Wednesday, December 20, 2006

Noise Floor? What Noise Floor?

For my performance review I thought it would be interesting to compare the self-noise of the electret capsules. I had noticed that the SoundProject Omni's had a slight buzz when I listened closely, so I wondered if there was any such subtle signature sound for each of the capsules. As it turned out, given the "test environment" of the Pieper warehouse, I could scarcely tell the recordings apart. But, just for fun and because I thought we needed another pretty picture on this site, I ran off a comparison sonogram of the four capsule mics side-by-side. It was produced from this recording (Test 5) which shares the same left to right order as the picture (Primo, Shure, SoundPro, Rapid).












In the sonogram, the purple-blues are the loudest frequencies and the yellow-whites are the softest. As the image contests, my ears weren't fooling me - in the droning hum of the warehouse, these capsules sounded practically the same. Any subtle differences might be attributed to the different rigs (headphone, parallel boundary) that these mics were mounted in. There are some small distinctions between the mics in the low freqency range (the fat blue bands) that may indicate that the Primo and SoundPro were a little bit quieter, but if one were to really want to test the noise levels of these mics, It might be best to find an anechoic chamber.

Brennan Alcott

added later:
Rob D's linear sonogram at -90dB to 24K:

click on picture to enlarge

1 comment:

Rob D. said...

Hi Brennan--

For whatever reason, I can hear more hiss in the first third and fourth segments. I did a sonogram in Amadeus and added it to your post. Its not as apparent as it would be were we using higher mic pre gain but with 90 dB of gain and a linear display, the step-up at three and continuation at four is apparent. The noise from segment 1 and 3 dosn't look as loud as it sounds but 4 shows the hazy cloud around 8K that is typically the visual evidence.

I agree with you that is not a big difference in these conditions and that is illustrative too. We were recording at Hi sens at level level "16" with the NH700 Hi-MD's.This is about the range where the self noise of electret mics becomes a non factor. But, the fact that the shure 183 is definitely quieter with 21dBA self noise, indicates that that the record level would have to drop to around "11" before the self noise of the other mics would be a similar non factor. It was a noisy barn compared to natural settings which often require gain settings in the 24-30 range and the self noise of the electret mics becomes very audibble.

The newcomer em158's we were experimenting with all semester has some great qualities to tout, but it doesn't look like self-noise lower than the WL-183's, as I hoped for based on specs, is one of them. Its extra brightness gives recordings made in noiser environments more bite and relief. NOise performance needs to be studied in quieter location as you suggest where we can crank the nh700 gain near maximum and then comopare the resulting sound files.

Thanks for taking your time for the study Brennan. You are right, Paying paying $80 each for WL-183's and $10 each for WM61a's would be hard to justify based on recordings made in a loud location like this. Rob D.