Two double-dip regions

[Mesh]

Hi all. Long term reader out of pure interest in the topic, but I would love some help on how to address two frequency regions with double dips in them.

The room is used strictly as a near-field listening room.

Setup
I am playing from a PC into a RME ADI-2 DAC, into a pair of Focal CMS40 and a Yamaha HS8s sub. I picked up the speakers mostly to see if my library could serve as a decent near-field listening room (510x310cm, and 250cm to ceiling with cinder block based walls and a hardwood floor), but have been using the DAC for headphone listening quite a while before that. Using REW measurements and a decent amount of GIK panels (all broadband even if they classify them as bass traps), I am quite happy overall as the room shows potential even if not ideal in size and have spent the year listening and feel pretty intrigued to upgrade.

For my front wall, I have 2 monster traps (60x120cm, 17cm deep) with a window between them and in the center of the room. In the front corners, I have 4 tritraps that straddle the corners from floor to ceiling. On both side walls I have 244 traps (60x120cm, 12cm deep) for first reflections. Due to a built-in bookshelf (a bit behind me), I can only have 2 on my right side while I have 3 on my left side (and overall get better REW measurements with that 3rd panel). For the back part of my room, I have 244 traps placed where I can and which produces the best measurements at my listening position (which I tested out before doing anything at all in the room). In total the back part has one 60x120cm panel and four 30x120cm panels. I actually have a 60x120 and a couple of 30x120 panels over, as I have been putting off setting up the cloud above me (ceiling is silly hard concrete and a pain to work with). Not that I think it has much impact, but I have a heavy Persian floor rug in the middle of the room, as well as a smaller sofa to my left a bit behind me (opposite to the build-in bookshelf). The back right wall is sadly not a 90 degree corner as it is cut diagonally and holds the door to the room.

Current situation and problems
Currently, I can control my lowest room mode (a smallish peak at 37Hz) using EQ, but I have what looks like two modes close to each other - one at 68 Hz and the other at 76 Hz - that cause a big dip and EQ ofc has no impact on. However, I can grab the frequency just between these dips and boost it to trick the brain to think there isn't missing level there, but would love to try and address this with a multi-sub setup if at all realistic to succeed. That is by far my biggest problem!

My second biggest problem is that I have another double dip region in the lower mids at 240-250 Hz and 310-330 Hz. Depending on monitor placement, I can move this region to be 310 Hz and 420 Hz instead, but the double dip always persists. I am not sure how to best address this one, but I suppose it could be that cloud that should get put up. It may also move around if I swap in other monitors and need to move them to get the best overall performance of them, so this region is a bit of an unknown to me how to work with.

Among my minor problems are a couple of peaks, but they all seem fine to just EQ: One at 100 Hz and one at 340 Hz. The latter is a bit tricky to work with as it is right next to the higher range of the dip, but still works to do if I just take REWs to validate that the results are good.

Beyond this, all looks really good with very even decay across all frequencies, except the 37 Hz issue which has a longer tail on a spectrogram but is decently fine to just EQ away level from to reduce the risk of it booming. My hard drive that held all my measurements from January has died and I lent my measuring mic to a colleague before Corona hit and haven't gotten it back yet, but I included a spectrogram (with my EQ from the RME DAC on) of what is almost the final result as I had that one in an email and maybe it helps a little bit? It shows the double dip in the lower mids as 310 and 420, so it doesn't perfectly match my final monitor placement. I can ofc try and get the mic back to do more measurements again if needed, but he lives in a different town so it is a bit awkward as we do not meet at work nowadays.

Plan so far and questions

• I'm pretty set with upgrading my monitors to Genelec 8030C.
• For sub, the big question is if I should aim towards 2 or more? I am not set on any particular sub yet, fwiw.
• For 2 or more subs, what would be a reasonable and good way to integrate them, bearing in mind that I have a very good RME DAC which is also used for headphone listing?

 

[Hipper]

Ceiling Reflections. As you are listening nearfield AND you have a slightly higher ceiling then normal it may be that you will not have any problems from the ceiling. What you could try is putting an acoustic panel face down on top of each speaker, sticking out at the front enough to catch any sound going to the ceiling. Listen to hear if that makes any difference.

To hear if the various peaks and nulls really affect the sound your hear, you could play some test tones. Here is the perfect test tone CD for this task:

https://realtraps.com/test-cd.htm

I don't use sub woofers so can't comment but others do report success with using up to four of them along with suitable bass management software.

 

[Mesh]

Thanks @Hipper - never knew that trick for testing, but now that you say it that sounds obvious to do! Procrastination probably got the better of me there, so hopefully that can sort out the lower mid-range issues.

I have run similar test tones through my system already, and it does affect the sound. How much actually translates to music listening is very hard to put into words.

 

[Hipper]

The beauty of the Real Trap test tones is that they go up 1 Hz at a time. I found that very useful to pin point problems that seemed to be smothered by external noise (traffic in my case).

Of course if you are going to do things properly you know you have to get your mic back!

I used copious GIK Soffit traps (twenty one to be exact!) as well as panel bass traps on front and back walls. I put some measurements up here, post 60:

https://www.audiosciencereview.com/.../share-your-in-room-measurements.13540/page-3

 

[Mesh]

Yeah for sure need to get that mic back! In light of the bigger problems the world has, it has just felt a bit petty to ask for, but I need to prioritize it.

I considered the Soffit ones also, but they do not go quite as deep in the bass as the Tri-traps so I went with the Tri ones instead. I have left space for some Soffit traps or additional monster traps though, so that is certainly an option if needed.

Btw, do wooden blinds (or any blinds really) offer any sort of meaningful diffusion if left semi-open? I only have flimsy metal blinds over my front wall window and they do nothing significant to the measurements. I've even considered mounting fixed (but fancy-looking) thicker self-made blinds from top to bottom that I can tilt each individual blind of.

 

[Hipper]

Soffits go deeper according to the tests on their site.

I don't know what blinds can do. It seems logical they would do something depending on how they were set. If you look at the design of diffusion products they use varying sizes of their gaps and protrusions to accommodate different frequencies. I would guess blinds, being the same size throughout, would only interfere with one frequency, or a very few anyway.

I tried some diffusion products but my room may be to small for them as I could hear them which I found annoying. I now just absorb. I even use GIK's 242 panels right on the outside of each speaker to prevent side wall reflections.

 

[ernestcarl]

A quick search yielded the ff. specs for your current equipment:
FOCAL CMS40
FR 60Hz - 28kHz

YAMAHA HS8S
FR 22 to 150 Hz
Phase Switch: Normal/Reverse
High Cut Control: 80 to 120 Hz
Low Cut Control: 80 to 120 Hz

It's difficult to figure out things without a good measurement microphone on hand so you really should either get yours back or perhaps get a new one eventually.

Leaving the question of room acoustics and treatment aside...

You did not mention your crossovers. Ideally, you would also need to carefully look at separate measurements for the sub and monitors. I'm not quite sure if the RME DAC you have is capable of bass management, but it seems unlikely; nevertheless, I don't think it would be optimal to use the Yamaha sub itself for bass management -- you either need a separate unit, or a better built-in one within the sub -- as in the case of a matching Genelec 7050C sub which is likely already optimized for the 8030C. It is possible (or more so likely) that the phase alignment between your current sub and monitors is not optimal. Fine tuning it requires the use of separate delay controls and flexible filters -- esp. if you are using a sub that's not specifically matched for your monitors. Theoretically, it is possible to overlap the monitors and sub without the use of additional HP and LP filters to better fill-in the area around 60-80Hz. BUT, this isn't always possible though...

Anyhows, even if you have a sub(s) that's not exactly matched for your monitors, as long as you can get the phase traces aligned as closely as possible like this:

You can overlap the sub(s) & monitors -- with maximum summation around a larger area seamlessly -- which would also help fill-in dips from either speakers' frequency response weaknesses around the xo region.

Easier said than done...

For two subs, you might want to try the ff. placement orientations:

I only have one sub placed mid-point along the front wall as circled in the above diagram.

 

[Mesh]

Soffits go deeper according to the tests on their site.

I don't know what blinds can do. It seems logical they would do something depending on how they were set. If you look at the design of diffusion products they use varying sizes of their gaps and protrusions to accommodate different frequencies. I would guess blinds, being the same size throughout, would only interfere with one frequency, or a very few anyway.

I tried some diffusion products but my room may be to small for them as I could hear them which I found annoying. I now just absorb. I even use GIK's 242 panels right on the outside of each speaker to prevent side wall reflections.

Their scales are a bit strange but the Tri go deeper than the 50Hz that their table stops at. It is splitting hair small differences though, as far as what I was told by the guy that helped me, but the Tri has slightly more effective absorption at the very bottom. The Soffits are overall by far the more effective though!

For the self-made blinds, I was thinking different depth and thicknesses, on them and ofc possibly it could help tilting some one way and others another way, and so on. I don't want to give up the window completely, but I think I will just leave it as it is until I get bored and do a mock-up prototype to stick in there and test. I actually don't have any diffusion at all, and also do not feel that to be a problem as things are now, but you know... turning all stones...

 

[Mesh]

A quick search yielded the ff. specs for your current equipment:
FOCAL CMS40
FR 60Hz - 28kHz

YAMAHA HS8S
FR 22 to 150 Hz
Phase Switch: Normal/Reverse
High Cut Control: 80 to 120 Hz
Low Cut Control: 80 to 120 Hz

View attachment 93093

It's difficult to figure out things without a good measurement microphone on hand so you really should either get yours back or perhaps get a new one eventually.

Leaving the question of room acoustics and treatment aside...

You did not mention your crossovers. Ideally, you would also need to carefully look at separate measurements for the sub and monitors. I'm not quite sure if the RME DAC you have is capable of bass management, but it seems unlikely; nevertheless, I don't think it would be optimal to use the Yamaha sub itself for bass management -- you either need a separate unit, or a better built-in one within the sub -- as in the case of a matching Genelec 7050C sub which is likely already optimized for the 8030C. It is possible (or more so likely) that the phase alignment between your current sub and monitors is not optimal. Fine tuning it requires the use of separate delay controls and flexible filters -- esp. if you are using a sub that's not specifically matched for your monitors. Theoretically, it is possible to overlap the monitors and sub without the use of additional HP and LP filters to better fill-in the area around 60-80Hz. BUT, this isn't always possible though...

Anyhows, even if you have a sub(s) that's not exactly matched for your monitors, as long as you can get the phase traces aligned as closely as possible like this:

View attachment 93118

You can overlap the sub(s) & monitors -- with maximum summation around a larger area seamlessly -- which would also help fill-in dips from either speakers' frequency response weaknesses around the xo region.

View attachment 93121

Easier said than done...

For two subs, you might want to try the ff. placement orientations:

View attachment 93122

I only have one sub placed mid-point along the front wall as circled in the above diagram.

Thanks @ernestcarl and good point - I missed the XO! The HS8s is crossed over at 100 Hz atm in a desperate attempt to try and reduce the problem below that. I let the CMS40 roll off as they naturally do. It is actually due to this I have a slight peak at 100 Hz as the sub and monitors sum a bit more than ideal there. I believe the CMS Sub crossed over at 80 Hz but it does make a bit of a difference to keep it at 100 Hz. The psychoacoustic curve in REW looks a bit happier at least as far as I can remember, as I had the sub at 80 Hz earlier.

The RME DAC has no bass management options sadly. I have considered the matching 7050 subs to the 8030C, but haven't even gotten to the point of checking if that puts me on a single sub only or if they can be daisy chained. I was so hesitant at picking up a sub that I went with something that was decent but on sale just to see how the room would behave. This was back when I was not even sure if having something more serious set up was feasible. If I can stick with my RME DAC rather than put something else in as DAC and bass management, then I'd gladly pay a bit extra for the subs as long as things don't get crazy expensive. My overall goal is to build a sensible setup and I don't really have a maximum budget. I won't be wasting money unless it also means it is way easier to set up, as I suspect matching subs would be, but if that's just not value... then change it is.

I will 100% get the mic back before I start doing any major changes. I am in no rush to do this, other than ofc the temptation of getting going. Before buying anything, I will have a serious test session focusing on phase. I have done some for sure, but more is certainly better.

 

[Mesh]

We all knew I had to go and get my mic this weekend, didn't we?

Initial problems (updated results):
- A double dip at 68 and 76 Hz (actually 71 for left channel and 77 Hz for right).
- A double dip at about 245 and 310 Hz (actually 246 for left channel, with no problem for right, and 296 Hz for both channels).

See the 1_24th smoothing graph with Black (L+R) where Blue is L, and Red is R.

New observations:
- A small asymmetry in speaker placement had major effect on top end when looking at both channels. My right speaker had gotten bumped 1cm further from the wall than the left, and the result was a sizable dip in the top frequencies. See the graph with Black (L+R) and Green (L+R with right speaker misaligned). I never spotted this before as each channel still measured very close to flat at the top end. Only when I ran an L+R test did the dip materialize. Funky.
- Putting a panel that sticks out a bit over the channel I test through to check if ceiling reflections are my issue (like @Hipper suggested earlier) had no tangible effects on my problems. I was really hoping the 246/296 Hz issues would improve. I even tested a panel sticking out quite a bit around the speaker and still got no major benefits. No graph included.
- Putting a large panel between my speakers and on my desk, as well as two smaller panels on the side of my desk (outside the speakers) to block off the surface as much as possible had slight effects. See the graph with Black (L+R) as the baseline without added panels on the desk and Pink (L+R) showing what laying out the panels did. You can probably ignore the top end drop-off, but I did repeat the test (as with all others!) and got similar results. Obviously, if I were to hang it up as the intended cloud, I would need to figure out what goes on at the top also. However, given what the Pink graph shows compared with the Black, I'm not so sure if I should. It gets better in some places, but worse in other. I lose a bit of level from a panel in the midrange also.
- Moving the mic forward 30cm or back 30cm started to affect the low end bump. The graph that is Black (my normal L+R) vs Purple (L+R, 30 cm forward) shows a slight improvement in the low end bump, but a small mid range bump and a small roll-off at the top. The graph that is Black (my normal L+R) vs Yellow (L+R, 30 cm back) show a deeper dip in the low end, a new dip in the higher mid range, and a slightly flatter top end (I assume this position is actually more on axis than what measures best overall - the beryllium tweeters can be a tad hot maybe?).
- Turning on the psychoacoustic smoothing shows that the overall result isn't terrible, even if I have the two double dip regions. The psychoacoustics smoothed graph Black (L+R) where Blue is L and Red is R shows this.

Test parameters:
- The monitors are allowed to roll off naturally. This produces the smallest dip in the low-end region, but sure - not a big difference if I cut them off higher up.
- The sub is cut at 100 Hz. I tested 120 as well as 80 and it is clearly best at 100 Hz.
- The sub only has a switch to flip the phase, and as I have it between my monitors it is clear that I do not need to flip the phase. I tried moving the sub around quite a bit, but it is clearly best between my speakers - perhaps as a direct result of the crude phase adjustment.
- I applied some low-end EQ to reduce the low end dip as much as I could, as well as to calm down the 100-200 Hz region which gets a bit hot. No other regions are touched by EQ. I normally add Sonarworks at the very end (as long as it measures better afterwards!), but have had that turned off in all testing.
- I tested moving my extra panels around to even more in the back of my room (at different locations), as well as over the front of my built-in bookcase (which btw have doors that do not rattle), but it had no tangible effects I would classify as improvements. The bumps moved a bit, but the graphs did not smooth out.

Bottom end:
- Sub is the only thing that can get my low end bump sorted it seems. It was completely unaffected no matter what.
- Mid dips and tops are not what I like to see, but I am still stumped as to how to address this.

 

[Hipper]

I'm struggling to understand the layout of your room. Can you put on here a plan of it please?

Where do you place the mic when measuring? I put it exactly where my ears will be. Others use a moving mic method.

I would advice concentrating on the 0-500Hz region first - basically bass up to what is called the Schroder or Transition Frequency, usually around 250Hz but room and contents dependent. It would also be useful to see a larger smaller range of dB on your graphs - 60-90dB would seem OK. This is to give a clearer impression of what's going on.

What do you use as an equaliser?

I'm not familiar with Sonarworks but it seems to be for headphones. Can it work with speakers too?

Anyway your biggest issue is clearly that dip centred on 70Hz. On the link I put on my post #4 I describe how, after trying everything else, I used an equaliser to correct a similar large dip even though we are told we shouldn't do that. It worked for some reason!

 

[Mesh]

Sure. I have attached it here. I sit towards the window at the top, with the mic placed where my ears otherwise are. My desk is not quite by the window and I am just inside the normal triangle at a third into the room.

It really is the region up to my 500:ish Hz that I look seriously at, but it bugs me with the jaggedness up to 1k. I have attached one of up to 1k and one up to 500 Hz.

The EQ is the parametric in the RME DAC. I have the chance to EQ each channel separately also, but haven't bothered with that yet. I could do that to try and even it out even more. I would prefer not having as much of an issue to work with to start with though. The low end dips completely ignore EQ (as expected) as soon as I hit the problem frequencies. All I can do is EQ as close to those as possible with some sort of significant effect. When I get too close to the no-effect frequencies, the boost has to be pretty significant, which hints at actually missing the frequency that 'sticks' but that the slope of the boost catches onto what can get EQed.

Sonarworks is primarily for speakers. They just happen to also have a simpler version for headphones also, and they make some money creating custom profiles for whatever headphone you send them. Their main business has always been for studios and speakers. I still don't use it in any of the measurements here, so that part has no impact.

 

[ernestcarl]

A suggestion: you might also want to look closely at or even post here your individual channel (besides summed) measurements.

Like this:

added +3dB offset to purple sub+sat averaged trace

You can see from the above that while both sub and mains are deficient in magnitude around their crossover points (due to placement & room acoustics), having their phases aligned as well as possible offsets that issue. We see a small cancellation dip at around 110Hz -- the point where the sub and mains' phases diverge too much. If you are widely overlapping and not applying a HP filter for your mains, this would also reveal imperfect summing due to phase coherency issues.

 

[Mesh]

A suggestion: you might also want to look closely at or even post here your individual channel (besides summed) measurements.

Like this:
View attachment 93764

You can see from the above that while both sub and mains are deficient in magnitude around their crossover points (due to placement & room acoustics), having their phases aligned as well as possible offsets that issue. We see a small cancellation dip at around 110Hz -- the point where the sub and mains' phases diverge too much. If you are widely overlapping and not applying a HP filter for your mains, this would also reveal imperfect summing due to phase coherency issues.

The channels were posted separated also. It is in Blue and Red below the Black line which is the summed one. Showing all three helps seeing the effect of the two separate channels.

Yeah, I get what you are saying with no HP filter for the mains. Sadly, it gets worse when I add an HP though, no matter where I overlap them. I expected 80 Hz to be the ideal place to do so for sub as well as mains, but it seems 100 Hz LP and no HP gives the smallest (but still large) dip. FWIW, the sub is placed exactly as suggested by Yamaha in relation to the speakers and neither side nor forward or backwards with the sub provide better results.

 

[ernestcarl]

With the sub muted or each channel swept in solo, I mean. Left & right, while separate, are still summed with the sub in your blue and red traces...

As Hipper alluded to with his request for a room layout, some major placement changes or re-arrangement experimentation may well be necessary. But if that’s absolutely not possible or proves ineffectual, EQ might be the only way to help somewhat.

 

[Mesh]

A, gotcha. Will get that sorted later. Night time now.

Multiple subs are still an option to cancel out the modal issue. I would widely prefer to address the issue before I look at separated channels for EQ, but I did already maximize what EQ can do to that area (without separating EQ for each channel that is).

I can add that I have done extensive testing of the room before it had a single item in it, so I know that the placement I have now is the least bad. We are at the point of fine tuning what can be.

 

[ernestcarl]

A, gotcha. Will get that sorted later. Night time now.

Multiple subs are still an option to cancel out the modal issue. I would widely prefer to address the issue before I look at separated channels for EQ, but I did already maximize what EQ can do to that area (without separating EQ for each channel that is).

I can add that I have done extensive testing of the room before it had a single item in it, so I know that the placement I have now is the least bad. We are at the point of fine tuning what can be.

No problem.

But if you want to examine an overlay of these measurements at your current listening position, ideally they would be individual channel measurements without any EQ (other than your sub xo at 100Hz) plus L+R (sub muted) and L+R+Sub summed swept sine responses just for the sake of visualization.

 

[Mesh]

I can start by confessing that I had given up on actually doing anything with the low end dip. That was why I had applied some EQ to minimize it while I focused my energy on placing panels to settle down the much harder to grasp region in the mids, but I have turned it off for this post and only looked at the low end dip.

NOTE: The graphs I mention below are both available raw and with 1/24th smoothing. I present the colors below as in the raw graphs.

General trend
It quickly becomes a rabbit hole of combinations if I consider no HP as well all three HP cutoffs (80, 100 and 120 Hz), and the three LP cutoffs (80, 100 and 120 Hz), so I can summarize by saying that there are no super strange effects at any of the combinations. The equipment largely follows the expected trends to my eyes, so nothing seems faulty. The room influences are what makes things hard for me!

Slight sub improvements
I hadn't tested it before, but in a state of forgetfulness I forgot to turn the sub to face forward (out into the room) and instead left it 90 degrees sideways along the front wall. Realizing I hadn't tested that before, I did so and there is a (very) slight improvement especially if I slide the sub a bit to the left of center in the room and flip the phase. Graph 1a shows the sub's behavior when facing forward, while Graph 1b shows the sub turned 90 degrees to the side (and a bit to the left with phase flipped). The 100 and 120 LP are a tad better overall this way, while it is clear that an 80 LP filter will just have a negative effect on my problem.

L and R are quite different
With the sub removed, L and R are quite different from 500 Hz and below. Clearly, I would benefit from individual EQ:ing of each channel. This is particularly evident above the low end dip region, so let's ignore that for this post. For the 150 Hz and below area, my right channel is much worse than my left channel. The right channel clearly contributes to making the low end bump worse, especially from 70-90 Hz. You can see this when comparing Graph 2a (left channel, where black is no HP filter, green is HP at 80, red at 100 and blue at 120) with Graph 2b (right channel - same coloring syntax).

Summing the parts
Let's focus on one example here, as each of these follow the same trends in an expected way given how Graphs 1-2 looked. With the sub now 90 degrees sideways and a bit left, a 120 Hz LP looks possibly a bit better than the 100 Hz used with a forward facing sub. Graph 3a shows No HP filter (black is sum, yellow is sub, while left is blue and right is red as usual). Graph 3b is 80 Hz HP filter, Graph 3c is 100 Hz HP filter, and Graph 3d is 120 Hz HP filter (all with the same coloring syntax).

The closest thing to an unexpected result is actually visible when you look at the summed up (sub and L+R) curves for 120 Hz LP filter. Graph 4a shows no HP included in black, while Graph 4b shows only the green 80 Hz HP filter, blue 100 Hz HP filter, and red 120 Hz HP filter. I simply don't know where that big dip for no HP at around 110 Hz comes from, but the trend is there from 80 to 120 Hz also, so there seems to be some logarithmic effect going on that at least I hadn't spotted until this comparison. Looking back at Graph 3a-3d, it could be explained by the right and left channels moving closer to 100 Hz in their peaks when going to 80-120 Hz HP filter compared with no HP filter, but I did not expect the difference to be this large!

However, sticking with the NoHP still is better for the low-end dip in two places (64-73 Hz and 79-87 Hz, and there is even an argument for 44-59 Hz being better without HP filter). The cost seems to be that dip at 110 Hz and peak at 98 Hz. NoHP or 100HP seems to be the two main options to start from when trying to EQ. Which is best probably depends on how the measurements respond to EQ.

In closing
I still feel that the best I can do with my current gear is to EQ each channel, which will help the low end dip a bit and the mid a bit more possibly. I could start repositioning the monitors also ofc, but that will put me back to square one and emptying out the room completely, and that doesn't seem likely before I actually upgrade! In the end, what I originally wanted to understand was how to best work with the two problem frequency regions as they both seem largely room/placement/damping related. Ofc, trying to do so with my existing gear is ideal just to get a serious feel for what I can affect and how is ideal as that is free (beyond the countless hours of testing, but that's also a bit fun when you learn stuff!). To really address the low end dip, I don't see a way out that doesn't include more than one sub to cancel that issue out. And I probably should get that cloud up just in case it has more serious effect when in place compared to when doing quick-and-dirty tests without mounting it. I mean... it should be better for that mid range, I would think... I just wish I had a clue if that is actually what needs to happen or if it is something else?

Aside from any reflections and thoughts on my take above, the three bullets from my OP are also still of great interest to get feedback on:

• I'm pretty set with upgrading my monitors to Genelec 8030C.
• For sub, the big question is if I should aim towards 2 or more? I am not set on any particular sub yet, fwiw.
• For 2 or more subs, what would be a reasonable and good way to integrate them, bearing in mind that I have a very good RME DAC which is also used for headphone listing?

Biggest unanswered question (was beyond this post's scope):

• What is the best way to try and smooth out the jagged midrange and especially the 300:ish Hz region with the big dip(s)?

 

[Hipper]

Unless this big dip in the 80Hz region is caused by some issue with the speakers or sub I would think the effort spent sorting the problem out now would be worthwhile even if you do upgrade to Genelecs.

I only use full range speakers, no subs, so I have no experience of positioning subs. Perhaps others can advice. I can only make a guess.

That guess is this. If your are running the subs up to 100Hz and your speakers above 100Hz then to sort out this dip we need to move around and measure only the sub whilst playing and measuring the REW signal through it. As your room is unconventionally shaped the subs final position may not be conventional.

In addition to moving the sub I would try a couple of different listening positions (mic positions) for each sub position. These would be on the 5.10 metre centre line, say your current listening position, and another at 1.70 metres from the back wall (this position is part of a set up known as 'The Thirds').

Draw a grid of the room - say every metre to start with - and measure from the sub in the various locations - by walls, and at the grid crossing points - some 23 or so positions plus the two mic measurements for each one - about 46 measurements in all! Number each spot on the grid (Aa, Bc etc. or similar system) and add are 1 or 2 for the two mic positions (e.g. 2Bc, 1Bc etc.).

Compare all these and find the best compromise for this 0-100-Hz region. You can set REW to only measure 0-120Hz say and with only one signal which will speed things up. These are only rough measurements to home in on the best location for the sub and your ears.

If all this does find a better frequency response, make a finer grid around that point to hopefully find an even better placement.

In regards to your PEQ, I see that your DAC has five bands to adjust. There are other PEQs, or equalisers with PEQs, which have more bands and the ability to make finer adjustments. My Behringer DEQ2496 for example has ten bands for each channel, or ten stereo, and can adjust the frequency as fine as 1/60 octave (e.g. I use such frequencies as 85.3Hz, 53.8Hz). You could get REW to generate the filters needed to fix any problems then manually implement them in a PEQ.

As you use a PC, perhaps you know of JRiver. As well as library management it has many other facilities including an infinite number of PEQs. There is other DSP software too.

 

[Mesh]

Yeah, no doubt the effort is helpful no matter what. It helps me see how much can be achieved by also separating the sub from the measurements, as for instance the somewhat surprising thing I mentioned last post wouldn't have been possible to explain otherwise.

I appreciate the thoughts on moving things around and the grid approach. That is what I did (although I believe I did to 150 Hz or if it was 200 even?) when the room was completely empty, so I have confirmed this to work the best it can with the gear I have (a new sub or two may activate new issues further down in the frequencies, however, so theoretically it could give a new best compromise).

In terms of new gear and specifically EQ, I'm willing to have a look at further bands if need be later. I would start from doing it in the box just to test things out first. The reason for not considering an EQ now is that multiple subs may require sub management also, and even if I would hate it I may then have to abandon my fancy RME DAC and go for a DAC that also has sub management. At that point, it would depend on what that new DAC provides in terms of EQ as many have more than the five bands that my DAC has. I'd love to NOT change out the DAC though, as it is an absolutely brilliant DAC (as proved in testing here at ASR for instance) that also is superb for especially easier to drive headphones or IEMs, and that is something I definitely need as well.

After I've cancelled out as much as I can using two (ideally four, but would start at two) subs, I can both get a grasp of what subs add to smooth out the room modes in the low end and how two do compared with one. I'm hoping at that point to either be well enough equipped with EQ already, but if I am not then I would add an EQ with more bands.

I need help from those with experience from two+ subs though, to figure out if I can keep my DAC in the chain or not and if it is realistic to hope for a major improvement with two subs.