Level Issue

[Edgeyboy69]

Hi, I'm relatively new to REW and have renewed interest in understanding what's going on with my 2 channel system. I am using a UMIK-1, HDMI out of a MacBook pro to a Devialet D Premier using it's HDMI input.

I've obviously done something wrong or changed something, as when I take the measurement(s), the SPL is low (not at 75db).

If I turn up the volume to match when using the generator to match the incoming signal it gets way too loud to the extent I'm worried it'll damage the speakers (B&W 802D3's). 50db is realistic maximum. From memory, it was ok on previous attempts.

What am I missing? Thank you for any input / help. Much appreciated.

 

[John Mulcahy]

If levels are low but you can hear the test signal it is the input that needs fixing, not the output. If the signal sounds loud it is not 50 dB!

To get correct levels shown on the graphs with the UMIK-1 the UMIK-1 must be selected as the input device and as the input and the cal file must be loaded. Note that this doesn't change the shape of the graph at all, just where it appears along the SPL axis.

 

[Edgeyboy69]

Thank you for the prompt reply. I have attached photos in the hope you can spot what is not quite right. I have checked and the UMIK-1 is selected as the input device and the cal file has been loaded,

 

[EarlK]

It all looks in order to my eyes.

For the heck of it, load a UMIK1 calibration file into that empty slot ( bottom right of your last pic ) called "Default Input" + then close + reopen REW.
- Set the UMIK as your Default input to see if there's any change.

Apart from that, you can always go to the "Preferences Menu" and then choose "Delete Preferences + Shutdown" .

Re-open REW and then reload your UMIK calibration into their respective slots.

Doing all that might clear-out-the cobwebs" ( so to speak ).

 

[Edgeyboy69]

It all looks in order to my eyes.

For the heck of it, load a UMIK1 calibration file into that empty slot ( bottom right of your last pic ) called "Default Input" + then close + reopen REW.
- Set the UMIK as your Default input to see if there's any change.

Apart from that, you can always go to the "Preferences Menu" and then choose "Delete Preferences + Shutdown" .

Re-open REW and then reload your UMIK calibration into their respective slots.

Doing all that might clear-out-the cobwebs" ( so to speak ).

Thank you. I tried those suggestions - unfortunately no change, Also uninstalled the REW app and then installed a "fresh" version. Unfortunately no change either,

 

[EarlK]

I suppose then that I'm out of ideas // unless the UMIK1 calibration file has been tampered with .

 

[Edgeyboy69]

I suppose then that I'm out of ideas // unless the UMIK1 calibration file has been tampered with .

Again, thank you. I re-downloaded the cal file. Unfortunately, no change

 

[John Mulcahy]

Can you attach an example measurement mdat file?

 

[Edgeyboy69]

Can you attach an example measurement mdat file?

Hi, This is one I did last week, showing the same SPL issue. Thank you

 

[Edgeyboy69]

Hi, So I hope it's now ok... I found a great tutorial on YT which is very clear and straightforward:

I now have a graph that looks like this... mdat also attached. Hoping it looks ok?

Is the "reference" point (?) for working out peaks and nulls 75db in a straight line across?

Now to start working on convulution and Roon..

Thanks again for all input

 

[John Mulcahy]

The target line can be flat at low frequencies, but it should follow the general trend of the response. What was done differently for the latest measurements? The June set have the default SPL offset value (86.5 dB), which would usually only happen if the sensitivity line was missing from the cal file or the cal file wasn't loaded when the measurement was made.

 

[Edgeyboy69]

The target line can be flat at low frequencies, but it should follow the general trend of the response. What was done differently for the latest measurements? The June set have the default SPL offset value (86.5 dB), which would usually only happen if the sensitivity line was missing from the cal file or the cal file wasn't loaded when the measurement was made.

I followed the instructions in the video regarding setting SPL and the generator, rather than trying to do it in the preferences section.I used pink noise full range. Seemed to work ok. The cal file was the same.

Regarding the line, what I was referring to was what SPL level to take as the average so to speak? On the 7th July chart, would that be around 70db?

Thanks again.

 

[John Mulcahy]

Around 72 - 74 dB by the look of it, with some room curve rolloff at higher frequencies.

 

[Edgeyboy69]

Around 72 - 74 dB by the look of it, with some room curve rolloff at higher frequencies.

Thanks John. For future reference, what is the methodology for doing this? Is it looking at a midpoint ?

 

[John Mulcahy]

More or less, If you apply heavy smoothing it is easier to see where the average level is to get a better idea.

 

[EarlK]

I'm still mystified as to why you had to do a manual SPL level setting.

Where's your calibration file come from ?

CSL ( Cross Spectrum Labs ) ??
- Maybe he stripped out the text that indicates sensitivity.

Have you opened this text file to see if that the 1st line ( "SENS Factor" ) is present ?

Here's a look at mine.

 

[Edgeyboy69]

This is the original one that was in the mac that I downloaded in March last year: Pretty sure it came from the UMIK-1 site. "Sens Factor =-1.582dB, SERNO: 7026682"
And this is the one I downloaded yesterday definitely from the UMIK-1 site using the serial number: "Sens Factor =1.418dB, SERNO: 7026682"

As you can see, they're different. Same with the 90 degree files too.

 

[EarlK]

What happens if you use the older calibration file ( that has a "minus" figure in front of the numbers )?

 

[Edgeyboy69]

What happens if you use the older calibration file ( that has a "minus" figure in front of the numbers )?

Good question. I'll try that and report back. Thank you.

 

[EarlK]

Edgeyboy,

At least one ( or both ) of those calibration files has to be wrong.

To choose the best one to use ( after relabel-ling the one having positive decimated numbers > as say ; test cal ) , load the one with the minus sign before the numbers into the UMIK1 preferences top slot ( seen below );

Then load the "test cal" having a plus value of numbers ( actual + sign is not necessary ) into the second slot ( having the word "MICROPHONE" included in the title.

Then open up REW's RTA window ( choose "SPL"> top left drop-down menu in window ) , engage the RTA while listening to nothing but room noise , look at HF area ( 10K to 20K specifically ) and you should see that area coming in at say + 20 db ( somewhat a realistic setting ) .

FREEZE the RTA display now

Go back to the preferences main page and choose a different calibration file ( use "test cal" now > the one with positive numbers ) .

Return to the RTA window , unfreeze the display ( make it active ) and you should see the overall SPL drop significantly ( 20 to 25db in my case ) .

I found the calibration file that had positive numbers gave a SPL readout ( above 10K ) in the minus zone .

That can't be correct > therefore choose to use the calibration file that includes a minus sign before the decimated numbers.

 

[John Mulcahy]

I found the calibration file that had positive numbers gave a SPL readout ( above 10K ) in the minus zone .

That can't be correct

Why not? It just means the pressure is below 20 uPa.