Extreme Crowd Noise

[PSEhub]

I've been testing in both lab and real environments and here is what I've found in terms of how loud crowd noise can be in dB before XPR7000 series units have issues. Extreme crowd noise is an even tougher environment than a lot of music concerts.

I broke my tests into 6 parts:

1.) Lowest allowed fixed mic gain with Noise Suppressor Disabled
2.) Lowest allowed fixed mic gain with Noise Suppressor Enabled

3.) AGC (Mars Mod) with Noise Suppressor Disabled
4.) AGC (Mars Mod) with Noise Suppressor Enabled

5.) Default AGC with Noise Suppressor Disabled
6.) Default AGC with Noise Suppressor Enabled


Immediately upon initial testing, 5 and 6 were weeded out as viable crowd noise settings, since they performed very poorly above 83 dB.

While noise suppressor offers extreme clarity in low-mid noise environments, the reality of dropped syllables starts to show in high crowd noise. Crowd noise is not only loud, but contains a very busy waveform with many frequencies.

Here are the results from 1-4 (5 and 6 promptly removed from additional testing)

Intelligible audio average cutoff level:

PMMN4069:
test 1, fixed, no suppressor = 96 dB
test 2, fixed, with suppressor = 91 dB
test 3, mod agc, no suppressor = 94 dB
test 4, mod agc, with suppressor = 88 dB

Internal XPR7000 series Mic:
test 1, fixed, no suppressor = 102 dB
test 2, fixed, with suppressor = 93 dB
test 3, mod agc, no suppressor = 97 dB
test 4, mod agc, with suppressor = 95 dB

NNTN8383 industrial noise canceling:
test 1, fixed, no suppressor = 109 dB
test 2, fixed, with suppressor = 95 dB
test 3, mod agc, no suppressor = 103 dB
test 4, mod agc, with suppressor = 94 dB


All dB measurements are an average across multiple tests of the level of crowd noise in which speech became no longer fully intelligible. If any syllables were dropped, the threshold was considered to be reached.

All test units had latest 2.4 firmware.

It is interesting that the PMMN4069 performs significantly worse than the XPR7000 series internal, regardless of orientation.
The INC mic performs significantly better than the internal mic by about 7 dB.

However, the noise suppressor dropped more syllables with the INC than the other 2 mics.

Best combination was lowest fixed gain, NNTN8383, with noise suppressor disabled.

If anyone else has experience with crowd noise settings and level of success, please post.

And don't use default AGC, it was unusable for this application and sounds terrible. That's why Mars fixed it.

 

[Mars]

I agree with PSEhub's advice: If operating in a high-noise environment, do not use AGC. Use the best linear gain (or negative gain) setting which permit intelligible communications.

AGC should be used for "average" or "everyday" type communications. It normalizes amplitudes. Audio quality while operating in a high-noise environment with AGC can be improved, if the user speaks very close to the microphone, in a loud voice. This causes AGC gain to "attack" and drop. But it is not nearly as effective as negative, fixed linear gain.

I'm surprised the noise suppressor didn't improve the clarity of communications in high-noise environments. I find it is best suited for MOBILE radios, as it does a fantastic job at eliminating vehicle/road noise. I do not use NS on my base radio.

I also found Noise Suppressor -- despite what Motorola claims in the "help menu", has a negative affect on the tone/clarity of ANALOG audio. They claim it's a digital-only filter. That's not true. This is why I leave it off unless absolutely needed. Also of note: NS is not menu-selectable. It must be engaged/disabled via CPS reprogramming.

 

[PSEhub]

Noise suppressor made the syllables that came through sound great, and chopped out the others.

Really good point about mobiles, mobile environment is a perfect application of the feature. Gets rid of any digital "wine" from harsh mobile background. Even ambe has a little.


Remember, my testing is in only one "fan-base" of noise. Even an apx7k had issues with extreme 1 sided at a few rare instances during the testing (threw it in the mix for the heck of it). As I'm sure others would agree, crowd noise is one of the toughest.

Even if NS makes 85% of the audio sound pretty, the other 15% of tx could be critical so dropped syllables are not an option and pretty must be sacrificed in this case.

The xpr7000 series has great mic omnidirectional sensitivity for voice (way better than xts for example) so it is user friendly and reduces the need for AGC when it isn't feasible (in terms of mic distance from user). The distance and angle the user talks is less noticeable on xpr7000 series compared to earlier models. Of course users should do their best, but I'm pointing out the improvement.