As there was a lot of discussion about microphone distortion here recently and I'm actually working on a pressure chamber which can produce >160dBSpl ... I spent an afternoon to do some microphone measurements.
Setup
A pressure chamber is a device with a speaker membrane with a very small and 100% sealed front chamber. This pushes the resonance frequency of the 8" speaker to about 350Hz in my case - the volume is that small. But you can produce huge pressure changes now!
Measurements as always done with proper gear (APx515, Class1 Calibrator, Benchmark AHB2 Amp etc).
Measurement frequency is 250Hz (to be sure we are in 100% pressure chamber behaviour and it's the recommended calibration frequency by B&K and GRAS), stepped level sweep, 31 steps.
Check of the chamber - how much THD produces the speaker.
Such a pressure chamber produces stupendous SPL at very little membrane movement - so there is very little THD from the speaker. To check my limits I did the first 2 measurements with a GRAS 1/4" 40BD capsule with 2 different "preamps" (actually just impedance converters, they are just called preamps but there is no amplification).
That's the result.
There is a LOT of noise and some strange stuff between 115 and 135dBSpl going on with these 1/4" capsules - legit THD measurements are only starting from 140dBSpl and up. So we produce 150dBSpl with <1%THD and 160dBSpl <3%THD. We don't know if that's the speaker or the refmic! (I have an idea ...)
But we are sure about staying under 1% <150dBSpl and 3% <160dBSpl
So how are our microphones performing.
Here you see the linearity graph. 160dBSpl is easy for our 1/4" capsule. Not so for the other microphones ...
M50 gets to 145dBSpl - spec is 140dBSpl
M215 around 140dBSpl depending on capsule sensitivity -> preamp distortion. Spec is 135dBSpl.
I didn't bother to look after the Behringer Mic spec as it's just here as deterrent example - but >130dBSpl is actually not bad!
But what about THD?!
Here it is - the truth about reference microphone THD:
As expected ... don't use cheap measurement microphones for delicate measurements. They can't do them. The Behringer is out.
Interesting is that the Earthworks has higher THD as the M215! Noise area is up to about 105dBSpl (there is a lot of noise in the pressure chamber cause of it's sensitivity!!!) but then we get valid measurements. At the specified 140dBSpl max it has already 2% THD - that's A LOT! Studio microphones are normally specified with 0,5% THD, somtimes with 1% when the manufacturer wants to push it a little.
M215 - 135dBSpl is where the preamp starts distorting - that's not a limit of the capsule! Actually - it is when you take the 0,5% THD rule ...
BUT - one of the M215 amplifiers produces more noise as it should. Noise should be less as the M50 (dark green) but number 1 has more noise (light green). Will have to investigate, maybe just dirt in the high impedance area (these mics don't get pampered).
You can calculate your THD from this graph for lower SPL. -20dB -> 1/10th of THD. Even the Behringer follows that rule, it's how a condenser capsule SHOULD behave in theory. And they do!
So what about higher order harmonics?
tbc.
Setup
A pressure chamber is a device with a speaker membrane with a very small and 100% sealed front chamber. This pushes the resonance frequency of the 8" speaker to about 350Hz in my case - the volume is that small. But you can produce huge pressure changes now!
Measurements as always done with proper gear (APx515, Class1 Calibrator, Benchmark AHB2 Amp etc).
Measurement frequency is 250Hz (to be sure we are in 100% pressure chamber behaviour and it's the recommended calibration frequency by B&K and GRAS), stepped level sweep, 31 steps.
Check of the chamber - how much THD produces the speaker.
Such a pressure chamber produces stupendous SPL at very little membrane movement - so there is very little THD from the speaker. To check my limits I did the first 2 measurements with a GRAS 1/4" 40BD capsule with 2 different "preamps" (actually just impedance converters, they are just called preamps but there is no amplification).
That's the result.
There is a LOT of noise and some strange stuff between 115 and 135dBSpl going on with these 1/4" capsules - legit THD measurements are only starting from 140dBSpl and up. So we produce 150dBSpl with <1%THD and 160dBSpl <3%THD. We don't know if that's the speaker or the refmic! (I have an idea ...)
But we are sure about staying under 1% <150dBSpl and 3% <160dBSpl
So how are our microphones performing.
Here you see the linearity graph. 160dBSpl is easy for our 1/4" capsule. Not so for the other microphones ...
M50 gets to 145dBSpl - spec is 140dBSpl
M215 around 140dBSpl depending on capsule sensitivity -> preamp distortion. Spec is 135dBSpl.
I didn't bother to look after the Behringer Mic spec as it's just here as deterrent example - but >130dBSpl is actually not bad!
But what about THD?!
Here it is - the truth about reference microphone THD:
As expected ... don't use cheap measurement microphones for delicate measurements. They can't do them. The Behringer is out.
Interesting is that the Earthworks has higher THD as the M215! Noise area is up to about 105dBSpl (there is a lot of noise in the pressure chamber cause of it's sensitivity!!!) but then we get valid measurements. At the specified 140dBSpl max it has already 2% THD - that's A LOT! Studio microphones are normally specified with 0,5% THD, somtimes with 1% when the manufacturer wants to push it a little.
M215 - 135dBSpl is where the preamp starts distorting - that's not a limit of the capsule! Actually - it is when you take the 0,5% THD rule ...
BUT - one of the M215 amplifiers produces more noise as it should. Noise should be less as the M50 (dark green) but number 1 has more noise (light green). Will have to investigate, maybe just dirt in the high impedance area (these mics don't get pampered).
You can calculate your THD from this graph for lower SPL. -20dB -> 1/10th of THD. Even the Behringer follows that rule, it's how a condenser capsule SHOULD behave in theory. And they do!
So what about higher order harmonics?
tbc.
Thanks for the work 👍
Another microphone distortion comparison can for example be found in the ARTA handbook, page 58: https://artalabs.hr/AppNotes/ARTA Handbook Version 2.4 English.pdf
Phantom power supply voltage can be a variable, as demonstrated by Isemcon:
Another microphone distortion comparison can for example be found in the ARTA handbook, page 58: https://artalabs.hr/AppNotes/ARTA Handbook Version 2.4 English.pdf
Phantom power supply voltage can be a variable, as demonstrated by Isemcon:
Linear THD scale, no specification of the class 1 ref mic ... don't get me started ...Phantom power supply voltage can be a variable, as demonstrated by Isemcon:
In theory H2 is dominating and higher order harmonics are fallen off pretty quickly, they only rise closer to the acoustical overload point. But do our capsules behave?
In short - yes. But - differently!
Here is H2 for all tested mics:
And H3:
H3 is about factor 10 lower as H2.
The area where we can do proper measurements is shrinking a lot now. While we still see THD from the Behringer (cause it's so high) we have influence of the preamp of the M215 at about 100mV Generator level -> 135dBSpl (could not change that scale quickly).
The M50 shows a very interesting behaviour - while it has a lot H2, higher harmonics fall off very quickly. That would be great for a recording microphone - but for measurements H2 also has to behave. It's a pitty, when you take the 0,5% THD rule it is only linear up to 130dBSpl.
Higher harmonics fall off even stronger so they don't bother us.
Yes, really. Don't waste time at debating about that. Go do measurements.
.
.
.
.
.
But as I know that there will be endless discussion anyways I printed a few more graphs - her you go
Number of harmonic is in the graph title.
And don't tell me you need H17 or data of noise which is factor 1000 smaller as the main harmonic - I don't give you that nonsense.
And cause it fits here - why are linear THD scales rubbish?
Here our comparison of the mics in linear scale from 0-5%.
While the Behringer is a little worse the other mics are pretty similar, no?
NO!
You have no useful measurement data below 120-125dBSpl for M50 and M215! You don't see where the noise area starts (about <105dBSpl) and which values are actual measurements. You don't see that the M50 has double the distortion of the M215. And it's very hard to see that the Behringer Mic has 10-20x the distortion.
PEOPLE are logarithmic! Just use the right scales. Please.
Here our comparison of the mics in linear scale from 0-5%.
While the Behringer is a little worse the other mics are pretty similar, no?
NO!
You have no useful measurement data below 120-125dBSpl for M50 and M215! You don't see where the noise area starts (about <105dBSpl) and which values are actual measurements. You don't see that the M50 has double the distortion of the M215. And it's very hard to see that the Behringer Mic has 10-20x the distortion.
PEOPLE are logarithmic! Just use the right scales. Please.
Interesting. There are plenty of microphone THD comparisons out there, but so far I hadn't stumbled across a comparison of the distortion components. Thanks. I also wonder how the behavior is depending on frequency.
As for my phantom power remark, I don't think what Isemcon threw in as a comparison object matters, or the graph scale (as long as we can see what's going on). It's the difference in higher SPL distortion between 12, 24, and 48V phantom power that's interesting. Doesn't matter for me, as with the EMX-7150, which I use, there's no difference beneath ~140 dB, and most I ever need to measure is 130-135 dB with PA speakers in a 2m GPM setup.. but, it seems to be a factor to look out for for academic research.
As for my phantom power remark, I don't think what Isemcon threw in as a comparison object matters, or the graph scale (as long as we can see what's going on). It's the difference in higher SPL distortion between 12, 24, and 48V phantom power that's interesting. Doesn't matter for me, as with the EMX-7150, which I use, there's no difference beneath ~140 dB, and most I ever need to measure is 130-135 dB with PA speakers in a 2m GPM setup.. but, it seems to be a factor to look out for for academic research.
Hi,
It's interesting. I have Dayton EMM-6 and Beyerdynamic MM1 and observed different distortion test results recently.
It was completely unexpected to me that some mics distort as much or even more than speakers. I mostly observed that H2 was elevated while measuring using Dayton EMM-6.
MM1 performed a lot better. Could do a comparison if anyone is interested in.
Regards
It's interesting. I have Dayton EMM-6 and Beyerdynamic MM1 and observed different distortion test results recently.
It was completely unexpected to me that some mics distort as much or even more than speakers. I mostly observed that H2 was elevated while measuring using Dayton EMM-6.
MM1 performed a lot better. Could do a comparison if anyone is interested in.
Regards
Does the 100% sealed front chamber have the test microphone inside?
Are the test microphones and connecting cables physically isolated somehow from chamber vibrations?
ONLY use 48V phantom power! 12V and 24V came from an ancient time and from battery powered devices. Nowadays they all produce 48V thanks to better DC/DC step up converters. (But beware, quality is not always the same).
THD over frequency - they should behave linear!
A pressure chamber is a lot - but for sure ot linear over frequency or wideband. So I can not perform this measurement in a meaningful way with this setup.
I did a quick check, linearised the chamber and did a run - it's linear inside the measurement parameters. But there are to many uncertainties to interpret too much into this. It's just that there is nothing mad going on and it stays about linear over frequency in this little range.
Yes - only the front part of the microphone. It goes through the lid and get's sealed with 2 layers BluTack.
Mic connects to the chamber lid. Cable is connected to the microphone. So both get some vibrations.
But I never detected influences, vibrations are not that high as you might think with these SPLs.
A good cable has no problem with vibrations - and you quickly get the difference to a bad cable
Thank you pointing this out!!!
I have been trying to tell people that for decades.
It's even much worse when there is only distortion in absolute numbers in dB.
I have detected the influences of vibrations on the response of cheap Behringer XM1800S vocal microphones which have poor (no..) mechanical isolation of the cartridge assembly from the external shell.Yes - only the front part of the microphone. It goes through the lid and get's sealed with 2 layers BluTack.
Cable is connected to the microphone. So both get some vibrations.
But I never detected influences, vibrations are not that high as you might think with these SPLs.
A good cable has no problem with vibrations - and you quickly get the difference to a bad cable
Stage vibrations through the mic stand and clip would cause broad band ringing that could not effectively be eliminated with HP filters or parametric EQ.
The ringing could be largely eliminated by putting carpet under the mic stand, damping the mechanical vibration transmission to the dynamic cartridge.
The BluTack may not be enough to mechanically isolate a mic that does not have an effective shock mount, the resulting transmission of that noise might look like distortion.
The Behringer H3 distortion rise looks quite different from the other mics, makes me suspect it might be a physical, rather than electrical difference.
Thank you for taking the time and effort in your professional schedule to perform these tests, and sharing them publicly for the benefit of amateur audio DIYers here.
These tests have many findings consistent with other microphone distortion studies
eg. MTG mic comparison, @A4eaudio 's microphone distortion study (TBA)
We have exchanged over a dozen messages via Conversation to get to this point. Please give me some time to analyze and summarize your findings for (my future reference &) other amateur audio DIYers who would like to performance frequency response and distortion / dynamic range measurements, as it relates to loudspeaker design taken indoors (nearfield + far-field) without access to anechoic chamber or pressure chamber (or G.R.A.S. scientific/laboratory grade microphones)
These tests have many findings consistent with other microphone distortion studies
eg. MTG mic comparison, @A4eaudio 's microphone distortion study (TBA)
We have exchanged over a dozen messages via Conversation to get to this point. Please give me some time to analyze and summarize your findings for (my future reference &) other amateur audio DIYers who would like to performance frequency response and distortion / dynamic range measurements, as it relates to loudspeaker design taken indoors (nearfield + far-field) without access to anechoic chamber or pressure chamber (or G.R.A.S. scientific/laboratory grade microphones)
SUMMARY:
IMPLICATIONS to come....
Microphone | | Manufacturer specifications | ||
Brand | Model | High Hz limit (+/-1dB) | Equivalent noise level (dB(A)) | Max SPL (% distortion) |
Behringer | ECM8000 | 20K | NL | NL |
MicW | M215 | 31KHz | 18 | 135(1%) |
Earthworks | M50 | 50KHz | 20 | 140 (NL) |
G.R.A.S. | 40DB | 70KHz | 44 | 166 (NL) |
| | | | NL- Not Listed |
Microphone | | Measurements courtesy @IamJF | ||
Brand | Model | 0.1% THD | 1% THD | 3% THD |
Behringer | ECM8000 | 95dB | 115dB | 130dB |
MicW | M215 (worse of 2 samples) | 120dB | 134dB | 136dB* |
Earthworks | M50 | 114dB | 134dB | 144dB |
G.R.A.S. | 40DB | >132dB^ | 152dB | ~162dB |
| | ^ :limited by self noise | | |
| | *:limited by mic electronics | | |
Microphone | | Measured H2/H3, courtesy @IamJF | ||
Brand | Model | 94dB | 102dB | 104dB |
Behringer | ECM8000 | 0.09/0.01/0.009/0.006 | 0.24/0.004 | 0.27/0.008 |
MicW | M215 (worse of 2 samples) | 0.05/0.04/0.09 | 0.024/0.016 | 0.03/0.01 |
Earthworks | M50 | 0.013/0.006 | 0.022/0.002 | 0.02/0.002 |
G.R.A.S. | 40DB | 0.28%^ | 0.12%^ | 0.08%^ |
| | ^limited by noise floor | | |
IMPLICATIONS to come....
