Speaker measurement with rew

[Fredd3]

Hi,
I got a setup with a pair of B&W 804d2 speakers and I have previously used them together with dirac. However I sold the device with dirac and I'm now back to messuring manually.

My goal was to try and find any issues and correct them with the dsp in roon.

My mic is a XTZ room analyzer 2 pro and it comes with a calibration file that I loaded into REW. But after doing a number of meassurments in mono I'm not sure how to interpet them. If I upload my session here, would someone be willing to have a look?

I did one messurement with mic pointing straight up and another forward at the same location both times, then swapped the channel L / R .

I also tested the meassurement in xtz room analyzer software and got some strange data points at 8 k, their support page suggests this is because the mic is not set to 24 bit, however I'm not able to select 24 bit in windows. Perhaps this causes issues also for REW. Not sure...

Regards,
Fredrik

 

[Hipper]

I can't open the file with REW. It says the 'File is not in mdat format' although the file name does end on '.mdat'.

 

[Fredd3]

Hi,
I rezipped it and uploaded. Seems like the other channel messurements got lost somewhere. Anyway compaired with the xtz one, can I make it look smoother like that? It's very hard to read anything from such a choppy line.

Regards,
Fredrik

 

[MakeMineVinyl]

Hi,
I rezipped it and uploaded. Seems like the other channel messurements got lost somewhere. Anyway compaired with the xtz one, can I make it look smoother like that? It's very hard to read anything from such a choppy line.

Regards,
Fredrik

Apply 1/6th octave smoothing in REW (ctrl-shift 6).

 

[Fredd3]

Apply 1/6th octave smoothing in REW (ctrl-shift 6).

Great, thanks! so with smoothing on I can see the heavy roll off in freq after 10khz. But I suspect this is a measurement problem and not real. What do you think? It should probably very noticeable otherwise? I have emailed xtz support to see if there is any solution to me being unable to use 24 bit mode.

/Fredrik

 

[MakeMineVinyl]

Great, thanks! so with smoothing on I can see the heavy roll off in freq after 10khz. But I suspect this is a measurement problem and not real. What do you think? It should probably very noticeable otherwise? I have emailed xtz support to see if there is any solution to me being unable to use 24 bit mode.

/Fredrik

I wouldn't worry about that since it could be room absorption in addition to rolloff from your speakers. I would be more concerned with getting the area in the middle around 600Hz flatter.

 

[Backflash]

Hi,
I rezipped it and uploaded. Seems like the other channel messurements got lost somewhere. Anyway compaired with the xtz one, can I make it look smoother like that? It's very hard to read anything from such a choppy line.

Regards,
Fredrik

Here is a good tutorial, try going as he shows and then adjust filters for your personal preference later, in your measurement I'd apply filters line at about 80db, lowering the area around 600hz to upper mid range levels brings a lot of clarity from personal experience.
Valley in 2-4k is your room reflections, no carpet? Drop in 10k range definitely the room, or distance from the speaker, how far was the mic? If you want to check if speaker tweeter work properly, just in case, get your mic directly at tweeter about 1ft/30cm away.

 

[Backflash]

Apply 1/6th octave smoothing in REW (ctrl-shift 6).

Got a question about that, isn't it too generous? My Adams look like top tier Genelecs if I apply that. Is there any consensus on what smoothing should be used in what situation?

 

[storing]

Here is a good tutorial

It's ok but also read the comment section: more interesting to (also) measure 1 speaker at a time, it's more common to place the microphone upright to measure the room not just the speakers (though I haven't verified differences).

 

[Backflash]

It's ok but also read the comment section: more interesting to (also) measure 1 speaker at a time, it's more common to place the microphone upright to measure the room not just the speakers (though I haven't verified differences).

It's a good tutorial of how to begin working with REW, additional knowledge is an abyss for this topic.
I did separate measurements, and if you correct one speaker at a time it becomes a mess, they begin to cancel each other in weird places on graph, I guess that's dependent on room a lot, milage may vary, maybe I messed up filters somewhere.

 

[Hipper]

Got a question about that, isn't it too generous? My Adams look like top tier Genelecs if I apply that. Is there any consensus on what smoothing should be used in what situation?

For proper measurement and adjustment, I use 'No Smoothing' for the region say 0-500Hz because I want to know the details and with the right tools you can be quite specific about targeting filters (e.g Parametric EQ), and something like 1/12 for the full range when I want to understand the higher frequencies and you can use wider ranges for adjustments (such as a Graphic EQ).

If you want to see the general trend then a 1/6 or 1/12 smoothing is useful.

 

[ernestcarl]

If you disable or clear out the microphone calibration file 10kHz and up rises -- dunno if there's an issue with the calibration file. A lot of reflections in your measurement... even after applying some IR windowing and smoothing, you've got several phase rotations and an unusually high GD -- it could probably be just the room.

I also don't know what to make of this strange peak prior t=0:

Envelope Time Curve

*note the strong reflection around 3ms which may be adding to the noise

 

[Fredd3]

After watching the video earlier I made a new measurement. I did left and right channel separately and 3 positions per channel in the sofa. The attached is from when I try to make a eq curve from it...

Seems like I don't get much low frequency out of my speakers... but the top looks much better

/Fredrik

 

[Fredd3]

If you disable or clear out the microphone calibration file 10kHz and up rises -- dunno if there's an issue with the calibration file. A lot of reflections in your measurement... even after applying some IR windowing and smoothing, you've got several phase rotations and an unusually high GD -- it could probably be just the room.

I also don't know what to make of this strange peak prior t=0:

Envelope Time Curve
View attachment 172440
*note the strong reflection around 3ms which may be adding to the noise

this is how it looks now. I did notice there seems to be an issue with the analog cable going into the preamp, sometime when I turned it around I got no sound at all, but I only noticed this on my last measurement. I should replace it or try and use the usb input on the preamp next time. Maybe that could explain some issues?

 

[RayDunzl]

I also don't know what to make of this strange peak prior t=0:

It may be "distortion".

Impulse Graph

Scroll down to "Distortion Components".

 

[Fredd3]

It may be "distortion".

Impulse Graph

Scroll down to "Distortion Components".

new measurement with usb cable instead. so this was in stereo, but looks like no distortion this time, right?

 

[Fredd3]

Is it not strange that I get almost no bass? Below 90 Hz the curve falls down soo fast.

 

[Backflash]

Don't know how your input device works exactly, if it even recognizes same way USB mic does, but could try this.

UMIK-1 + Windows Audio Enhancement

Dear All, We've heard from some of you that Windows update is enabling Audio Enhancements on microphone inputs by default for their new builds. Features like Acoustic Echo Cancellation (AEC), Active Noise Compensation (ANC) will create some issue with your measurements. You can check your PC...

www.minidsp.com

 

[Fredd3]

Don't know how your input device works exactly, if it even recognizes same way USB mic does, but could try this.

UMIK-1 + Windows Audio Enhancement

Dear All, We've heard from some of you that Windows update is enabling Audio Enhancements on microphone inputs by default for their new builds. Features like Acoustic Echo Cancellation (AEC), Active Noise Compensation (ANC) will create some issue with your measurements. You can check your PC...

www.minidsp.com

hi, it does work the same way. I will try this, thanks!

 

[Fredd3]

I changed the mic to use channel L instead of R. that gave proper bass in the data.