concerned about imaging? Did you mean above 2k? I agree that I would expect a more regular downward slope, but I believe this is just a of my room, Head Simulator, and mic. This is a very good point, though. When I measure the Ear Simulator through REW, the silicone ear is detached and is facing left, directly towards one speaker one meter away. The microphone's enclosure also blocks direct sound from the speaker on its right side. The SPL Meter mic on the other hand is probably extremely omnidirectional and its path is not blocked to any speaker. I will measure the Ear Simulator microphone again, this time in the same orientation as the SPL Meter, and without the Head Simulator.