Room measured, how best to proceed with EQ?

[aquavit]

Hello!

After lurking for a while and trying to absorb as much information as possible I finally bit the bullet, bought a Umik 1 mic and a miniDSP 2x4HD in preparation for improving my sound.

I've got myself to the stage of having measured my room and am ready to apply EQ for uploading to the miniDSP. In line with what seems to be the advice on here I was planning to EQ only the subwoofer integration. For guidance I have been following this YT video:

very clear and easy to follow instructions. However he is treating the complete system (mains and sub) with what he calls a "house curve" derived from the Harman curve. I don't think I need all this.

This is the graph of my averaged out readings, I've no idea if this is good, bad or "normal":

My room measures 5.6M x 4.7M x 2.8M. System comprises Audiolab 6000A, Dali Oberon 1's, Dali E-9F subwoofer.

Any advice as to how proceed from here would be gratefully appreciated.

 

[Rednaxela]

If you have just started to learn, I would disconnect the subwoofer and park the integration objective for the time being.

Just try and optimize your Oberons first.

Once you feel you have a good handle on the various measurement techniques, EQ strategies, pros and cons of house curves etc. only then take the next step and start working on subwoofer integration.

This is the graph of my averaged out readings, I've no idea if this is good, bad or "normal":

Make it your goal to be able to answer this question yourself. If you follow the above you will reach this point in no time.

Not the answer you were looking for perhaps, but hope it helps anyway.

 

[ZolaIII]

You need more then that. Those are small speakers so you would really benefit from using two sub's one per each speaker.
I would use them with port's plugged (to minimise cabinet refractions and what it does to uper mids) and without grilles (when you don't use system please put them back on). Best crossover point is about 160 Hz but that would work only with two sub's which would make them more friendly to drive, lower the THD and future improve the mids (little more tight and up from lower mids). That's what you can do about speakers.

Regarding the room try to minimise the deep in 70 to 90 Hz region if you can by better placement. Rest is pretty much usual. When you do that PEQ only the first spike at 63 Hz and do the measurements again. Now you can apply self generated full range PEQ with adopted curve to reference calibration SPL point or desired one (if you won't use equal loudness normalisation you account it in for this).
When you decide how to go and get to point of EQ-ing room mode (63 Hz spike) ask for the "target curve" and to SPL as explained.
Best regards and have fun.

 

[aquavit]

Thank you for the replies!

I have gone down the route of trying to optimise the sub first of all, there seems to be conflicting opinions on this - some say to do this and others say just optimise the mains/complete setup.

Two subs may be problematic (WAF and limited space) as is location of just a single sub, both the mains and sub are in a fixed position dictated by furniture placement. This is in a living space and will remain optimised for that, I can accept that it may not be optimal for SQ hence the desire for room correction.

DALI say the Oberon 1's have been voiced for grilles ON, note they also state wide dispersion with no toe-in - this is how I am using them. They may well be replaced in the longer term (they weren't intended for here to begin with) but for now I want to make the most of what I've got.

This is a very steep learning curve but I am at the stage where I have applied EQ to the sub template courtesy of the miniDSP correction filters. I now need to check this with a sound test but I've hit a problem as I'm not getting any noise out of the subwoofer. I have connected the Pre-Out of my amp to Inputs 1 & 2 of the miniDSP and the subwoofer cable (twin RCA) to Outputs 3 & 4 of the 2x4HD. Attached screenshot shows how I have configured the miniDSP control panel. I'm not sure what I'm missing here ...........

 

[ZolaIII]

You first do leveling individually L & R then add Sub to it and level it up. Then do measurements separately. Problem is Obren 1's are small speakers so you would need to cut them too high for only one woffer and then you will know those frequencies are comming from it (60~80 Hz upwards).
As for now try to do 24/12 dB mains/sub crossover at 90 Hz.
Then measure and let's see (including; RT60 decay, clarity and waterfall plots prior leveled up of course). And get the damn grilles off as measured. How much PEQ's per bank/chenel does that 2x4 suport?

 

[aquavit]

Ok, this is what I have - taken from my listening position which is off axis and grills removed.

First individual speakers no sub, Var smoothed and average:

Next is mains plus subwoofer, crossover on the sub amp set to 90Hz, volume on the sub around halfway:

And this is the mains + sub averaged and var smoothed:

 

[ZolaIII]

Put about - 6 dB on left speaker. Don't set the crossover on the sub but MiniDSP high pass the mains 24 dB slope per octave and 12 bd low pass the subwoofer. Then match it's volume as close to mains and staying just a little bit up (currently it's way above but you didn't high pass mains so all on it's own). Can you move sub to the side where is right speaker as it exhibit less of the deep between 70 to 90 Hz? If you can do it and then let's PEQ 63 Hz room mode on the sub (you put it on the third and forth chenel PEQ bank's. I meant RT60 decay times (right behind RT60) and do the 1\6 smoothing. You can change smoothing in graphs setting for already done measurements and in settings to set that as default when doing a new ones.
Once you have done the above do the measurements again.

 

[aquavit]

Thanks Zolalll, let me digest that before moving move forward. For reference the sub is already located to the right of the RH speaker, speakers are equidistant from side walls. This photo may help:

 

[ZolaIII]

Placement seams fine, close to back wall which ain't great with back ported speakers but it is how it is. Is the other side of the room different shaped or is your listening spot of the centre? I am trying to understand why L had +6 dB which is very significant (50% louder subjectively).

 

[aquavit]

Placement seams fine, close to back wall which ain't great with back ported speakers but it is how it is. Is the other side of the room different shaped or is your listening spot of the centre? I am trying to understand why L had +6 dB which is very significant (50% louder subjectively).

LH does look suspicious, I wonder if my error in moving mic around - I may redo them to double check. As my listening position is off-axis it places me nearer the RH speaker and I compensate by adjusting the balance on the amp. For the test I centered the balance control. Listening position is marginally nearer the speakers than the centre of the room. The rear of the speaker cabinets are 17cm from the wall, I could move them marginally forward (DALI recommend 1cm - 50cm).

 

[ZolaIII]

LH does look suspicious, I wonder if my error in moving mic around - I may redo them to double check. As my listening position is off-axis it places me nearer the RH speaker and I compensate by adjusting the balance on the amp. For the test I centered the balance control. Listening position is marginally nearer the speakers than the centre of the room

You need to start using 2x4 for all as that it has including levels. I guessed armchair is the listening spot. Thing is if we do it to that spot it won't sound very good elsewhere. I need to find of axes measurements for Obren 1's to see what can be done off axes and what's the sweat spot on angle. I guess moving the chair to center isn't that much to ask? Ideal would be moving sofa to that position and making a bit wider setting for more people for both music and TV but who knows how that plays with your functional layout.

 

[aquavit]

You need to start using 2x4 for all as that it has including levels. I guessed armchair is the listening spot. Thing is if we do it to that spot it won't sound very good elsewhere. I need to find of axes measurements for Obren 1's to see what can be done off axes and what's the sweat spot on angle. I guess moving the chair to center isn't that much to ask? Ideal would be moving sofa to that position and making a bit wider setting for more people for both music and TV but who knows how that plays with your functional layout.

I hear what you're saying but, for better or worse, the room layout is fixed. Here is some info on Oberon 1 https://www.i-fidelity.net/testberichte/hifi/dali-oberon-1/seite-5.html

 

[ZolaIII]

Never mind, they also skimped it. Did found partial one's for Obren 5's hire so let's say it's good for 10+° horizontal.
Not even a chair to a center as listening position?
You can do it off axes but it will complicate a lot of things (from timing to phase correction) and it will sound as crap pretty much everywhere else.

 

[Jprime84]

for better or worse, the room layout is fixed

Always gotta maintain board approvals!

 

[Keith_W]

LH does look suspicious, I wonder if my error in moving mic around - I may redo them to double check. As my listening position is off-axis it places me nearer the RH speaker and I compensate by adjusting the balance on the amp. For the test I centered the balance control. Listening position is marginally nearer the speakers than the centre of the room. The rear of the speaker cabinets are 17cm from the wall, I could move them marginally forward (DALI recommend 1cm - 50cm).

Welcome to ASR!

1. It appears as if the massive volume discrepancy with the L speaker being +6dB with respect to the right is due to adjusting the balance of the amp due to your off axis listening position. The first step should be to adjust the balance so that both speakers are the SAME VOLUME at the listening position. I second @Zollall's suggestion to reconfigure your listening position so that it is in the center. You don't have to keep it that way permanently, you can move your furniture around when you want to listen.

Speaking of which, what is your aim for room correction? Do you want to optimize for the main listening position only? Or do you want a "general purpose" optimization that will work no matter where you are in the room?

2. Are you using a microphone tripod and keeping your mic in the exact same position between the sweeps? Resting the mic on a cushion or books will change the frequency response quite dramatically. Make sure you use a mic tripod, preferably one with a very long boom because the tripod itself can alter the frequency response.

3. You asked whether you should correct bass only, or the whole system, and you are confused about the contradictory replies. I asked the same question on ASR a few months ago - thread here. As you can see from that poll, ASR members are divided on the issue. Since that thread, I have changed my approach and I now correct full range. I suggest you read that thread to gain an idea of the pros and cons for each approach, and then try it out for yourself and see which one you prefer.

4. In case you want to know, I calculated your Schroder frequency to be 128Hz using the following inputs: dimensions 5.6m x 4.7m x 2.8m, T30 of 0.3s. The formula for the Schroder frequency (Fs) = 2000 * sqrt (T30/V); where V = room volume. This means your transition zone ends at 4Fs, or 4 x 128 = 512Hz. If you wish to correct "bass only", the normal practice is to correct up to the Schroder frequency, in your case it is 128Hz. If you wish to correct full range, a different strategy is needed above 128Hz because the sound no longer exhibits minimum phase behaviour, i.e. it can not be corrected by inversion.

5. I am surprised by how little output you get from your subwoofer! It looks as if there is reasonable output till 40Hz, and then it drops off a cliff. I did some googling and I found this published FR for the Dali E-9F sub:

So it really looks as if your sub does not produce any meaningful bass below 40Hz.

You can decide whether this is an issue for you or not. Personally I would not be too happy with a huge chunk of my bass missing, but this might be more of an issue of not looking nice on measurements but may be inaudible depending on what music you listen to.

 

[aquavit]

Thank you Keith, a very comprehensive reply and a lot to unpack!

To cover a few of your points, there is no way I will be moving furniture to and fro every time I want to listen to music, this is a house to live in and the reproduction of music comes secondary to that. Now, if I had a dedicated room .........

As my listening position is off-axis and other seating areas are all off-axis I am looking for a "general purpose" optimisation, I enjoy the sound of my speakers but the "boomy" bass is the main issue and the sub-woofer was added to take the strain off what are very compact speakers. I have opted to try room correction in an effort to maximise what I have before/if upgrading.

I have this morning done a fresh measurement sweep with the mic on a proper floor mounted stand, following this I had a bash at sub integration following the tutorial from miniDSP https://www.minidsp.com/applications/subwoofer-tuning/sub-integration-rew It seems from your comments on the Shroder frequency I may have to trim down to 128Hz from the currently set 200Hz? I have set the crossovers at 100Hz, Dali suggest 90Hz for the Oberons but 100 sounds better to me.

Here are some measurements from today:

I also re-ran individual speaker measurements with balance zeroed and the mic better mounted:

A little alarming about the FR of the E-9F sub, although I'm not a bass head should I be worried?

I have very little experience in this game and very little to compare with which to make valid judgements, so I'm on a very steep learning curve.

 

[Keith_W]

Hello Stephen, to answer your questions:

1. Do you need bass freqs between 20Hz - 40Hz that your system currently does not produce? My answer: I can't answer that for you, because it depends on what music you listen to. If you want to find out, run your music through a real time analyser (like the built-in RTA in JRiver). This analyses the amplitude of each frequency in real time. If your typical music selection has a lot of signal between 20-40Hz, then you should want to reproduce it. HOWEVER, this comes at the cost of a new sub, or even multiple subs. And also, your brain is good at reconstructing the missing bass due to the "missing fundamental effect". As you might guess, the answer to whether you "need" that bass varies from person to person. You have to balance your OCD and your music against your budget and domestic situation. Only you can do that.

2. Wanting to optimize for a large area AND getting rid of boomy bass with a single subwoofer is a problem which DSP may not be able to solve. The usual approach is to use multiple subwoofers to even out the bass response in the room. If you have a single sub, correcting the frequency response in one position will introduce problems elsewhere in the room. The larger an area you wish to correct, the less effective DSP will be (although some DSP is still better than nothing). If you want to correct over a large area, I would suggest drawing a map of your room marking out the areas where you wish to correct. Then when you take the measurement, you need to perform a sweep at every listening position you wish to correct for, and then average out the curves, then load that into your DSP software and correct that.

Bear in mind there are some simple measures you can take to alleviate boomy bass, e.g. opening doors and windows. Obviously the downsides are that you annoy your neighbours and cold weather, insects, and traffic noise get in, but you can decide whether that trade-off is worth it for you.

3. If I am reading your measurements correctly, it appears that your Dali Oberon's are able to reproduce frequencies down to 40Hz! And your subwoofer is the same, it only goes down to 40Hz. Are you sure you took a measurement of the speaker only with no sub? I tried Googling for the FR of the Dali Oberon 1's and I am not sure if the image posted by @Zollall upthread was the FR of the Oberon 1's since his image is unlabelled. His image shows that a roll-off below 100Hz, which is the kind of performance I would expect from a small speaker like yours.

Obviously, if you have three speakers (i.e. 2 speakers + a sub) capable of going down to 40Hz, that makes point 2 in this post invalid. You would be able to use all 3 speakers to even out the bass in the room. Are you SURE your speakers don't roll off at 100Hz? This is important because it will change the conversation.

4. Re: choosing where to correct to. Yes, theoretically your Schroder freq is 128Hz and this is where you should correct to because above this frequency, bass no longer exhibits minimum phase behaviour in the so-called "transition zone". However, if you look at your transition zone (between Fs and 4Fs, i.e. between 128Hz to 512Hz in your case) you will see the peaks and dips gradually becoming finer and spaced more closely together, corresponding to gradual loss of minimum phase behaviour. I wouldn't be so strict about the calculated 128Hz Schroder frequency, because this calculation assumes the speed of sound is 343m/s, and the speed of sound varies with temperature, humidity, elevation, and so on. I also don't know your T30, I guessed at it after seeing your T60 is 500ms. Regardless, the Schroder frequency is not a number you need to religiously adhere to, it is a rough guideline. If you decide on "bass only" correction, you can go up to 150Hz - 200Hz with no problem.

 

[ZolaIII]

That's more like 130\150 Hz crossover regarding sub raw/corrected. I said 90 as minimum from their measured responses ideal is 160 Hz but with pair of sub's in 2.2 setup, with one it's minimumal one.
Now balance is better with difference of about 3.5~4 dB between L & R @ 1 KHz.
Set the REW correction range to only 57 to 72 Hz no smoothing and PEQ the room mode only to a target level about 78 dB! Enter the PEQ to output chenels of 2x4 that feed the sub's. Now you're bass is not boomy. And you can proceed to measure again and full range correction to desired target. 40 Hz all do not very low (it's low bass but no sub bass) is enough for most materials 30 Hz is for pretty much all and 20 and under well for brave ones regarding neighborhood.
@Keith_W it's from Australian Audio measurements and I don't doubt them tho those aren't complete so I looked up for tweeters of axis from Obren 5's picture someone posted hire.
Judging on that those have 10+° horizontal sweat spot and could fill the sofa from mid field quet nicely.
You two have fun with this.

 

[Willem]

The measurements are still somewhat puzzling, with the channel inbalance and the lack of bass below about 40 Hz. However, from what I read the claimed extension is indeed only down to 37 Hz. Perhaps moving the sub even more into a corner will help there. It is an admittedly small sub, but even so. I have a somewhat comparable KEF Kube 8 and fully pushed into a corner it will go down to 25 Hz even though the claimed response only goes down to 34 Hz.
So I think you first have to do a bit more measuring to really figure out what is happening. Beyond that, I can only recommend a second sub, for more power and a response that is more evenly distributed across the room. Subs don't have to be identical, so you could opt for one that goes down somewhat lower.
For equalization of two subs across the entire room I would go for Multi Sub Optimizer. It is hard to do, but the result is much better than the alternatives. You already have all the necessary hardware, other than a second sub.
I sympathize with your desire to keep this as a room for living. That is also my take on these things. Even so, if you go the multiple sub route, the subs really don't have to be that large.

 

[aquavit]

Here's a better graph of the sub measurement taken this morning, uncorrected no smoothing:

FR according to DALI: Oberon 1 51-26,000. Sub E-9F 37-200

Edit to add, when doing a measurement this morning with the volume too hight the windows were rattling. Do I need more bass than this??