Room acoustics - a full breakdown and analysis

[aerosnack]

Hello all,

I am looking for some advice on room acoustics. I have done some research, lots of measurements, identification and some experimentation. Read lots of threads so I have enough idea to have an approximate grasp but far from a mastery at any single aspect.

It would be great to run this all by yourselves, to understand where I should concentrate my efforts.

Use

I currently work from home a majority of the time, and spend most of that listening to music. A very wide range from female vocals to rock through to techno and dub.

I use a combination standing / sitting desk, so there are two vertical Z positions that I rotate through in a day.


Speakers

I currently have two ADAM Audio A4vs on my desk at around 10degrees below my earline with 15degrees angular horizontal difference (meeting around 40cm behind my head). They are around 130cm apart with me being around 90cm from a virtual line connecting the two. I have blocked the LF ports. High passed at 100Hz (24dB/oct LR). They are on ‘iso acoustics’ pucks.

I also have a BK XXLS-400-FF around 1.5m from me on my right. Low passed at ~100Hz (24dB/oct LR + PEQ)

The Adam audios specifically seem troublesome to integrate from a technical standpoint, with very high vertical lobes from the directivity data in the 1-2kHz region:

https://www.audiosciencereview.com/forum/index.php?threads/adam-a4v-monitor-review.36704/


Room

Room is a dual purpose office / bedroom. See image:

There is some scattered furniture but no acoustic treatment. It is also very close to a square, so the LF peaks are pronounced. There are however two very heavy curtains with 5mm acoustic felt for the windows.


Processing

I currently use the inbuilt DSP in the A4Vs for basic room correction via the 5 PEQs, tuned via REW. This is relatively limiting. I use a separate t.racks for the sub to try and take the big modes out.

I have taken order of a MiniDSP flex 2x4, so hoping that the big room modes (<200Hz) can be tackled with Dirac. (The conversation of bass traps in the bedroom went down like a lead balloon, unsurprisingly.)


Room measurements

I annoyingly have lost the EQ’d data, but I have MMM data as below for the listening position. I have tried identifying modes as per room dimensions. Where possible I have tried verifying with slabs of Rockwool RW3.

I have also overlaid this with the vertical directivity, also showing desk bounce angle:

RT60 data:

Experiments

• Adam speaker positioning and angles. This currently works well with a wide sweet spot for me.
• Subwoofer crawling game. Found a spot with a nice flat response which increased punch significantly.
• REW EQ balancing. Gone well with a lot of removal round 150-400Hz for the Adams and correction at 1-2kHz. Smoothed the bass. Pretty happy with the results here (even if the Adam software continues to be glitchy)
• I have tried placing 50mm of melamine foam both directly under the woofer, in the desk bounce path. While the results were clear in the impulse response at lack of spike at 1.1mS, the frequency/RT60 results weren’t quite as I hoped. I also tried placing on the desk, but the foam pad had less of an effect and would have to be too large to cover all listening positions

Red: No foam. Orange with 150x200x50mm of melamine foam as per below diagram:

Next steps

• Place 50mm and 100mm absorbers in the room. The 50mm will be on the side walls (10% coverage). 100mm will be on the rear wall (12% coverage). Hopefully this will remove some of the reverberation. I have been in touch with GIK acoustics, but sadly their products are coming out a little expensive (based in UK).
• Place 600x1200x400mm absorber on top of wardrobe. This is the only WAF location identified.
• Tune with Dirac, trying both full range and up to 300Hz.

Questions

• Do the next steps make sense? Would 10/12% wall treatment be enough for an audible difference?
• How should I tackle desk bounce? Is it worth it in the grand scheme of things?
• What will give the biggest benefits for listening enjoyment / bang for buck?

Thanks all

 

[aerosnack]

Interesting, no comments. Perhaps it’s too wide of a question set.

Any feedback is appreciated!

 

[amirm]

Next steps

• Place 50mm and 100mm absorbers in the room. The 50mm will be on the side walls (10% coverage). 100mm will be on the rear wall (12% coverage). Hopefully this will remove some of the reverberation. I have been in touch with GIK acoustics, but sadly their products are coming out a little expensive (based in UK).
• Place 600x1200x400mm absorber on top of wardrobe. This is the only WAF location identified.
• Tune with Dirac, trying both full range and up to 300Hz.

Dirac is your best friend there for low frequencies. The other bits you list will be like spitting in the wind in that region. They simply won't be effective below 100 to 200 Hz.

 

[mitchco]

I was just about to post Digital Room Correction, Amir beat me to it. Think wavelengths. Bass traps, even tuned Helmholtz resonators do (very) little below 100 Hz.

While Dirac is a goto, it does not perform low frequency excess phase correction like other DRC software does. All rooms exhibit non-minimum phase response and John's REW whitepaper does a good job of explaining this: https://www.roomeqwizard.com/help/help_en-GB/html/minimumphase.html

Most folks who have performed an A/B comparison, with and without excess phase correction comment that the clarity of the bass response is improved with excess phase correction in addition to the bass sounding more even. But it is easy to determine ones preference using the right tools and level matching filters with and without excess phase correction.

Good luck!

 

[Axo1989]

Hello all,

I am looking for some advice on room acoustics. I have done some research, lots of measurements, identification and some experimentation. Read lots of threads so I have enough idea to have an approximate grasp but far from a mastery at any single aspect.

It would be great to run this all by yourselves, to understand where I should concentrate my efforts.

Use

I currently work from home a majority of the time, and spend most of that listening to music. A very wide range from female vocals to rock through to techno and dub.

I use a combination standing / sitting desk, so there are two vertical Z positions that I rotate through in a day.


Speakers

I currently have two ADAM Audio A4vs on my desk at around 10degrees below my earline with 15degrees angular horizontal difference (meeting around 40cm behind my head). They are around 130cm apart with me being around 90cm from a virtual line connecting the two. I have blocked the LF ports. High passed at 100Hz (24dB/oct LR). They are on ‘iso acoustics’ pucks.

I also have a BK XXLS-400-FF around 1.5m from me on my right. Low passed at ~100Hz (24dB/oct LR + PEQ)

The Adam audios specifically seem troublesome to integrate from a technical standpoint, with very high vertical lobes from the directivity data in the 1-2kHz region:

https://www.audiosciencereview.com/forum/index.php?threads/adam-a4v-monitor-review.36704/


Room

Room is a dual purpose office / bedroom. See image:
View attachment 330882


There is some scattered furniture but no acoustic treatment. It is also very close to a square, so the LF peaks are pronounced. There are however two very heavy curtains with 5mm acoustic felt for the windows.


Processing

I currently use the inbuilt DSP in the A4Vs for basic room correction via the 5 PEQs, tuned via REW. This is relatively limiting. I use a separate t.racks for the sub to try and take the big modes out.

I have taken order of a MiniDSP flex 2x4, so hoping that the big room modes (<200Hz) can be tackled with Dirac. (The conversation of bass traps in the bedroom went down like a lead balloon, unsurprisingly.)


Room measurements

I annoyingly have lost the EQ’d data, but I have MMM data as below for the listening position. I have tried identifying modes as per room dimensions. Where possible I have tried verifying with slabs of Rockwool RW3.
View attachment 330884
I have also overlaid this with the vertical directivity, also showing desk bounce angle:
View attachment 330886

RT60 data:
View attachment 330885

Experiments

• Adam speaker positioning and angles. This currently works well with a wide sweet spot for me.
• Subwoofer crawling game. Found a spot with a nice flat response which increased punch significantly.
• REW EQ balancing. Gone well with a lot of removal round 150-400Hz for the Adams and correction at 1-2kHz. Smoothed the bass. Pretty happy with the results here (even if the Adam software continues to be glitchy)
• I have tried placing 50mm of melamine foam both directly under the woofer, in the desk bounce path. While the results were clear in the impulse response at lack of spike at 1.1mS, the frequency/RT60 results weren’t quite as I hoped. I also tried placing on the desk, but the foam pad had less of an effect and would have to be too large to cover all listening positions

View attachment 330883

Red: No foam. Orange with 150x200x50mm of melamine foam as per below diagram:

View attachment 330881
Next steps

• Place 50mm and 100mm absorbers in the room. The 50mm will be on the side walls (10% coverage). 100mm will be on the rear wall (12% coverage). Hopefully this will remove some of the reverberation. I have been in touch with GIK acoustics, but sadly their products are coming out a little expensive (based in UK).
• Place 600x1200x400mm absorber on top of wardrobe. This is the only WAF location identified.
• Tune with Dirac, trying both full range and up to 300Hz.

Questions

• Do the next steps make sense? Would 10/12% wall treatment be enough for an audible difference?
• How should I tackle desk bounce? Is it worth it in the grand scheme of things?
• What will give the biggest benefits for listening enjoyment / bang for buck?

Thanks all

You've investigated fairly thoroughly, I like the superimposed FR and directivity example.

Also looks like a nice work-space facing the bay window.

Despite doing a lot of work from home, I don't actually have desktop speakers, rather I position floor-standers close to the front wall and the desk further back. I expect there are budget reasons why you wouldn't do the former and living space reasons why you may not do the latter. Just means I don't have experience to offer integrating small two-ways with a sub as you are doing.

That said, how did your sub crawl work? Those two bass suckouts between 40 and 100 Hz may not really respond to a boost from EQ. A different sub position and an additional sub would be where I would go next (if your budget stretches sufficiently for the latter). Moving the desk back toward the bed may be worth trying (I expect you've played with the REW room modelling tool, if not I'd do that first). Amcoustic offers another good tool to investigate room modes (which I used in conjunction with the room modelling).

 

[Alice of Old Vincennes]

I suspect any solution shall require subtle subterfuge with a normal mate not attuned to our eccentricities.

 

[Keith_W]

Well done for investigating your room issues so thoroughly. I did a quick calculation of your Schroeder frequency using the following inputs: 5.2m x 4.7m x 2.7m and T30 of 0.4s, and came up with Fs = 159.6Hz, so yeah it looks like your graph is spot on.

People much more knowledgeable than myself have already recommended you use DSP. I was going to jump in myself and say, wait for your MiniDSP and use Dirac.

BTW is there a good reason we are not recommending membrane absorbers for the OP?

 

[Bjorn]

50 mm absorbers are too thin and will only work in the highs. Broadband absorption panels should be placed surgically at reflection points and not just spread around the room. If you have the needed distance, using quality diffusion at some places is generally better than only using absorption. Mainly for later arriving reflections. If you want to go with low cost, you could build some poly diffusers. They don't offer temporal diffusion, but offer good spatial diffusion if they are big enough and used together correctly.

For the frequency area below 100 Hz it requires much surface treatment for it to be highly effective. Unless you can do that, using EQ here is a decent compromis. But above 100 Hz I would recommend treatment if possible.

 

[aerosnack]

Wow, thanks all. Looks as if all the buses arrived at once!

Amirm: That’s the same I’ve arrived at. I’m hoping to manage the 200hz+ decay by some absorption, and below (and maybe above too?) with Dirac. The MiniDSP has arrived now, I just need to get it set up.

Mitchco: low frequency excess phase correction sounds interesting, I haven’t come across this concept before. I will try it out but looks like a fun rabbit hole!

Axo1989: thanks, it’s got a good view too! The speaker positioning is pretty fixed, due to needing access behind for curtains and keeping room behind too.

The frequency data for the bass is actually pre sub-crawl. I used the method of stacking the woofer at head height in the chair then taking small area MIM at woofer positions. The new frequency response is about half the distance from peaks to nulls. And being far more defined due to less excitation. Much happier with now! The data does correlate nicely with the room simulations and representative wavelengths / distances.

Alice: and a great deal of time!

Keith_W: thanks! I have used this example as ‘live’ learning. Quite enjoyed trying to figure it all out.

Bjorn: cheers. I will look into the poly diffusers, they look super straightforward to make. Will have to assess acceptance factors.

In terms of first reflections, there is unfortunately no possibility to cover those due to furniture. They are also quite a long way off axis, so I would imagine energy would be relatively small. It was the rear wall which I was most concerned about. Placing a pair of 0.6 x 1.2 x 0.15 absorbers behind my head does seem to clear up the ‘wall bounce’ quite a bit. Definitely prefer the clearer forward facing image. Hence hoping to capture this effect with a less intrusive option with the installed 50/100 panels

 

[Bjorn]

Yeah, a 15 cm absorber behind your head is a good choice. Later arriving reflections from side walls are less detrimental but still not good with various music material IMO. But furnitures, bookshelves or something to break it up is surely better than nothing.

 

[fpitas]

It should be noted, your brain separates out reflected sound from direct sound, going by time-of-arrival. So, it's generally not necessary or desirable to EQ down reflections. I placed some absorbers and a lot of diffusers, but that gets into personal taste.

 

[aerosnack]

Had a few plays with Dirac Live now, and it definitely seems to help across the board! Aside from being a massively powerful EQ, the imaging is much wider with definition taking a step up too.

So far I prefer the full frequency correction, but I am still curious about absorption taking up some of the heavy lifting for the MF/HF range.

For those that are curious, here is the target curve I am playing with at the moment (seems to be a thing amongst Dirac Live users? Will play with it more, perhaps also delay and XO at some point. When I can be bothered to re-do the measurements again anyway!

And the impulse response. Although I can't seem to work out why the corrected X axis is offset by 8ms? You can see the rear wall slap at approx. 25ms in the original data though. That should be reduced slightly as we 'busy up' that wall with coincidental diffusers and absorbers.

 

[Alice of Old Vincennes]

Had a few plays with Dirac Live now, and it definitely seems to help across the board! Aside from being a massively powerful EQ, the imaging is much wider with definition taking a step up too.

So far I prefer the full frequency correction, but I am still curious about absorption taking up some of the heavy lifting for the MF/HF range.

For those that are curious, here is the target curve I am playing with at the moment (seems to be a thing amongst Dirac Live users? Will play with it more, perhaps also delay and XO at some point. When I can be bothered to re-do the measurements again anyway!

View attachment 334796

And the impulse response. Although I can't seem to work out why the corrected X axis is offset by 8ms? You can see the rear wall slap at approx. 25ms in the original data though. That should be reduced slightly as we 'busy up' that wall with coincidental diffusers and absorbers.

View attachment 334797

You may over dampen higher frequencies. I stopped correction at 5000 hz and may lower. Wide dispersion speakers sound much better when reflected sound arrives later. Consistent with how our hearing evolved. Amir warns about excessive damping for wide dispersion speakers repeatedly in his speaker reviews.

 

[markus]

While Dirac is a goto, it does not perform low frequency excess phase correction

Dirac Live ART does.

 

[keks8430]

While Dirac is a goto, it does not perform low frequency excess phase correction like other DRC software does.

Interesting, what is your source?

 

[reggiegasket]

Interesting setup, and some great data. I've no dsp experience, so feel free to dismiss this entirely but I've always found speakers quite close to my ears hard to set up, and have tended to push them further and further away to sort it out. If it was my room (and it looks like a lovely room by the way...) I'd experiment with the speakers off the table, wider apart, on stands, and without the window glass directly behind them. I'd hope this would widen the soundstage, integrate better with the sub (and also work better if you take a break on the bed...) and not have glass directly behind them. With the speakers in front of a solid wall, as far into the room as is practical, I'd then put some diffusors on said wall, as in my experience what sits directly behind a speaker is a critical part of the jigsaw, and a source of many problems. I'd not worry too much about the rest of the room, as it's pretty big.

If the little ADAMs don't play nicely with that setup then I'd try some bigger speakers/drivers, to give it more oomph and presence, and allow me to rely less on the sub, either by lowering the low pass frequency, or just turning the sub volume down a bit.

That's my 10p

 

[Keith_W]

Dirac Live ART does.

I read an old article by @mitchco - Dirac Live walkthrough where I note that Dirac Live creates mixed phase filters. I asked a follow-up question in that thread whether this was still true in 2024. I don't know the answer.

Is Dirac Live ART available as a software package for PC/Mac? I would imagine it would be pretty difficult to perform that "ART" function if you don't have FIR filters.

 

[DWPress]

He's got a miniDSP 2x4 - just on the hardware side it is pretty limited in what it can do in the bass region and baked in basic Dirac is the tool he has to work with. If source is computer some additional EQ work can be done before to help the minDSP do the best it can.

Just a thought for the space if WAF works: Maybe a small tri-fold room divider like used in office spaces (minimal frequency absorbtion) but could potentially help tighten up the space and rear wall signal return. Fold it up and slide it under the bed at the end of the day.

 

[Kal Rubinson]

Is Dirac Live ART available as a software package for PC/Mac?

Yup.

 

[DWPress]

Yup.

Are you sure Kal? I just looked, hoping for a trial or beta. Looks like ART is still only available via hardware.