REW - Help Needed with Subwoofer Integration!

[maezun]

I increased the gain on the subwoofer until it was just over the target (0.8dB/8ve aligned at 80dB @ 1000Hz ie. the SPL at which I set the speakers using "Pink Random" speaker cal in REW generator). Did I do this incorrectly?

As I intended to EQ only below 200Hz, obviously I needed to align the target with the rest of the response (as opposed to letting REW "calculate the target level"). The EQ I've been using for the past few weeks is in post #17 above:

Is it really that bad? This is the response post EQ:

View attachment 514408

Have I been doing the EQ incorrectly? In any case, I have applied the EQ in Roon using a convolution filter.

 

[ZolaIII]

@Keith_W they are and I told him to do counter PEQ too sub with it fully exited and mesure again with PEQ in place active so we can exclude them (first three peeks and two deeps) and he is done at least half the job with that in place. I am worried about 150 to 500 Hz and want to see (surely you do too) how much is phase shift and how much are room refraction and of course try to get it back as close to 0 dB as posible. 550 Hz is combined room and box but even cutting woofers a little bit will lower boot port output and related cabinet refractions. If he even get them to -3 dB then they will be good enough (-18 dB under woffer response) to interfere less with FR. Please guide him and as you took a good look at the measurements are the highs focused (not beeming)? I don't think that's posible with listening position so close to wall behind tho easy enough to correct with mild accustic treatment there. I always recommend curtains as they are least obstructive, easy to maintain and don't look bad.

 

[ZolaIII]

@maezun did you do SPL matching between L&R speaker first? You need to do that first and then SPL match to them stereo or little bit above. And that's it. If you shave top of sub response to EQ it then you rematch it afterwards. I don't know what rew generated regarding noise. We use white focused noise not pink more shattered focus one. Simply said white will sound sharper (as it is because it's more focused). It's important to do room fundamental EQ first becose with it out of picture so are it's related harmonics and largely response irons out for first two peeks and deaps behind it. Only then (with that out of picture) you mesure and see what else should be done regarding EQ-ing rest of the response. We need to see what can be done to get back lower mids as close as possible to 0 dB.

Rest will be easier.

 

[maezun]

@maezun did you do SPL matching between L&R speaker first? You need to do that first and then SPL match to them stereo or little bit above. And that's it. If you shave top of sub response to EQ it then you rematch it afterwards. I don't know what rew generated regarding noise. We use white focused noise not pink more shattered focus one. Simply said white will sound sharper (as it is because it's more focused). It's important to do room fundamental EQ first becose with it out of picture so are it's related harmonics and largely response irons out for first two peeks and deaps behind it. Only then (with that out of picture) you mesure and see what else should be done regarding EQ-ing rest of the response. We need to see what can be done to get back lower mids as close as possible to 0 dB.
View attachment 514417
Rest will be easier.

No, I didn't SPL match the speakers (not really sure what you mean tbh). It's literally just a sweep with a boom mounted UMIK-1 at the MLP, as centrally positioned and level as I could manage with a tape measure. The gain of the sub is set so it sits above the target and it's phase is based off a few test sweeps.

That's about it, mate. I'm not setting up a studio or anything, so perfect razor flat fr isn't life or death. It's only for Roon / Tidal ffs!

I don't really want to spend money on acoustic treatment, either - I had a look at GIK and thought their prices were hilarious.

Really, I just wanted to get @Keith_W 's opinion on correct subwoofer phase as a follow up to his method suggestion from earlier in the thread.

 

[ZolaIII]

L and R SPL will be different because R picks a lot from side wall and corner so you need to dial it down to match SPL of left one.

 

[maezun]

If I measure the left speaker at 75dB (Pink Random / Speaker Cal 500-2000), the right will also measure 75dB.

Is there an SPL difference between these 2 sweeps, then? I haven't adjusted volume, I literally just swapped channels:

 

[Keith_W]

@maezun your post-EQ curve looks good. I agree with @ZolaIII, you have a dip 150-500Hz. I looked back at your older measurements (posted 21 Jan) which had individual L/R speaker measurements, and the dip is already in each speaker:

Yes, this can be fixed. But it's not going to be easy. You will need quasi-anechoic speaker measurements, which you will not be able to take in your small room. And you also need quite extensive DSP knowledge if you are going to do this properly.

Basically, the problem is that we don't know if it's the speaker or the room which is causing this dip. I would wager it is the room since the excess phase group delay shows a heck of a lot of spikes in this region:

If it is due to your room, which I suspect it is, then inversion is not the correct approach. It would be: get a better room, use room treatment, rearrange the system, etc. But you have already indicated that you are unable to do this.

A very simple shortcut is to simply apply 1/1 smoothing above 200Hz and invert the response over a target curve. I don't normally tell people to do this, because you don't mess with those frequencies unless you know what you are doing. But it won't take you too long to try, and it might help.

 

[ZolaIII]

@Keith_W thanks it's the room but even so it's pressed down to how should speakers react based on B&W 607 S2 behavior for same woffer in similar but not the same enclosure. I still think root of the problem is R speakers placement and level matching won't help much either regarding deep (he still needs to do it anyhow).

 

[maezun]

Funnily enough, I actually followed @OCA 's "REW (Room EQ Wizard) Top Tricks: Convolution with Inversion" YouTube video guide last week.

I've created the convolution filter but I've yet to listen to it. I did run in to one problem, though, (with "regularisation %") as I am using a different version of REW to the version OCA uses in the guide. I also stopped short of the rePhase stuff - definitely beyond my capabilities currently!

A very simple shortcut is to simply apply 1/1 smoothing above 200Hz and invert the response over a target curve.

Sounds interesting. How would I go about doing this?

 

[ZolaIII]

Yes it is take a look at that swing it's everywhere.