REW EQ Right Approach

[rcarlbe]

I've taken my best first shot EQ'ing my system with REW and would sincerely appreciate any feedback. Tower speakers 2.0 system 11x14 room. I limited equalizing between 10-200hz to avoid over-equalizing the speakers instead of addressing the room. I have a big dip at around 72hz, but raising the valley 4-6 dB created boominess. I just couldn't touch it, it seems. I am well aware of room treatment, but I am very limited in this realm. I can't move speakers or listening position. If EQ can work a little magic here, I'd like to learn how. Ok, next I cut down the peak to the right of the valley at around 80hz to coincide with my target house curve. Then I raised broadly around 120hz and basically just tied it back in by 200 hz. Is there anything more I can accomplish or test here? I'm using Roon and headroom is set to -3 with no clipping. Thanks for any observations and suggestions.

 

[rkbates]

I'm using REW for the first time myself so very keen to see what users think on this one. I saw some advice which said don't touch the big dips - at the lower frequencies they are likely caused by reflections (partially) cancelling the initial sound, so if you boost the original sound you also boost the reflection and get nowhere. The other advice I used was to take a few measurements at different locations to help identify whether the dominant factor is the speaker or the room. I ended up doing this, then equalizing the average response. The waterfall plot can also be used for this.
From my reading so far I would have done exactly what you have done - hopefully the more experienced members can help a bit more.
https://www.audiosciencereview.com/forum/index.php?threads/room-equalisation-for-ijuts.28618/
https://www.audiosciencereview.com/...rial-from-a-genelec-product-specialist.24201/
The miniDSP site is also handy https://www.minidsp.com/applications/auto-eq-with-rew
The REW manual is excellent as well

 

[radix]

In another thread on room treatments, there's been some discussion about how to use REW to figure out what should be treated.

On suggestion was to get an absorber (e.g a pillow) of some sort and hold it around your measurement mic (e.g. L, R, top, back) and if that decreases your notch, you know it's a room issue and more-or-less where it's coming from.

You can do various things for room treatments, like canvas prints (maybe stuffed with material) or curtains, etc., it does not need to be big obnoxious baffles or diffusers. Even small changes can make a few dB difference. Your big notch is about -10 dB, so even a 3-5 dB difference is a big fix. I tried a few other methods too (the string method and simply walking around with a big pillow).

Another thing to look for in REW is the spectrogram or waterfall to see if you have any large frequency delays (e.g. much over 200 msec). This can cause reverb or smear out the sound.

 

[rcarlbe]

Appreciated. I intend to continue to refine my moving mic measurements. I did a pink periodic noise test, and I'll do a sweep next to check waterfall etc. Thanks again.

 

[rcarlbe]

In another thread on room treatments, there's been some discussion about how to use REW to figure out what should be treated.

On suggestion was to get an absorber (e.g a pillow) of some sort and hold it around your measurement mic (e.g. L, R, top, back) and if that decreases your notch, you know it's a room issue and more-or-less where it's coming from.

You can do various things for room treatments, like canvas prints (maybe stuffed with material) or curtains, etc., it does not need to be big obnoxious baffles or diffusers. Even small changes can make a few dB difference. Your big notch is about -10 dB, so even a 3-5 dB difference is a big fix. I tried a few other methods too (the string method and simply walking around with a big pillow).

Another thing to look for in REW is the spectrogram or waterfall to see if you have any large frequency delays (e.g. much over 200 msec). This can cause reverb or smear out the sound.

Cool. I totally agree just a few dB in that hole would be huge. I really want to nudge it up. I'll look up the string method and see what that is. Part of the issue, I assume, is I listen with my back of head 18" from a wall of windows. The windows nor my head can move/modify. My speaker position options are 0 to 12" away from the wall. I have sealed speakers, which helps. No horizontal leeway--spkrs hit couch or intrude into doorway.

 

[radix]

Cool. I totally agree just a few dB in that hole would be huge. I really want to nudge it up. I'll look up the string method and see what that is. Part of the issue, I assume, is I listen with my back of head 18" from a wall of windows. The windows nor my head can move/modify. My speaker position options are 0 to 12" away from the wall. I have sealed speakers, which helps. No horizontal leeway--spkrs hit couch or intrude into doorway.

Are curtains possible? Or even if you have vertical or horizontal blinds, those might help a tiny bit. Anyway, best of luck.

 

[YSC]

Tried with EQ APO and using harman curve, 3 measurement average is thhe blue line, red is predicted target. can't do further measurement after EQ since complains about loud sweeps comes, but after EQ the speakers sounded like it's much more open with image depth and engaging bass, prior to EQ it sounded more into the face for pop songs, no idea which attribute is related, maybe the 1.5-6k broad dip?

 

[anotherhobby]

That doesn't look like or sound like a null that's going to respond to EQ. I have also have some that don't. Most reasonable room treatments also don't do that much at 72 Hz (even thick ones). If moving your speakers isn't an option (which you will likely just find a new null anyway), then the most effective way to deal with it is by either adding multiple subwoofers or by not looking at the graph and instead listen to some good music.

 

[YSC]

That doesn't look like or sound like a null that's going to respond to EQ. I have also have some that don't. Most reasonable room treatments also don't do that much at 72 Hz (even thick ones). If moving your speakers isn't an option (which you will likely just find a new null anyway), then the most effective way to deal with it is by either adding multiple subwoofers or by not looking at the graph and instead listen to some good music.

thanks, for Bass I am pretty satisified, moving or placing extra sub isn't an option so I will just leave the 70-80hz null there. just discovered the Equalizer APO seems able to do L and R independent EQ, and the L/R channel measurement shows that my left speaker have the 1-6k null of 5db and the right speaker is perfectly flat, I think it's the abundant of stuffs literally stacked right next to the left speaker (No, removing the queen of home's stuff away is not an option) which I am thinking should need some EQ or bring down the 6-10k peak? or should I just leave that alone and the EQ won't do good as the ear is more sensitive to the direct sound?

 

[ernestcarl]

Tried with EQ APO and using harman curve, 3 measurement average is thhe blue line, red is predicted target. can't do further measurement after EQ since complains about loud sweeps comes, but after EQ the speakers sounded like it's much more open with image depth and engaging bass, prior to EQ it sounded more into the face for pop songs, no idea which attribute is related, maybe the 1.5-6k broad dip?
View attachment 184625

Presuming the speakers are anechoically neutral, a lot of the peaks and dips above the room's transition zone are best corrected with acoustic treatment and good positioning.

FDW 4 cycles at 0.85 m distance -- speakers toed on axis with 90 degrees calibrated mic

Sorry, could not remove the ceiling cloud and rest of the other room treatment. However, just that large desk boundary is already enough to cause many of the measurement non-linearities often seen in our single-point sweeps.

With treatment, less EQ correction / compensation is now going to be applied -- and only if desired:

FDW 4 cycles