Putting THD in the perspective

[Rja4000]

Glad it worked out for you. Steep learning curve, but remarkably versatile, easily my favorite measurement software for use with soundcards and external ADC/DAC.

Yes.
I now run it from my own .Net code and store data in SQL DB.
The graph I published here is done with this code.
I'm still struggling with different DAC and ADC (for different sampling frequencies), but I'll get through it.
Then I'll seriously start to use it.
But that's quite off-topic here, sorry. I'll come back to that in another thread (when ready).

 

[capslock]

Another interesting reading about attempts to measure IMD on microphones
https://www.gearslutz.com/board/gea...41571-microphone-distortion-measurements.html

Note that they use 2 speakers with 2 DACs and 2 amps, to avoid the IMD to be produced by those.

I also found some measurements in Eartworks M50 review on Earthworks website (pdf)
https://earthworksaudio.com/wp-content/uploads/2017/07/EarthworksM50Review_AXJuly2017.pdf

View attachment 34112

View attachment 34111

Best performance is THD around 0.17% or -55dB

(Note: I copied those graphs from the freely-available-for-download document.
So I guess that's allowed by copyrights. If not, this should be removed.
This is extracted from an AudioXpress review)

Sorry to resurrect an old post here. Did anyone figure out why this graph seems to asymptotically approach 0.17%? Is this SPL-independent distortion in the microphone (unlikely) or rather the inherent distortion of the loudspeaker? My understanding from Stuart's article is that the loudspeaker was operated at constant level and the SPL seen by the microphone was increased by moving it closer to the cone. So if this distortion at the far field limit was not subtracted, this is what we see in this graph?

 

[KSTR]

@capslock,
Quoting SY in the AX article:
"Figure 3 shows the distortion vs. SPL. This was obtained using a 6” loudspeaker driver measured near field (about 2 mm distance to the microphone from the cone), with an amplitude-swept sine wave.
[...]
Using one of my reference microphones, I confirmed that the distortion of the source was under 0.3%, so the measured distortion here really is that of the microphone."

 

[capslock]

That is exactly what confused me. It seems to say that those 0.17% at lower SPL are real, while from the rest of the description I understand that no subtraction was done, so those 0.17% may well be the distortion of the driver rather than the microphone. And then what is measured at higher SPl should be whatever the reading is minus 0.17% (assuming distortion does not change with distance, which should be true in the piston region).

 

[SIY]

That is exactly what confused me. It seems to say that those 0.17% at lower SPL are real, while from the rest of the description I understand that no subtraction was done, so those 0.17% may well be the distortion of the driver rather than the microphone. And then what is measured at higher SPl should be whatever the reading is minus 0.17% (assuming distortion does not change with distance, which should be true in the piston region).

Not sure what's confusing. I ran a baseline distortion measurement with a low distortion condenser mike to confirm that the driver's distortion at 130dB (near field) was well below the measured distortion with the mike under test. So the measured distortion mostly arises from the mike, not the acoustic source.

I do not think a direct subtraction is necessarily valid unless the spectral compositions are the same in amplitude and phase.