Acoustic treatment with slats - before-and-after measurements

[ernestcarl]

Sure, but the point was to describe what happens with quality treatment to answer the question posted. Proper diffusion will create less sparse reflections and pack them closer together. Both diffusers and absorbers attenuate reflections. Slats neither diffuse or absorb well. The periodicity with slats is also an issue.


Good treatment doesn't add more absorption than necessary. Better to get it right from the beginning. Multichannels will create it own issues with lobing and combing.

I generally would already agree... I haven't found lobing or comb filtering to be a significant audible issue, though, at my current MCH setups. In that past, I did run into some issues with bass cancellations, but I've figured ways to reduce it.

Traces 1-3 are measurements at the rear couch of the listening room. Trace 4 is at a front office desk in the middle half of the room.

Yes, level is definitely reduced as one sits away from the aligned center seat due said lobing and comb filter phase cancellations. Yet, in my assessment, nothing heinously egregious enough to push me away from MCH upmixing.

 

[janbth]

Thanks for your reply! I am not an engineer, but I think the professional experts on the forum here will likely agree with me when I say that 9mm thick felt is simply incapable of absorbing the longer-wavelength lower frequencies like the ones you are looking to get better control of - which, if I read your measurements correctly, is why the room response at those lower frequencies appears to have changed in your before and after, but not necessarily improved. I would guess that most of the differences you've measured below about 2kHz would be more impacted by the diffraction of the slats than the absorption of the felt.

Yes, OK, I think you have a good point. Now that you mention it, I think that diffusion could possibly explain much of the reduced sound energy that I measure and hear down to a few hundred Hz. Sound that is diffused to the left would move out into the open area of my L-shaped living room, while sound diffused to the right would be absorbed by the curtains by the windows.

 

[janbth]

Seems like there's something weird going on in your L+R measurement -- could be operator error. The bass appears abnormally high as opposed to what one gets with a simple vector average or "aligned sum" between left and right.

View attachment 269650

View attachment 269651


*Generally, decay in the bass (when it is not elevated up so much) is actually okay; and, I do see indeed a marked improvement in the clarity above 500Hz after treatment. Still not perfect -- and a little uneven -- but better than before.

View attachment 269653

@ernestcarl Thanks for looking into this! Note that the L and R meaurements would not sum up to All. For the L and R measurement I have removed the subs - by the M33-specific tweak I described above in post #47. (Crossover set to 40 Hz, sub volume at minimum, -60 dB.) The All measurements plays the full system (crossover at 120 Hz, subs quite high at -6 dB). The bass does indeed look dominating, but it sounds more balanced than it looks. The sub volume has been turned up relative to what comes out of Dirac (with a +6 dB low frequency boost curve from meldhau.net as target). The sub level was set so that the *perceived* volume below crossover was the same as above. I really don't understand why I need to do this last step, but without it the sound of my system would actually be a bit bass-shy.

 

[janbth]

There are no acoustics test data on their site that I can find, maybe I’m missing something?

Click on the datasheet for the first panel, called BASIC. Page two of the pdf has the absorption plots.

 

[ernestcarl]

@ernestcarl Thanks for looking into this! Note that the L and R meaurements would not sum up to All. For the L and R measurement I have removed the subs - by the M33-specific tweak I described above in post #47. (Crossover set to 40 Hz, sub volume at minimum, -60 dB.) The All measurements plays the full system (crossover at 120 Hz, subs quite high at -6 dB). The bass does indeed look dominating, but it sounds more balanced than it looks. The sub volume has been turned up relative to what comes out of Dirac (with a +6 dB low frequency boost curve from meldhau.net as target). The sub level was set so that the *perceived* volume below crossover was the same as above. I really don't understand why I need to do this last step, but without it the sound of my system would actually be a bit bass-shy.

I'm a bit curious, can you mute the LR mains and send a quick full-range sweep signal only to the subs and attach it here as well -- if necessary (and possible), just physically unplug them from the amplifier. Thanks!

 

[janbth]

I'm a bit curious, can you mute the LR mains and send a quick full-range sweep signal only to the subs and attach it here as well -- if necessary (and possible), just physically unplug them from the amplifier. Thanks!

Yes, absolutely, I could try that tomorrow as it’s getting late here now. So you want one sweep with sub1+sub2 playing, L and R unplugged? I guess I should set the crossover to max = 200 Hz?

 

[ernestcarl]

One more thing I notice: it looks as though your room in conjunction with the subs are producing and sustaining more of this visible late reflected-decaying energy:

 

[janbth]

One more thing I notice: it looks as though your room in conjunction with the subs are producing and sustaining more of this visible late reflected-decaying energy:

View attachment 269741

View attachment 269742

Interesting - this is not normal? How is it supposed to look?

 

[ernestcarl]

Interesting - this is not normal? How is it supposed to look?

Hmmmn, if we were to theoretically take away the time domain "distortion" while at the same time keep the magnitude all the same... it's going to look close to something like this:

Extracted minimum phase version

 

[janbth]

@ernestcarl Here is the measurement of the subs. I also added a separate sweep of Sub1 (front) and Sub2 (back).

Unplugging the mains worked, but then an unexpected problem occurred: There was no acoustc reference signal! Fortunately, I found a workaround: I have a couple of Bluesound Pulse Flex 2i hooked on to the system as "extra" L and R speakers, controlled by Roon. Since I play recorded sweeps from Roon, I was able to start the sweep with the Flex L unmuted, mute it while the subs were playing, and then unmute to play the stop signal. A low clicking sound was heard when muting the L, but I dont think it was loud enough to disturb the measurements.

 

[janbth]

After a suggestion from @Absolute above in this thread, I made a new "All, after" measurement - this time using the same volume as the before measurements. The FR curve then has matching levels at low frequencies. It looks like this:

The reduction from about 700 Hz to 9 kHz may be easier to see in this graph.

 

[ernestcarl]

Unplugging the mains worked, but then an unexpected problem occurred: There was no acoustc reference signal!

Thanks for the effort, janbth! However, acoustic reference may not have been necessary as I still needed to apply a time offset to make a guesstimate of the correct alignment:

The delayed "maximum phase" peak -- honestly, not sure if I'm even using that terminology correctly here -- in the step response graph (~90ms) is coming from the subs -- specifically, magnitude energy coming going below 26 Hz where the roll-off is accompanied by a rapid rise in GD -- likely, this is partially caused by the room itself.

Very long FIR filtering applied to the sub could level that down, but I don't think it's worth it -- and, besides, your current processor will not have that ability anyway. Much of the more problematic, unevenly extending reflection-decay energy occurs right above 150 Hz where this kind of room correction DSP isn't well suited for... In fact, more liberal use of thick, broadband acoustic treatment will be a much better choice, as other's have already advised.

As to why you prefer the "highly" elevated bass... well, I do have some "theory" as to why that may be: it may very well be in order to compensate for the broad drop in clarity starting below 500Hz or so in your measured room response. Notice that in heavily treated studio spaces they often will show "flatter" FR measurements across the board -- and if there is a preferred elevation in the bass, it's usually only up by a few dB. Think about it, if the bass response is absolutely, resoundingly crystal clear and balanced along with the rest of the spectrum, there really should be no need to elevate it by so much. Eh, I could be wrong with some of the details, but I'm sure you already got the gist of it.

 

[janbth]

Thanks for the effort, janbth! However, acoustic reference may not have been necessary as I still needed to apply a time offset to make a guesstimate of the correct alignment:

View attachment 269941

The delayed "maximum phase" peak -- honestly, not sure if I'm even using that terminology correctly here -- in the step response graph (~90ms) is coming from the subs -- specifically, magnitude energy coming going below 26 Hz where the roll-off is accompanied by a rapid rise in GD -- likely, this is partially caused by the room itself.

Very long FIR filtering applied to the sub could level that down, but I don't think it's worth it -- and, besides, your current processor will not have that ability anyway. Much of the more problematic, unevenly extending reflection-decay energy occurs right above 150 Hz where this kind of room correction DSP isn't well suited for... In fact, more liberal use of thick, broadband acoustic treatment will be a much better choice, as other's have already advised.

As to why you prefer the "highly" elevated bass... well, I do have some "theory" as to why that may be: it may very well be in order to compensate for the broad drop in clarity starting below 500Hz or so in your measured room response. Notice that in heavily treated studio spaces they often will show "flatter" FR measurements across the board -- and if there is a preferred elevation in the bass, it's usually only up by a few dB. Think about it, if the bass response is absolutely, resoundingly crystal clear and balanced along with the rest of the spectrum, there really should be no need to elevate it by so much. Eh, I could be wrong with some of the details, but I'm sure you already got the gist of it.

This is valuable feedback - tanks a lot! What I take away from this is that I need to pay more attention to the step response next time I do the sub integration. Previously I have looked mostly on the frequency response, and phase in the crossover region. My SVS subs has a number of DSP possibilities, and I will play with them to see what I can do.

I'm not sure I follow the argument regarding the bass elevation issue. If "clarity" means early-to-late sound energy ratio, then low clarity (late dominates and builds up a sound pressure that takes time to decay?) should presumably make me *reduce* the level? The way I tuned this was playing sine tones for a number of frequencies in the bass, and listening to how loud they were perceived compared to sine tones of higher frequency. So this would be a steady state situation.

 

[ernestcarl]

This is valuable feedback - tanks a lot! What I take away from this is that I need to pay more attention to the step response next time I do the sub integration. Previously I have looked mostly on the frequency response, and phase in the crossover region. My SVS subs has a number of DSP possibilities, and I will play with them to see what I can do.

I'm not sure I follow the argument regarding the bass elevation issue. If "clarity" means early-to-late sound energy ratio, then low clarity (late dominates and builds up a sound pressure that takes time to decay?) should presumably make me *reduce* the level? The way I tuned this was playing sine tones for a number of frequencies in the bass, and listening to how loud they were perceived compared to sine tones of higher frequency. So this would be a steady state situation.

In acoustics measurement software such as REW, there are multiple views available to the user. The step response is merely one of these. REW manual explains each section to some degree… there you can also find an explanation for the “clarity” values — some if these terms and related technical use isn’t as intuitive as one may think. Other details, though, you will only find in other books/papers/articles or discussion forums such as ASR. But in general, it is recommended to have the C50 higher than 10 dB across the board.

It is true that the steady state frequency response magnitude/SPL is the most consulted, but you will miss a lot by just looking at any one single view for too long. It’s a good thing that one can often find a correlation between the measurement views; however, each one is more heavily biased at showing certain aspects of the measured IR or transfer function sooo… take each snapshot shown to you by me, or others for that matter, with a grain of salt.

Listening to generated test tones and noise can be very useful as well, but there are several other things that will influence your hearing.

For one thing, while we mostly focus on the “direct sound”, all the reflections and/or later arriving sounds (so including time aspect) shouldn’t be overlooked. It’s possible that part of the loudness difference in your sine tone test comparison with select high frequencies was influenced heavily by their relatively longer sustain in the room — this kind of information may be more readily apparent in REW’s normalized wavelet spectrogram views. Then, there’s also your listening levels (playback volume) which play an important role in determining the perceived loudness — search for the topic “equal loudness curves”.
…

Unfortunately, room acoustics and psychoacoustics are quite big subjects and it’s very easy to miss the forest for the trees… myself included here, BTW.

As always, Floyd Toole’s book should be a good start in learning more (Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms).

Interestingly, a section in the same book shows preferred room listening curves, I’ll try to find the image…

 

[janbth]

In acoustics measurement software such as REW, there are multiple views available to the user. The step response is merely one of these. REW manual explains each section to some degree… there you can also find an explanation for the “clarity” values — some if these terms and related technical use isn’t as intuitive as one may think. Other details, though, you will only find in other books/papers/articles or discussion forums such as ASR. But in general, it is recommended to have the C50 higher than 10 dB across the board.

It is true that the steady state frequency response magnitude/SPL is the most consulted, but you will miss a lot by just looking at any one single view for too long. It’s a good thing that one can often find a correlation between the measurement views; however, each one is more heavily biased at showing certain aspects of the measured IR or transfer function sooo… take each snapshot shown to you by me, or others for that matter, with a grain of salt.

Listening to generated test tones and noise can be very useful as well, but there are several other things that will influence your hearing.

For one thing, while we mostly focus on the “direct sound”, all the reflections and/or later arriving sounds (so including time aspect) shouldn’t be overlooked. It’s possible that part of the loudness difference in your sine tone test comparison with select high frequencies was influenced heavily by their relatively longer sustain in the room — this kind of information may be more readily apparent in REW’s normalized wavelet spectrogram views. Then, there’s also your listening levels (playback volume) which play an important role in determining the perceived loudness — search for the topic “equal loudness curves”.
…

Unfortunately, room acoustics and psychoacoustics are quite big subjects and it’s very easy to miss the forest for the trees… myself included here, BTW.

As always, Floyd Toole’s book should be a good start in learning more (Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms).

Interestingly, a section in the same book shows preferred room listening curves, I’ll try to find the image…

Yes, I already own a copy of Toole's book, and have read (quickly) through it once. There is a learning curve, so I plan to read it again when I get time. Is this the image you are referring to? I guess I'm in the "untrained" part of the population...

 

[ernestcarl]

Yes, I already own a copy of Toole's book, and have read (quickly) through it once. There is a learning curve, so I plan to read it again when I get time. Is this the image you are referring to? I guess I'm in the "untrained" part of the population...View attachment 270019

Yes, and thank you! Would have posted it earlier but had to leave the house and am on mobile.

 

[Blumlein 88]

Are there any calculators that output a the best spacing for wood slats with a given width?

Random distance between slats is better than having the same distance but there probably is an even better setup from an acoustic's standpoint.

Quadratic diffusors use a pattern of spacing and variable depth.

Calculate Quadratic Residue Diffusor ( given size )

www.mh-audio.nl

Here is a bit of an explanation.

What Is Quadratic Diffusion? Sound Diffusion Explained – Acoustic Fields

One method of diffusion that is widely known but not well understood is quadratic diffusion. Quadratic diffusors have been around for years.

www.acousticfields.com

 

[juliangst]

Quadratic diffusors use a pattern of spacing and variable depth.

Calculate Quadratic Residue Diffusor ( given size )

www.mh-audio.nl

Here is a bit of an explanation.

What Is Quadratic Diffusion? Sound Diffusion Explained – Acoustic Fields

One method of diffusion that is widely known but not well understood is quadratic diffusion. Quadratic diffusors have been around for years.

www.acousticfields.com

I know QRD diffusers but does this also work with the slat wall with absorption behind?
Those slat walls have equal depth slats while a QRD diffuser has different well depths

 

[Blumlein 88]

I know QRD diffusers but does this also work with the slat wall with absorption behind?
Those slat walls have equal depth slats while a QRD diffuser has different well depths

Basically I'd say no. There are some you could make with equal depth and unequal spacing that work over a narrower range, but they aren't going to work very effectively I don't think. I'd say a solid slightly curved diffuser would likely work better.

 

[juliangst]

Basically I'd say no. There are some you could make with equal depth and unequal spacing that work over a narrower range, but they aren't going to work very effectively I don't think. I'd say a solid slightly curved diffuser would likely work better.

I don’t really need the benefits of diffusion with those slat absorbers. The slats are just for the looks.
But I don’t want the slats to have any negative effects on mids/highs.
Their main purpose should be bass absorption