Is this a 'good/decent/could be better' room equalization?

[RDC]

Hi Guys,

I'm fairly new at this whole room equalization and treatment concept. I think I got a pretty good room response now, I could be wrong, so I would love to hear what you think.
Measurements taken with REW and a calibrated mic. REW file available here (zip).

Room dimensions are 21 x 11.5 x 9 feet (6.5 x 3.5 x 2.7 meters). I've got diffusion panels behind the speakers en behind the listening point. I also have 4" thick absorbers behind the listening point. Rug on a wooden floor. No side wall, ceiling treatments or bass traps possible. ML Montis and SVS SB-2000 Subwoofer with a crossover and PEQ (IIR-type) set via a MiniDSP DAC. I still need to try some other crossovers, as I only tried it at it's current setting of an (probably unwise) arbitrary 100Hz.

SPL 10-200 Hz


SPL 20-20000Hz


Waterfall 10-200Hz


Waterfall 20-20000Hz


I'm not sure why the images are getting smaller every time...must be a forum thing, because they are all the same size. But they might also be attached, not sure how this works on here.

Anyway, thanks for your time!

RDC

 

[Speedskater]

Lots of experts don't think much of room EQ.

 

[RDC]

I'm not sure what I can do with that comment alone.
Why don't they think much of room EQ?
And who are these experts?

But maybe I forgot a crucial part, the base measurement, so you can see what is altered.


FYI: the low frequencies volume are mostly tamed by volume adjustment to the subwoofer and the XO in the final result.

 

[twich54]

RDC, hopefully Jonathan will chime in here for he has a very good take on all of this.

 

[JonFo]

And As surely as a moth is drawn to flame, a thread with 'room equalization' in the title will see me pop-up

RDC, thanks for sharing the source measurement files, helpful to see greater detail and things like the group delay metrics.

First, a couple of questions that you'll hopefully see today.

What is the wiring scheme relative to the new miniDSP box?

Last we discussed your setup, it was not possible to have an external active crossover remove the low's from the Montis when using the PrimaLuna in 2ch mode.

If the miniDSP is being fed by the mono output of the Prima Luna, then you have a doubled bass scenario, as the Montis is still reproducing content below 100Hz. That might explain the huge hole at 40Hz as likely the phase of the Montis woofers is opposite the SVS at that frequency. Could also be a nasty room mode, but only a separate measurement of the Montis alone would tell us that. So if you have a chance, grab that measurement as well.

 

[JonFo]

Phase is critical

OK, my intuition was right, as shown in the graph of the phase relative to SPL, we immediately see sharp phase swings at the xover point (100Hz), which indicates the distance (delay) and phase need to be tweaked.
This is why there is an SPL dip at the xover and a mighty huge suckout at 40Hz.



The Group Delay graph is extremely helpful in fine tuning delays between speakers, one wants as smooth a GD plot as possible. I spent an hour tweaking delays for my mid-bass modules the other day, and much of that was spent looking at GD.
Your GD graph also highlights the fact the timing is off at the xover point, that bump at 100Hz could be flattened out by setting the correct delay values in the miniDSP.

 

[JonFo]

Spectrogram shows resonances

The Spectrogram as well as the waterfalls show significant resonances below 40Hz, the room is over-energized there. So some EQ below 40 to cut that would help.
If even after EQ there is still significant resonance, then large (and many) bass traps will be required.

This view also shows the aforementioned xover/phase issues at 40 and 100Hz. Also shows a weak mid-bass range between 100 and 200Hz, possibly from the sub canceling.



The strands pre-and post signal at 50Hz confirm you live in a country with 50Hz mains Likely from a fridge or AC motor (it's hot over there this year I hear). But could be some hum from the valve amp.

 

[amey01]

I really want to hear your system Jon.

 

[RDC]

Hi Jonathan, thanks for your valuable feedback!

What is the wiring scheme relative to the new miniDSP box? [...] If the miniDSP is being fed by the mono output of the Prima Luna, then you have a doubled bass scenario, as the Montis is still reproducing content below 100Hz. That might explain the huge hole at 40Hz as likely the phase of the Montis woofers is opposite the SVS at that frequency. Could also be a nasty room mode, but only a separate measurement of the Montis alone would tell us that. So if you have a chance, grab that measurement as well.

I do not use the Primaluna subwoofer-out anymore in this scenario. The source DAC delivers the signal to the analog ins of the MiniDSP. There I feed both L&R-input signal to the sub output, so it gets the full range combined L&R input signal. To prevent overlap in frequencies I choose in the MiniDSP to XO between L&R and Sub at 100Hz (24dB slope). So I applied the XO-over to both L&R and Sub. This should result is having no overlap in bass.


I'll grab the sub and Montis-only measurements without and with EQ later today.

[...] we immediately see sharp phase swings at the xover point (100Hz), which indicates the distance (delay) and phase need to be tweaked. This is why there is an SPL dip at the xover and a mighty huge suckout at 40Hz. [...]
Your GD graph also highlights the fact the timing is off at the xover point, that bump at 100Hz could be flattened out by setting the correct delay values in the miniDSP.

If I understand you correctly, I will not change the phase (set at 0 deg) on the sub, but just add a delay to the sub-output on the miniDSP to flatten out the bump at 100Hz in the GD graph.

So some EQ below 40 to cut that would help. If even after EQ there is still significant resonance, then large (and many) bass traps will be required.

I already applied EQ over that range to the sub.

Jonathan, do you think I need to try out other XO-points (and possibly slopes)? Or is 100Hz (24dB slope) already good enough as an XO-point?

I'll add the new measurements and REW-files when I get time to measure again.

Thanks again!

 

[RDC]

I did some changes and measurements. REW-files can be found here. The zip contains the following REW measurements:

• Aug 19 14_03_49 - Full range LR (Montis) only - no PEQ - no XO
• Aug 19 14_07_30 - Sub only - no PEQ - No XO
• Aug 19 14_12_23 - Base - full range LR+SUB - no PEQ - No XO
• Aug 19 14_56_38 - Base Measurement MiniDSP Prof 2 - XO 90Hz - PEQ ON

Looking at both of the graphs of 'only the sub' and 'only the speakers', I decided to use a XO-point at 90Hz as they seem to cross each other around that point. That flattens the respons a bit around that range, but still has a dip at 92Hz, but less than at 100Hz. I might try lower XO-points later on and see what changes. I changed the phase, directly on the sub, taking measurements each time to look at the changes in the GD and SPL graph. I got it down quite a bit, without getting too much rise at others frequencies. I haven't added a delay in the miniDSP yet. All the measurements of today are made with this phase setting at the sub.

Previous measurement - GD


Change after changing phase on sub


Not sure how it happened, but now I've got a huge peak at 52.5Hz. In the 'only sub' and 'only speakers' SPL graphs you can see that both have a peak there. Is this a room mode? Any idea how to get rid of that? I didn't have that before the phase change on the sub.

Before/after


At certain parts of the room, it now sound overly 'bassy' and is very uncomfortable to stay at. At the sweet spot it sound pretty good.

@Admin: why is the forum messing with the size of my images? They are all 800px wide, but somehow get scaled down. It seems to work fine for JonFo.

 

[JonFo]

I really want to hear your system Jon.

Anytime you are in Atlanta area Amey, open invite to hear it.

And I really want to visit Australia someday, ya'll have some great places to see.

 

[JonFo]

@Admin: why is the forum messing with the size of my images? They are all 800px wide, but somehow get scaled down. It seems to work fine for JonFo.

Hum, maybe because:





Seriously, no idea why yours progressively scale down.

Working on the responses to the main topic.

 

[JonFo]

... The source DAC delivers the signal to the analog ins of the MiniDSP. There I feed both L&R-input signal to the sub output, so it gets the full range combined L&R input signal. To prevent overlap in frequencies I choose in the MiniDSP to XO between L&R and Sub at 100Hz (24dB slope). So I applied the XO-over to both L&R and Sub. This should result is having no overlap in bass.

Thanks, and that is the correct wiring scheme and basic miniDSP config for your 2.1 setup.

... If I understand you correctly, I will not change the phase (set at 0 deg) on the sub, but just add a delay to the sub-output on the miniDSP to flatten out the bump at 100Hz in the GD graph.

I already applied EQ over that range to the sub.

Jonathan, do you think I need to try out other XO-points (and possibly slopes)? Or is 100Hz (24dB slope) already good enough as an XO-point?

I would indeed recommend a lower XO point, as the Montis has a decent low-end. So 80Hz would be something to try, and even 70Hz might work.
With 24db/octave (4th order) slopes, that should keep the sub in it's optimal range and likewise the Montis.

Please post the miniDSP EQ settings, I might have some feedback on those.

Thanks for posting the raw Rew measurements, studying those now.

 

[amey01]

Anytime you are in Atlanta area Amey, open invite to hear it.

And I really want to visit Australia someday, ya'll have some great places to see.

One day for sure. Likewise if you're in Australia - you're welcome for an evening of good music.

 

[JonFo]

OK, back at it, got dragged away to go play tennis on a (rare) cool (71F) August afternoon.

Your sub location is suspect, as that huge 40Hz null is present in the sub-only metric. The peak at 50hz, and the crest at 30 hz is likewise a boost from location, must be near a three surface intersection, right?

I know locations are limited, but ideally, a single sub is placed along the front wall, centered on that wall. If two subs, then the 2nd is centered along the back wall.



I'd recommend spending an hour moving the sub around and re-taking this same no-xo no-EQ metric until it looks as even as possible, it will never, ever be perfect, so don't get frustrated, just find the location that has the least amount of swings.

 

[JonFo]

... Change after changing phase on sub
View attachment 19856

Not sure how it happened, but now I've got a huge peak at 52.5Hz. In the 'only sub' and 'only speakers' SPL graphs you can see that both have a peak there. Is this a room mode? Any idea how to get rid of that? I didn't have that before the phase change on the sub. ...

The peak at 52Hz is a room mode energized by the L/R Montis, as seen in this spectrogram derived from your metric of those two speakers alone.



It is basically similar to the sub curve, but offset a bit due to position. This tells us that this pattern of peak and null are indeed room-induced.

So what we want to do, is find a combination of locations where the sub is NOT reinforcing the same pattern. Now the XO will mitigate this, as the Montis should not be producing as high output at 52Hz with a 4th order XO.

Back to my previous post, find the ideal sub location, then use EQ to knock down the peaks. So in a future PEQ setup in the mini DSP, set a high Q (sharp notch) cut filter for the L/R outputs at 52.5Hz.

 

[JonFo]

...
Before/after
View attachment 19858

At certain parts of the room, it now sound overly 'bassy' and is very uncomfortable to stay at. At the sweet spot it sound pretty good. ...(

Yes, that overly 'bassy' thing will happen based on room modes, the closer to a wall/floor or wall/all/floor intersection, the worse it will be.

But one thing to consider here is to possibly move the prime listing position to get out of the worst null/peak thing. I had to move my main position by a foot to fix some nasty nulls, and that's in a room designed to be 'optimal'.

One more question, I gather room construction is typical EU concrete walls, right? That might explain the huge low-end resonances, as that's typical of solid walls and floors. EQ cuts will be your friends here.

 

[RDC]

WOW! A lot of information and stuff to try out. Thanks!

Placement of the sub is limited. It's now placed about a quarter length in on the back wall (behind the listener). I can try two or three spots along the back wall, just to see the difference. Unfortunately, due to how my living room is situated, I have no space left over on the front wall.

Do I need to adjust the phase on the sub, or leave it as it is for now?

My walls are brick and mortar and floor and ceiling are wooden beams and wood floorplanks. They bounce!
Most buildings over here from the early 80's and upwards are almost all concrete boxes. Mine is from the 70's.

 

[RDC]

So in a future PEQ setup in the mini DSP, set a high Q (sharp notch) cut filter for the L/R outputs at 52.5Hz.

I use REW to create the EQ-files to import. Is there a way to set a high Q for that frequency within REW EQ? I can't find it. It only seems to do it automatically.

[Edit 1] I think I found it, with the modal analysis and resonances panels. Is it best to select only the 'high' resonances, or the whole lot when creating a filter?
I still haven't found out how to adjust the Q (high notch). I can only use what REW generates.

[Edit 2] Nope, using resonances doesn't change a thing to the EQ-filter created. I'm stuck.

[Edit 3] Found it! EQ button on top...that's some strange user interface REW has. Just dabbling with it at the moment, see how it reacts in REW.

Only a PK let me set a Q. What's a number for a high Q? is 0.1 high, or is 50.0 high? I just see the whole plot move up or down, not just at 52.5Hz.
With a type 'Modal' I can set the T60 target, If I set T60 to 200ms it only dips around 52Hz. Isn't that what we want? And when I switch from modal to PK, it sets the Q automatically based on the modal.

 

[JonFo]

WOW! A lot of information and stuff to try out. Thanks!

Placement of the sub is limited. It's now placed about a quarter length in on the back wall (behind the listener). I can try two or three spots along the back wall, just to see the difference. Unfortunately, due to how my living room is situated, I have no space left over on the front wall.

Do I need to adjust the phase on the sub, or leave it as it is for now?

My walls are brick and mortar and floor and ceiling are wooden beams and wood floorplanks. They bounce!
Most buildings over here from the early 80's and upwards are almost all concrete boxes. Mine is from the 70's.

Real quick, as I'm in the middle of other stuff.

Move the sub to be centered on rear wall.

Since it's physically quite separate from the fronts, you must set up delay on the sub output. Start with 1ms for every 1 foot(sorry not metric, do your conversions) between the sub and the mid-point between the ML's.