• 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!

Room EQ, do's and dont's

MRC01

Major Contributor
Joined
Feb 5, 2019
Messages
3,424
Likes
4,030
Location
Pacific Northwest
What kind of mics are you all using for these measurements? My Rode NT1As are fine for music but I'm finding they're not accurate enough for these measurements and I don't have customized correction curves for them. I could use some suggestions - thanks!
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
After time aligning both responses and setting IR window to 3.5ms, as that is when first noticeable reflection peak occurs:

Cheater...

My phase was displayed with a 500ms right window (default).

With 3.5ms it looks almost flat here.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,498
Location
Monument, CO
What kind of mics are you all using for these measurements? My Rode NT1As are fine for music but I'm finding they're not accurate enough for these measurements and I don't have customized correction curves for them. I could use some suggestions - thanks!

Most mics are not accurate enough for measurements, intentionally so, with rolled-off bass and colored midrange to do things like emphasize vocals or certain instrument frequency bands. My friend has a Shure 81, generally considered a very flat mic, but it wasn't close to my Earthworks M30 when we compared them.

I use mostly a CSL-calibrated UMIK-1. It's frequency response after cal is very close to my Earthworks M30, a much more expensive mic that also requires a preamp and all that jazz. The UMIK-1 cannot match the dynamic range, low self noise, and such of the M30 but for about $100 USD the UMIK-1 plus REW (free) provides me with equal or greater room measurement capability than the $10k+ USD system of M30, preamp, and professional SW I was using for years before. And it is much easier/faster to set up and use.
 
OP
Krunok

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Cheater...

My phase was displayed with a 500ms right window (default).

With 3.5ms it looks almost flat here.

Lucky you - I obviously have lot of reflections, which is no wonder as my second bass driver is firing vertically to the floor. :D

I followed the guideline posted here. If I don't apply gating phase curves just keep going down forever.. :(

I'm not sure what should I do now with these phase graphs. Any advice?
 
OP
Krunok

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
What kind of mics are you all using for these measurements? My Rode NT1As are fine for music but I'm finding they're not accurate enough for these measurements and I don't have customized correction curves for them. I could use some suggestions - thanks!

Daytonaudio iMM-6 calibrated measurement microphone.
 

Biblob

Addicted to Fun and Learning
Joined
Sep 13, 2018
Messages
633
Likes
600
Let me say this thread has been very helpfull to me already. Many thanks for sharing all your experiences and knowledge. One of the reasons I keep coming back here.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
If I don't apply gating phase curves just keep going down forever..

That's what I see with my JBL, Infinty, and other's speakers measured mid or far field... The MartinLogans measure ungated (500ms) like the others measure with tight gating.

The JBL throws a pretty flat response measured nearfield, so, with the limited data I have, can only come up with wide dispersion and the bounces/interactions with the room boundaries as a reason for them not to project that quality on to the listening position, when measured with a long time window on the right side of the impulse.

I'm not sure what should I do now with these phase graphs. Any advice?

I've lost the plot here...

Do you have a reason for the difference left/right at the low frequencies?

For me, it's a non-rectangular room. Left speaker fires into a bigger space - the left rear corner of the room is open to breakfast/kitchen areas.

Do you hear a problem? I didn't.
 
OP
Krunok

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
That's what I see with my JBL, Infinty, and other's speakers measured mid or far field... The MartinLogans measure ungated (500ms) like the others measure with tight gating.

The JBL throws a pretty flat response measured nearfield, so, with the limited data I have, can only come up with wide dispersion and the bounces/interactions with the room boundaries as a reason for them not to project that quality on to the listening position, when measured with a long time window on the right side of the impulse.

Oh, so it was you who cheated with your MLs! :D

And what is automatic correction doing with such falling phase curve? Does it look the same after the correction or autmoatic correction somehow manages to do something with it?

I've lost the plot here...

Do you have a reason for the difference left/right at the low frequencies?

For me, it's a non-rectangular room. Left speaker fires into a bigger space - the left rear corner of the room is open to breakfast/kitchen areas.

Do you hear a problem? I didn't.

I don't hear any problems. LF amplitude of both speakers measures ok when I measured them separately and together. I can easilly follow bass guitar lines in every song I listen which was previously not the case and the drum bass is well defined and not boomy. I am simply asking, as step response is less than perfect, is there something I should do about it and how, or I should just leave it as it is?
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
And what is automatic correction doing with such falling phase curve? Does it look the same after the correction or autmoatic correction somehow manages to do something with it?

On both ML and JBL it brings the woofers forward in time a little.

Can't do too much, there aren't enough taps available to me for resolution or (?) or time adjustment.
 
OP
Krunok

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
On both ML and JBL it brings the woofers forward in time a little.

Can't do too much, there aren't enough taps available to me for resolution or (?) or time adjustment.

Or it doesn't do anything at all.

From what I have understood your MLs anyhow have flat phase response so no need for correction there. On the other hand, I can hardly imagine an algorithm facing that never ending phase curve of your JBLs and my Castles and doing something usefull about it. Unless, of course, you have phase response of JBLs before and after correcction to prove me otherwise. I mean.. there are many room EQ vendors claiming they are correcting phase as well but I don't recall seeing any independently measured proof of it.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
From what I have understood your MLs anyhow have flat phase response so no need for correction there. On the other hand, I can hardly imagine an algorithm facing that never ending phase curve of your JBLs and my Castles and doing something usefull about it. Unless, of course, you have phase response of JBLs before and after correcction to prove me otherwise. I mean.. there are many room EQ vendors claiming they are correcting phase as well but I don't recall seeing any independently measured proof of it.


Blue - JBL before/after - 6ms window
Green - ML before/after - 500ms window
Red - ML before/after - 6ms window

1551379229444.png
 
OP
Krunok

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Blue - JBL before/after - 6ms window
Green - ML before/after - 500ms window
Red - ML before/after - 6ms window

View attachment 22828

Thanks for posting this Ray!

Let's put aside your MLs who obviously have natural tendency of flat phase response, but can you please post before and after step response that goes with this phase correction of your JBLs?

Btw, is this based on both speakers response on sine sweep? Have you ever checked if this result with your JBL speakers is consistent in terms that it can be repeated with similar results if you repeat the measurement?

Also, are you able to hear the difference between corrected and uncorrected phase reponse?

I must admit I'm a little bit sceptical about this "time/phase correction" of room eq software. I think it is a little bit stretching into snake oil teritorry as I find it hard to believe phase can be accurately measured in home environment.
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
Let's put aside your MLs who obviously have natural tendency of flat phase response, but can you please post before and after step response that goes with this phase correction of your JBLs?

Impulse

Blue - with correction, at the listening position, both speakers

1551381669399.png


Step

1551381599519.png


This could be a little unfair or misleading - the JBL are located above the listening position, not at ear level, but AcourateDRC corrects the LP response, so...
 
OP
Krunok

Krunok

Major Contributor
Joined
Mar 25, 2018
Messages
4,600
Likes
3,065
Location
Zg, Cro
Impulse

Blue - with correction, at the listening position, both speakers

View attachment 22830

Step

View attachment 22829

This could be a little unfair or misleading - the JBL are located above the listening position, not at ear level, but AcourateDRC corrects the LP response, so...

Well, no doubt corrected looks better!

Do you get similar graph when you repeat the measurement?
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
Do you get similar graph when you repeat the measurement?

Yes. Well, some additional information below.

24 measurements at different SPL:


Impulse

1551383548152.png


Step:

Gets a little confused with the different levels. The four outliers are the lowest SPL, too close to the noise floor. Once well aove the noise floor, the results are rather congruent.

1551383757876.png



SPL

Ambient noise intrudes at low levels at low frequencies

1551383826616.png


Distortion is level vs noise dependent, too. Only at the highest SPL do the distortion measurements begin to rise out of the noise (Low SPL at the top (high distortion measured is actually just noise), high SPL at the bottom, where the THD measures begin to bunch up, and rise, with higher SPL, and some actual meaningfult THD from the speakers)

1551384147425.png
 
Last edited:

FrantzM

Major Contributor
Forum Donor
Joined
Mar 12, 2016
Messages
4,339
Likes
7,737
Can we make of this thread a "sticky"?
WoW just WoW!!
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,986
Location
Riverview FL
Then there's the off-axis situation:

Step response when about 3 feet to the left of the listening position, both speakers playing, right speaker is delayed (more distant), the left speaker is the reference;

1551385931461.png


And impulse:

1551385956341.png
 

mitchco

Addicted to Fun and Learning
Audio Company
Joined
May 24, 2016
Messages
640
Likes
2,397
@RayDunzl looking good Ray!

@Krunok I tried a time domain experiment here: https://audiophilestyle.com/ca/ca-a...tereo-mains-using-audiolense-r712/#timedomain I did find I could tell the difference under blind conditions. But it is not the fairest as there are three parameters being adjusted for there, time alignment of drivers, excess phase correction, and major low frequency reflections being taken care of.

Having tried it a few different ways, I find linearizing the phase of a passive XO system, in which the drivers were not time aligned was next to impossible to hear a difference. Time aligning the drivers (using linear phase XO) and linearizing the phase was audible. Then correcting for low frequency room reflections improved the clarity of the bass response. Note Audiolense is the only DSP software I know of that corrects for major room reflections (usually the front wall and rear wall first reflection).

Given the power and precision of modern audio DSP, I hope for more research in this area with controlled subjective listening tests.
 
Top Bottom