• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

UMIK-2 First time Room EQ measurements.

  • Thread starter Deleted member 35357
  • Start date
D

Deleted member 35357

Guest
Just got interested into Room EQ & doing measurements so quickly got hold of a new UMIK-2 to pair it with REW to get started with some simple room measurements for equalization for some hopeful better sound results.

Wanted to get this started on my old setup to see how it performs and then get any possible recommendations on what needs to be done to upgrade this setup I have with better speakers or just EQ if it was only thing needed. But seeing my initial results with the UMIK-2 I might just need to get a new pair of speakers overall + potential other things.

But before that it's good to get familiarized with REW & PEQ with APO for better in room results. I read the tutorials over here and watched some videos etc to get started.

Measured my practice tuned system with a improper microphone & placement (to see how far off those would have given a result) and then started over with doing measurements for a 20hz-25Khz sweep and try to correct it some with adjusting sub gain & doing some quick equalization using the 7-band Analogue EQ I have between my DAC (Topping EX5) and my powered speaker system. (5.1channel setup for only 2 channel Audio)
It's a simple thing, the Analogue EQ, it's useful for some coarse adjustments but not much. (does help a little as I have other sources than my PC that can't use EQ with current setup)

UMIK-2 showed that I had my SUB gain set way to high. (thanks bad improper other mic)
While the Frequency response elsewhere was quite all over the place. Nothing flat anywhere.

Tried to correct as best as I could but the 7-band EQ isn't reaching where it should. Didn't help much.
221216 UMIK-2 initial measurments.png

Gray line is what a mess it makes was after using a improper mic to measure.
Black Line is what I could do with adjusting SUB GAIN & the 7-band EQ to straighten it up as best as I could.
I could not do much with the peaks & valleys than this with that.
the 10Khz peak could not be flattened with the 7-band EQ.

I then tried out the EQ that REW will spit out. (going for a 74dB target from the black line measurement)
221216 UMIK-2 initial EQ prediction.png


The prediction was really close to the actual results from this REW calculation EQ:
221216 UMIK-2 initial EQ results.png



Where does one go from here is the question, keep going and just try for better results?
I tried to measure the final results from a few positions around the listening position and got quite different results from where I had the UMIK-2 when I did the corrections.

Seems I didn't place the mic in a proper position as the included stand wasn't a good fit to use in the chair and decided to place it as close as possible to it on the desktop. It wasn't optimal.

Holding the mic around the area where I sit had lots of variance to the earlier results:
221216 UMIK-2 initial EQ results mic posistion change.png

Green is on table, while the blue lines are for holding the UMIK-2 around the seating position after everything was done.

Anything I should know? Missed something critical here or there?
I did measure all speakers at the same time doing all the measurements. Doing individual speakers isn't optimal as the SUB is most likely only getting left or Right channel information while the centre speaker is outputting the other side. (just using duplicate L & R channels)
Thought it would be best.

Waterfall before EQ is applied:
221216 UMIK-2 initial waterfall before PEQ.png
 

staticV3

Master Contributor
Joined
Aug 29, 2019
Messages
7,536
Likes
12,004
Anything I should know? Missed something critical here or there?
1. Don't do single sweep measurements for room EQ. They're useless. Instead, do MMM. YouTube guide 1, guide 2, REW settings:
image.png
(set low and high cut according to your system)

2. Don't EQ to a flat response. Speakers in a room should slope down slightly. About 5dB from 100Hz to 10kHz is a good place to start:
Genelec 8341A SAM™ Studio Monitor Powered Speaker CEA-2034 Spinorama Predicted In-room Respons...pngNeumann KH-420 G Measurements Predicted in-room Frequency Response anechoic Active 3-way studi...pngRevel F328Be CEA-2034 Spinorama Predicted In-room Frequency Response Measurements.png
 
OP
D

Deleted member 35357

Guest
I did measure the mic in seating position holding it straight in front of my eyes, and then I went back and held the mic from behind my seat in my head position.

They are the "after EQ" purple and blue lines supplied above versus the green on table measurements.
 
OP
D

Deleted member 35357

Guest
What are your speakers by the way?
Some old Altec Lansing 5.1 system., (THX) ADA995. (In 2 channel mode)

I'm planning to upgrade it but wanted to use the time to learn this stuff with it and see what kind of upgrade is reasonable to get and have then go back and reference from what results the previous (this system managed).

I'm thinking of getting some Genelec or Neumann, the smaller kind if they are a good choice as replacements. But it might take some time before I decide what I want to go with. not sure how much I want to spend yet on it, if only going for a 2.1 channel thing or get 5.1 straight away etc.
But that's several months from now most likely.
 
OP
D

Deleted member 35357

Guest
Can you upload the measurements some where? I want to have a look at them.

I hope this works:

*removed link as they were bad measurements* (100MB)

I did several measurements that are kinda redundant as I was adjusting the SUB Gain & 7-Band EQ. Comments might not be to helpful to you to sort it out.
 
Last edited by a moderator:

staticV3

Master Contributor
Joined
Aug 29, 2019
Messages
7,536
Likes
12,004
OP
D

Deleted member 35357

Guest
But I'm missing on how to actually do the measurement?
Do you do 32x single measurements and then somehow average them or is there a automated process? (I didn't get through the document yet to where they show how to actually do it)
All just discussed the principle, not the practical details, of what I managed to read yet.

It's late around here so can't do more measurements in the middle of the night.
Will have to measure tomorrow if I find the practical details by that time.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,424
Likes
7,940
Location
Brussels, Belgium

staticV3

Master Contributor
Joined
Aug 29, 2019
Messages
7,536
Likes
12,004
But I'm missing on how to actually do the measurement?
I sent you a PDF with detailed guidelines, two separate "How to MMM" videos on YouTube, and a screenshot showing the correct generator and FFT settings for REW.
 
OP
D

Deleted member 35357

Guest
They're not useless. Far from it.

But currently it's a mess with nothing labeled. I have no idea what's going on. @Nighthog
Measurement nr 4 is basically the start line from where to begin. (I was looking if 192K, 48K sampling had a difference of measurement)
I decided to use 192K for everything.
All the unlabelled ones below are me adjusting the 7-band EQ to see what it does (learning it's effects on the results)
Only the nr. 17 result is of importance as I was mostly done adjusting it and tried out changing the speaker locations a little, not much, but tried different directions mostly. (the ones saying speaker position, nr, 18-21)
It didn't have much effect as I wanted so I tried adjusting the EQ again a little with nr. 22-24.

I used measurement nr.25 for the waterfall & EQ, I tried it as mainly a test-drive on how it works out.
Nr.26 shows the result of the EQ REW spit out for measurement nr.25. (prediction is close to the actual result)

Nr. 27 I placed the mic in my face to measure the difference from where I measured everything else.
Nr. 28 I went behind my chair and held the mic in then position my head would be when leaning back in the chair watching movies listening music etc.
 
Last edited by a moderator:
OP
D

Deleted member 35357

Guest
I sent you a PDF with detailed guidelines, two separate "How to MMM" videos on YouTube, and a screenshot showing the correct generator and FFT settings for REW.
There was another video around that I found that shows where you access the settings you wanted me to use.
Between reading documents, answering questions, watching a video or two + providng a upload gets you distraced on where you find the options you wanted to be used.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,424
Likes
7,940
Location
Brussels, Belgium
I checked measurement number 3 in REW and here is what I recommend. The ones with 0 dB values need to be tuned based on what you hear with pink noise playing as what we measure above 100Hz is not really what we hear. Use negative values only. Keep decreasing the value till the offending frequencies are not present anymore in the pink noise.


PK Fc 7000 Hz Gain 0 dB Q 2.000
PK Fc 492.0 Hz Gain 0 dB Q 2.000
PK Fc 60.40 Hz Gain -12.70 dB Q 2.690
PK Fc 81.20 Hz Gain -4.10 dB Q 5.520
PK Fc 103.5 Hz Gain -6.20 dB Q 7.092
PK Fc 122.0 Hz Gain -6.60 dB Q 4.690

I don't think MMM measurements are necessary. Good luck!
 
OP
D

Deleted member 35357

Guest
Did MMM according to the settings provided earlier with pink noise, but they are at to low levels. (Will have to redo tomorrow at higher SPL, can't do higher than this at the moment without complaints)

This is with sitting in chair holding the Mic in face area moving it around and then being behind the seat and holding the mic around the head area in generality as I moved it around.
Somehow it was averaged 500+ measurements, it did them quicker than I thought.


MMM.png
 
Last edited by a moderator:
OP
D

Deleted member 35357

Guest
I checked measurement number 3 in REW and here is what I recommend. The ones with 0 dB values need to be tuned based on what you hear with pink noise playing as what we measure above 100Hz is not really what we hear. Use negative values only. Keep decreasing the value till the offending frequencies are not present anymore in the pink noise.


PK Fc 7000 Hz Gain 0 dB Q 2.000
PK Fc 492.0 Hz Gain 0 dB Q 2.000
PK Fc 60.40 Hz Gain -12.70 dB Q 2.690
PK Fc 81.20 Hz Gain -4.10 dB Q 5.520
PK Fc 103.5 Hz Gain -6.20 dB Q 7.092
PK Fc 122.0 Hz Gain -6.60 dB Q 4.690

I don't think MMM measurements are necessary. Good luck!
Sadly I had adjusted the GAIN on the SUB later so these values aren't usable for current settings on SUB & the Analogue EQ.
Sorry for the mess with the data.

Thanks for the work though. It wasn't in vain entirely!
Might need to provide better measurements before you do it again.

Figured out how to do the MMM thing now.
 
OP
D

Deleted member 35357

Guest
Yeah the measurements looked waaaaaaay to bassy.

I will gladly take a look with something a bit more organised.
Watched Dune with those measurments...
Rumble rumble. Was sure something went wrong. Though it was nice at times but quick volume adjustment between hearing dialogue and when the rumble started was not fun.
 
OP
D

Deleted member 35357

Guest
1. Don't do single sweep measurements for room EQ. They're useless. Instead, do MMM. YouTube guide 1, guide 2, REW settings:
View attachment 250503
(set low and high cut according to your system)

2. Don't EQ to a flat response. Speakers in a room should slope down slightly. About 5dB from 100Hz to 10kHz is a good place to start:
View attachment 250504View attachment 250505View attachment 250506
Should I use all speakers when doing it?
The videos tell to do so, but the settings you had was for only Left channel pink noise in the generator settings.

And is the slope just a preference target? Seems people in general use their own curve from what I saw around with variance.
Are there reference targets, or is it choose to taste?
 

staticV3

Master Contributor
Joined
Aug 29, 2019
Messages
7,536
Likes
12,004
Should I use all speakers when doing it?
I'd do each speaker separately. If you can't EQ your channels separately, you can still AVG all measurements and use that to design your EQ.
But having measurement for each speaker might reveal some flaws that would be obscured when running them all at the same time.

And is the slope just a preference target?
It's the predicted response of an anechoically flat speaker with wide directivity in a typically reflective room.
Here are some preference targets that you could try:
https://www.audiosciencereview.com/...ker-target-responses-in-csv-txt-format.16401/
 
Top Bottom