REW first scan - mystery

[staticV3]

I suspect the mic.

Yet OP confirmed that this issue is plainly audible as well.

 

[Sokel]

This is all I have (which I do not know how to use).
Originally I wondered about measuring the DAC output voltage at the REW graph drop-off frequency area but was not sure which two pins to connect to and how to get it off DC Volts onto AC Volts.

Is the pin number reference I found correct?

Yes, correct.
Measure with AC, see my edited and also check the other post.

 

[Sokel]

Yet OP confirmed that this issue is plainly audible as well.

No objection to that but no connected mic I have seen in my life has reached -40dB self noise.

 

[1066]

There's something else to check too as in an acoustic measurement like that the level shown there is crazy and I suspect the mic.
While mic is connected just hit the RTA window at REW and show it to us.

Normally it should show your room's noise floor.

Thank you all very much for your help and patience - will take some time to set up for REW again.
I saw an RTA button in REW - but was not sure what it was for.

Back later.

 

[Keith_W]

There is something very weird here. First, take note of the noise floor. The Revels roll off at about 40Hz, below which it drops to the noise floor, which is about -40dB relative to the speaker measurement. This, I believe.

But look at what happens above 5kHz (left speaker, red) and 8kHz (right speaker, green). It drops to -140dB relative to the main speaker. There ain't no noise floor which is that low. There is some problem with the microphone.

What I can't figure out is why you think the top end is rolled off as well.

(EDIT) beaten to it already!

 

[staticV3]

Again: normal DMMs cannot be used to measure frequency response much higher than 1kHz, so not useful here.


This is also different on each DMM model so there's no way of knowing if OP is measuring a DAC defect or the DMM's inherent roll-off.

 

[Keith_W]

This is a total wild guess, but I wonder if there is a typo in the mic calibration file. Maybe a decimal point is missing or something.

 

[Sokel]

Again: normal DMMs cannot be used to measure frequency response much higher than 1kHz, so not useful here.
View attachment 436692

This is also different on each DMM model so there's no way of knowing if OP is measuring a DAC defect or the DMM's inherent roll-off.

Again, you're right but we don't ask for precision here.
Even if we read 2V at 100Hz and 1V at 10kHz there's enough that would be an -3dB error by the DMM, that's a long way from -40dB.

Edit: the shown DMM is True RMS as I see, usually these are probably precise up to 1kHz and the losses from then on are small, I just tested mine.

 

[sam_adams]

I got this graph and am puzzled.

MacBook.

Make sure that the sample rate in Audio MIDI Setup.app is set to 24 bit, 48 KHz for both the DAC and the UMIK-1. Make sure that the sample rate in REW is set to 48 KHz.

There may also be an issue with the DAC muting the outputs when there is no signal on one of the channels. The workaround for this is to output to both L+R channels but disconnect the DAC output channel that you are not presently trying to measure.

 

[staticV3]

Even if we read 2V at 100Hz and 1V at 10kHz there's enough that would be an -3dB error by the DMM, that's a long way from -40dB.

My AN870 drops from 1.0052V at 100Hz to zero (0.0000V) at 10kHz.

DMMs like these just don't work up high.

 

[Sokel]

My AN870 drops from 1.0V at 100Hz to zero (0.0000V) at 10kHz.

DMMs like these just don't work up high.

Just for the fun of it:




80Hz



10kHz



20kHz.

We know there will be losses, I'm with you at it.
But it will at least show some output.

 

[staticV3]

We know there will be losses, I'm with you at it.
But it will at least show some output.

Yours does. Mine doesn't.

How should OP know where on the spectrum his DMM falls and whether he's measuring a DAC defect or DMM roll-off?

It's a complete wild card. Without a known good reference, any result is useless.

 

[just1n]

This is a total wild guess, but I wonder if there is a typo in the mic calibration file. Maybe a decimal point is missing or something.

Agree it’s worth exploring. I would measure a test tone with the calibration file.

 

[Sokel]

Yours does. Mine doesn't.

How should OP know where on the spectrum his DMM falls and whether he's measuring a DAC defect or DMM roll-off?

The shown difference is 100dB, that's a pre's full attenuation range normally.
That's not a roll off, that's a no-signal. At all.

Even the slightest reading (200mV for example) at 9kHz throws the DAC fault off and he can search elsewhere.

 

[Keith_W]

Agree it’s worth exploring. I would measure a test tone with the calibration file.

Nah it's just a text file with numbers in it. Open it up with a text editor and take a look. Cal files should correct +/- 4dB at most. If he sees something like -140dB then that's the culprit

 

[staticV3]

Even the slightest reading (200mV for example) at 9kHz throws the DAC fault off and he can search elsewhere.

Or his DMM is like mine and shows no output at all, not because there is none but because the DMM doesn't reach high enough.

We just don't know. Why is this so hard to understand?

 

[Sokel]

Nah it's just a text file with numbers in it. Open it up with a text editor and take a look. Cal files should correct +/- 4dB at most. If he sees something like -140dB then that's the culprit

Or he can just upload it here.

 

[1066]

This is a total wild guess, but I wonder if there is a typo in the mic calibration file. Maybe a decimal point is missing or something.

This is the UMIK-1 with 0º pointed at MacBookPro screen from 30cm using MacBook's left speaker.
It does not show the cliff-shape response that I got with hifi / F206 speakers measurement.

I will try the other suggestions from contributors in turn.
I just need a bit of time as not very familiar with these things.


Does this show the UMIK-1 is working correctly?

 

[Sokel]

Or his DMM is like mine and shows no output at all, not because there is none but because the DMM doesn't reach high enough.

We just don't know.

Would it hurt to test it?
That's all I'm saying. It can be like mine and get lucky.

I'll stop now,I think I made my point.

(educational though, I was thinking of DMMs based on mine which is clearly wrong)

 

[staticV3]

Does this show the UMIK-1 is working correctly?

Yes, that looks about right.