First go at room measurements using UMIK-1 and REW

[C0mbat]

Oooh.... you have an almost square room ... very similar to mine. My room is 4m x 4m.

Your room super huge mode at ~47Hz looks very similar to mine (@ 43Hz).

Allow me to share a few ideas to try:

• make the speaker-wall distances different (speaker-sidewall <> speaker-frontwall)
  • simultaneously keep the left-side & right-side symmetrical
• listening position away from the backwall (i think your diagram already shows it)
• use EQ to tame down those >10dB modes
• my square room can sound a little honky/boxy ... if so ... you can try to use EQ to tame down the 300Hz-600Hz region between 1dB to 2dB (Q anywhere from 1.0++)

Hope these ideas help...

Thanks for this. On your point 1, I'm limited with my options. I can move the speakers forwards or back a few inches but that's it. I posted a picture of my room and speakers above previously.

Im not sure what you mean by bullet point 2?

 

[Grandzoltar]

Understanding LPF And HPF Settings

Assuming the crossover acts as a brick wall is a common mistake that makes it really difficult to integrate your speakers and subwoofers. When setting the HPF and LPF filters yourself, you need to know about slope.

us.kef.com

 

[OCA]

Anyone got any tips?

You can set up REW's "Configurable PEQ" according to Wiim's PEQ Gain/Bandwith limits. Once done, REW's "Match response to target" feature will generate the optimized filters for your intended target curve suitable to Wiim DSP. Here's a sample below. Q (Bandwith) range is correct but I couldn't find gain limits:

I guess Q format will be Classic Q but you should still check (even REW and rePhase use different default Q types).

 

[Grandzoltar]

Being that the f3 of your speakers is 80hz I would experiment with a hpf of 90hz or 100hz and a lpf of 70hz or 80hz aligned with kefs explanation

 

[C0mbat]

Being that the f3 of your speakers is 80hz I would experiment with a hpf of 90hz or 100hz and a lpf of 70hz or 80hz aligned with kefs explanation

Thanks for the clarity. I'll try that. I'm a bit confused though because Kef have said in an article on their website to try a LPF of 50hz and a HPF of 70hz so I'm already higher than their recommendation.

I did try a LPF of 70hz (instead of 60hz) and a HPF of 80hz and that filled the 80hz dip a bit (not fully) but caused 65hz to bump higher. However, that might be better given that I can more easily deal with bumps vs dips.

Now that I know that trying to use eq to reduce dips isn't a good idea, then I'll first play around in trying to sort that 80hz dip with both the crossovers and speaker/sub placement. I assume that there's a possibility that I may never correct it without trying room treatment but I get the feeling that the dip isn't severe enough to worry too much about it?

 

[C0mbat]

You can set up REW's "Configurable PEQ" according to Wiim's PEQ Gain/Bandwith limits. Once done, REW's "Match response to target" feature will generate the optimized filters for your intended target curve suitable to Wiim DSP. Here's a sample below. Q (Bandwith) range is correct but I couldn't find gain limits:

View attachment 347289

I guess Q format will be Classic Q but you should still check (even REW and rePhase use different default Q types).

Thanks. I'll take a look more closely over the next week.

 

[C0mbat]

Just one other thing. Should I be concerned that the frequency range below 1k is 5-10db above the frequencies above that? Should I be trying to bring everything above 1k up by using the shelving filter?

 

[OCA]

Just one other thing. Should I be concerned that the frequency range below 1k is 5-10db above the frequencies above that? Should I be trying to bring everything above 1k up by using the shelving filter?

I wouldn't EQ beyond the room transient frequency (Schroeder's) which is around 200Hz for typical rooms. If you apply frequency dependent windowing to your response, you will see that the high frequency response will tilt down which is much closer to what you will hear.

Check @Keith_W 's thread "Should we correct for Schroeder....." for more info.

 

[C0mbat]

I wouldn't EQ beyond the room transient frequency (Schroeder's) which is around 200Hz for typical rooms. If you apply frequency dependent windowing to your response, you will see that the high frequency response will tilt down which is much closer to what you will hear.

Check @Keith_W 's thread "Should we correct for Schroeder....." for more info.

Thanks for this. I couldn't find his post but I get that changing the eq on above 200hz might not make sense. However, if you look at my above (yesterday) measurement in yellow, you'll see that everything above 1k is 5-10 db lower than everything below. So I'm not talking about smooth out the bumps with peq. I'm just wondering if I would need to raise the whole lot using one of my slots with a shelving filter? I should I not be worried about that permanent dip above 1k?

 

[staticV3]

Thanks for this. I couldn't find his post but I get that changing the eq on above 200hz might not make sense. However, if you look at my above (yesterday) measurement in yellow, you'll see that everything above 1k is 5-10 db lower than everything below. So I'm not talking about smooth out the bumps with peq. I'm just wondering if I would need to raise the whole lot using one of my slots with a shelving filter? I should I not be worried about that permanent dip above 1k?

Your in-room response is supposed to slope down. It should not be flat.

 

[DJBonoBobo]

@C0mbat 1/6 smoothing is not ideal to see the actual bass response. Look at "var" smoothing for the bass.

 

[C0mbat]

@C0mbat 1/6 smoothing is not ideal to see the actual bass response. Look at "var" smoothing for the bass.

Oh right. This is different to what others recommended. I'll take a look at that. Thanks.

 

[DJBonoBobo]

Oh right. This is different to what others recommended. I'll take a look at that. Thanks.

You can chose different smoothing options for different goals. 1/6 or psychoacoustic can be helpful for some tasks, but no smoothing shows bass modes most precise.
Var smoothing combines little smoothing in bass with high smoothing in treble region. Personally i use it most often, but not exclusively.

 

[C0mbat]

Your in-room response is supposed to slope down. It should not be flat.

Brilliant! Thanks for this. Really interesting and educational. This is a rather large rabbit hole!