Need help with left-biased soundstage (early reflection?)

[tuga]

Have you tried crossing the axis in front of you?

 

[Bjorn]

These measurements have little relevance to acoustics in small rooms. Like someone said: Graphs are for marketing. People don't understand what they are showing and how to utilize the data.

You can build some cheap absorption panels yourself that will attenuate specular reflections much better. The key is low air flow resistivity, sufficient thickness, and higher frequencies can be attenuated better from acute angles with a "well" or "pyramid" surface. Commerical foam products are generally poor because they are very band limited.

Not saying this is the only solution though. But it might improve things.

 

[edechamps]

Thanks for the suggestions.

I was wondering if there was a way to double down on my acoustic-foam-on-MondoTrap solution, since that's the only approach that seems to make a difference compared to everything else I tried.

I couldn't find detailed absorption data for the basic, cheap foam tiles I used for my test. However, I found an identically-sized product that does come with data:

Ouch, that doesn't look good. Looks like the acoustic foam tiles I used are not thick enough - they're only 2-4cm in thickness.

It does look like I can get much better performance simply by using thicker foam, e.g. something like this which is a constant 10cm thick:

That particular foam panel is not profiled. From my research I couldn't find any evidence that the foam profile makes any difference to the acoustic performance (in fact, at equal peak thickness it tends to harm it, simply because there is less foam). I would have intuitively expected the foam profile to help with diffusion, but I couldn't find any confirmation of this. (Apparently diffusion doesn't really help anyway on such short distances.) So flat panel it is.

 

[Bjorn]

Again, you're looking at measurements that have no relevance to small room acoustics. Sorry to be blunt, but there's no point in looking at measurements you don't understand.

 

[edechamps]

Again, you're looking at measurements that have no relevance to small room acoustics. Sorry to be blunt, but there's no point in looking at measurements you don't understand.

That's the only data available, so that's what I'm using. I'm well aware these measurements are for random incidence and my case is not random incidence (it's more like 45° or something), but the relative comparison still seems somewhat relevant. In any case, even ignoring these measurements, it stands to reason that a 10cm flat slab will absorb significantly more than 2-4cm wedges. So that seems like the most promising approach to solve the "absorber doesn't absorb enough" problem, IMHO.

Not saying I won't consider anything else, but I have to start somewhere. This potential solution is easy to try out and is a continuation of an approach that's already proven to be an improvement (according to ETC data and listening tests), so I think it makes sense to start there.

 

[JoachimStrobel]

You need to measure first the speakers invidually both with and without the panels. And then compare left and right, both with and without panels. That's two different comparisons. I'm not sure what you're showing.
You can't measure speakers simultaneously FIY.

Echo that. Measure each driver from each speaker with a short window in REW.

 

[edechamps]

Naive question, but did you try to swap the speakers L<>R?

So I went ahead and did some measurements with speaker swaps to rule out response inconsistencies in the speakers themselves:

To my surprise, there's actually a tiny bit of inconsistency there - PM6096629, the speaker I was using as my left speaker, has about 1.1 dB more output around 4 kHz, a critical region. It's repeatable, so not caused by random placement errors during the swaps.

That's a small error, and when listening it doesn't seem to make much difference if I swap the speakers, but clearly it's not helping. I have an additional pair of 8030As that I'm normally using as surround speakers that I have measured to be more consistent; I'll use them as my front speakers from now on. Hopefully, that combined with the upcoming acoustic treatment experiments will be enough to get rid of the problem. Fingers crossed!

 

[Chromatischism]

Or just pull the right speaker closer to you, an inch or two at a time.

Easy "timing" experiment.

But that will also change the level.

 

[levimax]

What is wrong with just balancing out the levels with either a balance control or DSP to center the image?

 

[edechamps]

What is wrong with just balancing out the levels with either a balance control or DSP to center the image?

That's what I'm doing in the mean time, but it doesn't really give satisfactory results. Sure, I can center vocals that way, and that helps, but imaging is still unclear depending on content and it's easy to notice that there is more "ambience" coming from the left than from the right. It just feels wrong, which is why I'm trying to address the root cause if possible.

I think adjusting the channel balance is an acceptable solution for fine adjustments to correct a subtle imbalance, say 1 dB or so. When faced with a massive 4 dB problem, it's just papering over the misery.

 

[andreasmaaan]

What is wrong with just balancing out the levels with either a balance control or DSP to center the image?

The perceived image shift is not only (or probably even primarily) a result of interchannel level differences, but also the fact that a strong early reflection is arriving from the left, which (I presume) is causing the OP's brain to interpret the direction of the source to be in the vicinity of the left speaker (the delay and spectrum of reflections provide important cues that our brains use to localise sound sources).

Adjusting the relative levels of the speakers will of course tend to compensate for this, but not as effectively as removing the reflection, and not without other side effects (which I presume are what @edechamps identifies as "unclear imaging").

 

[levimax]

That's what I'm doing in the mean time, but it doesn't really give satisfactory results. Sure, I can center vocals that way, and that helps, but imaging is still unclear depending on content and it's easy to notice that there is more "ambience" coming from the left than from the right. It just feels wrong, which is why I'm trying to address the root cause if possible.

I think adjusting the channel balance is an acceptable solution for fine adjustments to correct a subtle imbalance, say 1 dB or so. When faced with a massive 4 dB problem, it's just papering over the misery.

OK, understood, good luck.... room issues can be frustrating.

 

[whazzup]

https://www.acoustic.ua/st/web_absorption_data_eng.pdf
FWIW, list of materials with their sound absorption coefficient.

 

[vavan]

I've read claims that time based room eqs (vs peq approaches) can help stabilize image wander due to asymetrical layout. Its worth a shot though in your near field case I imagine it might only help over a very tight listening space. Does dirac offer a limited trial, to check the concept?

That's exactly what I was going to suggest when started reading the thread. Not only I have asymmetric setup but also slight but infamous parasound/alps imbalance and dirac helped to mitigate it

 

[racerxnet]

Another source for coefficients:http://johnlsayers.com/Stuff/AbsorptionCoefficients.htm

 

[edechamps]

I just received the large slab of 4"-thick acoustic foam I ordered.

Disappointingly, the ETC shows little difference compared to the thinner wedges. In fact, the wedges seem slightly more effective, likely because I can cover more surface area with them.

I did some more listening tests with the foam wedges and after fixing the slightly imbalanced speaker pair. The imbalance has now been reduced to just 1.5 dB, which I think I'm gonna have to accept and move on. That's before EQ; I plan to re-EQ this weekend, there's a good chance the remaining 1.5 dB might go away after EQ simply because the left-right responses will be aligned over the bottom half of the spectrum. We'll see.

The imaging seems to have improved too, but I can't really quantify that so it might just be placebo for all I know.

 

[Duke]

Have you tried tuga's suggestion of more aggressive toe-in? That would reduce how much the left speaker "illuminates" its wall.

The contralateral reflection of the right speaker would become stronger, but the ear is better at processing reflections which arrive at the opposite ear from the first-arrival sound.

Speakers which are more directional, ideally down to a lower frequency, might also be beneficial - once again the idea being to reduce the amount of energy in that early-onset left-side ipsilateral reflection.

 

[trl]

Can't you just get something like https://m.thomann.de/ro/the_t.akust...MIj5eh4L2X7gIVUOqyCh3hLAbQEAQYAyABEgJfqfD_BwE or even cheaper and find a way to place it temporarly on top of existing panel from the left side (where the sound bounces from)?

It should absorb trebles and midrange, while mid-low will be absorbed by the initial panel from the bottom.

 

[tmtomh]

I just received the large slab of 4"-thick acoustic foam I ordered.

Disappointingly, the ETC shows little difference compared to the thinner wedges. In fact, the wedges seem slightly more effective, likely because I can cover more surface area with them.

I did some more listening tests with the foam wedges and after fixing the slightly imbalanced speaker pair. The imbalance has now been reduced to just 1.5 dB, which I think I'm gonna have to accept and move on. That's before EQ; I plan to re-EQ this weekend, there's a good chance the remaining 1.5 dB might go away after EQ simply because the left-right responses will be aligned over the bottom half of the spectrum. We'll see.

The imaging seems to have improved too, but I can't really quantify that so it might just be placebo for all I know.

Glad the problem has at least been reduced - 4dB down to 1.5dB is significant, and even though it's probably about reflections rather than absolute levels, I would imagine (or at least guess) that using EQ as a brute-force way to deal with the imbalance in L-R reflections gets much easier when it's only 1.5dB rather than 4dB.

I do have one final suggestion. It might be dumb, and it's just a variation on something suggested earlier, but it would be easy, free, and instantly reversible if it didn't work: since the problem is an asymmetry in reflections, and it appears you've done all you can do in that space with absorbing the L channel reflection, how about trying to slightly increase the R channel reflection? Not by putting a whole huge 4'x8' panel on the R side, but rather by trying something on the order of a 2' x 2' (give or take) sheet of MDF or even a couple sheets of heavyweight matte board or similar, on your desk, leaning up against the side of that black wire shelf unit next to the R speaker? Low frequencies would pass right through, but a chunk of the midrange and treble frequencies that matter in this case would bounce off.

I know you want to keep your window view, but it looks like maybe you could catch a bit of the first reflection if the material protruded just a few inches out from the front edge of your shelf unit. Heck, you might even be able to get the sonic effect you're looking for with a piece of plexiglass or glass. Normally such things would be bad precisely because they are so reflective - but in this case a small surface area that produced just a little bit of extra reflection off that R speaker might be just the thing to give you that final 1.5dB worth of soundstage balance you're looking for.

If it worked, you could shape, paint, or otherwise dress up the material to make it more attractive.

Just a thought.

 

[dc655321]

Apologies for the somewhat off-topic question, but can someone explain how absorption coefficients can be > 1?