EQ Advice

[Logan Nolag]

So I just picked up an RME 802 FS AE and I’ve been experimenting with its built in Room EQ feature. After several hours of testing in REW this is my best result. I have a 2.2 system with symmetrically placed subs just to the left and right of the main speakers.

The purple trace is with no eq just a high/low pass at 80hz. The blue trace is the fully EQed result.

My biggest issue is right around 80hz. I was able to get my left speaker and left sub to integrate pretty well but my right speaker seems to produce a massive null around 80-90 hz all on its own even without subs.

The massive peak right next to the 80hz null is also coming from my right speaker. When I combine the left and right subs together with the right speaker it results in that double null on either side of 80hz.

I tried literally every combination of polarity and phase in my svs subs. I took measurements from 0-180 degrees phase in 10 degree steps with polarity set to both positive and negative and then set the subs to the result that integrated best but no matter the phase or polarity I couldn’t completely eliminate the null and this was the best I could do.

I was able to eliminate the null and peak in the right speaker by running no crossovers on the subs or speakers at all but it would come back after applying eq.

 

[SigbergAudio]

Can you upload the .mdat file from REW?

 

[Logan Nolag]

Can you upload the .mdat file from REW?

Sure. I had to split the measurements into two files because of the upload limit.

 

[Keith_W]

My biggest issue is right around 80hz. I was able to get my left speaker and left sub to integrate pretty well but my right speaker seems to produce a massive null around 80-90 hz all on its own even without subs.

I had a look at your measurements - I saw that huge 80Hz null in your right speaker (absent in your left). This is post EQ:

Red: L sub + L speaker
Green: R sub + R speaker
Black: L+R subs with L+R speakers (I staggered the SPL for clarity)

Note that the huge 80Hz null in the right speaker/sub is compensated by the left speaker/sub. The end result does not look too bad, although there is still a fairly wide "double dip" between 60-150Hz that is still visible after applying ERB smoothing (not shown).

You could probably obtain a better result if you used a filter design tool like MSO. If you were curious you could try your hand at designing that filter manually, but it's much easier to simply use that tool.

 

[janbth]

@Logan Nolag Are you using bass management? What are the crossovers? It looks like you have a stereo-sub setup with L-sub integrated with L and R-sub integrated with R. If so, I would suggest you try a "regular" bass managed setup: Integrate L+sub with R-sub to play well together first, then integrate this sum with L+R. I think this will make it easier to EQ/smooth out the 80-90 Hz region.

 

[Logan Nolag]

@Logan Nolag Are you using bass management? What are the crossovers? It looks like you have a stereo-sub setup with L-sub integrated with L and R-sub integrated with R. If so, I would suggest you try a "regular" bass managed setup: Integrate L+sub with R-sub to play well together first, then integrate this sum with L+R. I think this will make it easier to EQ/smooth out the 80-90 Hz region.

Yeah I have two subs. I’m using the RME to cross them over at 80hz.

I followed the instructions on their forums to make a 24db crossover by stacking 2 eq bands set to a q of 0.7. The instructions also say to invert the main speakers polarity but that just seemed to make things worse.

And yes I tried to integrate the left sub with the left speaker and the right sub with the right speaker.

These measurements were done with the subs set to positive polarity and the phase at 30. I arrived at those settings by brute force measuring my right speaker+right sub at every combination of positive and negative polarities between 0-180 in 10 degree steps.

I actually just tried another experiment where I measured everything together also with the subs set to both positive and negative and the phase set from 0-180 in 10 degree steps and the best result turned out to be positive polarity and 0 phase. Those settings pretty much eliminated the null around 90hz.

 

[radiomike]

The null is visible in both the EQ'd and non EQ'd measurements for both your speakers and subs. It's slightly shifted up in frequency for the sub.
In my opinion, it's probably it's a function of your room + speaker location + measurement location. A diagram of your room indicating dimensions and the locations of speakers and measuring location might be helpful here.
It seems very unlikely that it's a processing issue, but you could rule this out by reversing the outputs of the RME or your speaker outputs. Then re-measuring. If the null moves with the change, then it's a processing thing, if it stays the same, then it's acoustic.

 

[Sokel]

Apart from EQ, lows should be much better with a 5dB increase, the way it is now sounds thin even to the eye.

 

[Pareto Pragmatic]

In my opinion, it's probably it's a function of your room + speaker location + measurement location. A diagram of your room indicating dimensions and the locations of speakers and measuring location might be helpful here.

That would be my guess as well, multiple factors in play adding up. Though given the crossover implementation, maybe something is a bit off there? Stacking, phase manipulation, lots of potential issues there.

For sure above 3k looks very phase-y to me, on the 1/12 graph. This can show up from having something close behind the mic, like a seat/couch back. But if that's not it....

I suggest a MMM approach to see what's what. Take 9-12 measures of the entire system, above/below/at the listening height/left/center/right of listening position, sum them, work from that. Then come back in with that as a starting point for tweaking a single position.

 

[janbth]

Yeah I have two subs. I’m using the RME to cross them over at 80hz.

So, you are crossing over right at the problematic region. Could be you are having a crossover problem, possibly in addition to a room mode problem around that frequency. I would try to move the crossover to see how things change.

... Those settings pretty much eliminated the null around 90hz.

Just a comment about terminology - not directed only to you, but it's a common mistake made by a great number of people here on ASR: A null is a region in space (i.e. your room). It's where the pressure is constant for a given room mode. If you put both a speaker and your measurement mic in a null it will lead to a dip in the frequency response. On the other hand, a high pressure zone is a region in space where the pressure fluctuates up and down. Put your speaker and mic here, and it will lead to a peak in the response.

Here is a quick diagram generated by Claude to illustrate nulls and high pressure zones. The example shows a snapshot in time of the standing wave pattern of the (1,1,0) mode:

 

[Logan Nolag]

Apart from EQ, lows should be much better with a 5dB increase, the way it is now sounds thin even to the eye.

I balanced the subs to the mains with the generator and the spl meter and they were the same. Are you really supposed to increase the subs over the mains by that much? Or is it just an adjustment to taste thing?

That would be my guess as well, multiple factors in play adding up. Though given the crossover implementation, maybe something is a bit off there? Stacking, phase manipulation, lots of potential issues there.

For sure above 3k looks very phase-y to me, on the 1/12 graph. This can show up from having something close behind the mic, like a seat/couch back. But if that's not it....

I suggest a MMM approach to see what's what. Take 9-12 measures of the entire system, above/below/at the listening height/left/center/right of listening position, sum them, work from that. Then come back in with that as a starting point for tweaking a single position.

That would make sense it is a desk setup and I was sitting in my chair directly behind the measurement mic which was between me and my laptop on a tripod. The mic was in the exact spot where my head typically goes and I was scooted back a bit behind my normal listening position when I made the measurements. I actually tried moving the mic and when I put it on the desk behind my laptop the 80hz dip is completely gone.

So, you are crossing over right at the problematic region. Could be you are having a crossover problem, possibly in addition to a room mode problem around that frequency. I would try to move the crossover to see how things change.


Just a comment about terminology - not directed only to you, but it's a common mistake made by a great number of people here on ASR: A null is a region in space (i.e. your room). It's where the pressure is constant for a given room mode. If you put both a speaker and your measurement mic in a null it will lead to a dip in the frequency response. On the other hand, a high pressure zone is a region in space where the pressure fluctuates up and down. Put your speaker and mic here, and it will lead to a peak in the response.

Here is a quick diagram generated by Claude to illustrate nulls and high pressure zones. The example shows a snapshot in time of the standing wave pattern of the (1,1,0) mode:
View attachment 556005

Didn’t realize that was the wrong terminology I’ll make sure to use dip from now on. I tried other crossovers and nothing really changed it other than allowing the subs to crossover around 150hz or more and keeping the mains lower but that seemed weird to have the subs playing that high. Also even with the sub low pass set high the dip came back after I applied EQ.

 

[Logan Nolag]

So I took all the advice from the thread and tried again. I was pretty much able to mostly clean up the crossover dip and I boosted the sub volume a little bit although that accentuated the remaining dips a bit more than before. I think this is probably the best I can get without rearranging the room or adding room treatments. The purple trace is the new measurements.

 

[Sokel]

So I took all the advice from the thread and tried again. I was pretty much able to mostly clean up the crossover dip and I boosted the sub volume a little bit although that accentuated the remaining dips a bit more than before. I think this is probably the best I can get without rearranging the room or adding room treatments. The purple trace is the new measurements.

View attachment 556618

Everything is within -/+ 3dB (10dB looks but I don't consider peaks) as it seems, declining slope at highs, not bad at all!

(a 50dB chart would be more helpfull as it's a kind of a norm here)

 

[Logan Nolag]

Everything is within -/+ 3dB (10dB looks but I don't consider peaks) as it seems, declining slope at highs, not bad at all!

(a 50dB chart would be more helpfull as it's a kind of a norm here)

I'm not sure exactly what a 50dB chart is but looking at the post from Keith_W where the scale goes from 50dB - 100dB I assume thats what you mean. Here's the new chart readjusted to that scale with the average line added in.

 

[Sokel]

I'm not sure exactly what a 50dB chart is but looking at the post from Keith_W where the scale goes from 50dB - 100dB I assume thats what you mean. Here's the new chart readjusted to that scale with the average line added in.

View attachment 556940

That's what I meant, yes.

There's some short of modulation after 3kHz or so, do you know how to gate measurements in REW?
It would be helpful to see if it's the room or something else.

Here's a quick way to remove some room up-high, just follow the steps I, II, III and reverse the two later ones to come back to the original room+speaker measurement:

Edit: you can also adjust any scale putting the mouse at the outer numbers using the mouse wheel or going at the upper right corner of the chart adjusting its limits.
Edit2: now that I see it's a L+R measurement, modulation may just be a few mm of mic misplacement from been dead at the center, so no big deal.

 

[Logan Nolag]

That's what I meant, yes.

There's some short of modulation after 3kHz or so, do you know how to gate measurements in REW?
It would be helpful to see if it's the room or something else.

Here's a quick way to remove some room up-high, just follow the steps I, II, III and reverse the two later ones to come back to the original room+speaker measurement:

Edit: you can also adjust any scale putting the mouse at the outer numbers using the mouse wheel or going at the upper right corner of the chart adjusting its limits.
Edit2: now that I see it's a L+R measurement, modulation may just be a few mm of mic misplacement from been dead at the center, so no big deal.

Here are the Left Speaker, Right Speaker and Both Speakers together with the MTW Applied.

 

[singular]

Both Speakers together with the MTW Applied.


View attachment 556984

It is not gated to your listening distance, is it?

 

[Sokel]

It is not gated to your listening distance, is it?

Maybe it's not, but MTW default settings are tight enough from 1kHz upwards.
It should smooth it.

Something else is at play here.

 

[singular]

Maybe it's not, but MTW default settings are tight enough from 1kHz upwards.
It should smooth it.

Something else is at play here.

Desk setup, front wall, subs under desk, rear wall not damped?

 

[Sokel]

Desk setup, front wall, subs under desk, rear wall not damped?

Probably right, never thought of desk.