How to get started with room treatments for the home

[sigbergaudio]

If you could just tell me what you think? Just in simple terms if it room related or the system? The way I see it that the problems are room based although two measurements from the (3 way are different).

It's hard to just read stuff out of these without having first hand information about the system(s), but I'm reasonably confident this is mostly the room. You won't get decay of several hundred milliseconds from the system alone (not from time alignment / DSP issues either).

 

[Trdat]

It's hard to just read stuff out of these without having first hand information about the system(s), but I'm reasonably confident this is mostly the room. You won't get decay of several hundred milliseconds from the system alone (not from time alignment / DSP issues either).

Yeh ofcourse we need to analyze so many of the other graphs, understand the room modes and SBIR and match it to the measurements. But at least now I know the system is not the issue at hand. In any case I will time align those subs individually.

I think next is targeted helmontz or some type of diaphragmatic absorber on the back wall which is 4m back to front. Where to place them is the problem. Lol!

I am adding 2d binary template on two absorber panels though just placing them infront rather making a sealed new absorber which might not actually add more for low frequency absorbtion.

Appreciate your input.

 

[radix]

Mate, if no one tells you its difficult to understand so many of the nuances of REW. Thats a great help thanks so much, so simple yet so complicated.

Ill look into what is audible regarding delay, I am also curious now so I will research deeper but will take on good merit that they are not a major concern.

I am folowing your posts carefully and please add your process and results(the 6db dip) so we can learn.

@Trdat, I figured there's likely lots of people in my boat about how to get started with room treatments and how to use the wonderful tool of REW to "do the right thing." I'm trying to find a process to identify the biggest band-for-the-buck treatments rather than just shotgun absorbers or baffles around the room.

Here's the process I've used so far my HT room. I do a REW SPL measurement and look for significant dips (e.g. 6dB or more below what looks more-or-less like the average). In my case, I have one around 330 Hz. I then setup the REW generator for that frequency and then use the RTA (real-time analysis) to measure the response. There should be a single peak right under where the SPL measurement was. I then take a thick blanket or pillow and move around the room and check large flat spots to see if that's contributing to the null. I was able to find a couple spots that seem to make a difference. Adding treatments also changes things, so in some cases it partially fixed one thing, then seemed to shift other problems around.

Right now, I only have some inexpensive 1/2" panels ($31 for 12), but the are good enough that when I put them up with double-sided tape I can see an SPL difference if I re-measure with REW. I just ordered a 2' x 3' x 2" ATS panel ($52), which I think will be easier to walk around with and place on a wall ceiling area to test for reflections. ATS Acoustics sells a bunch of DIY materials so you could make your own panel for less than buying it. I think if I need several panels, I might order materials from them (at least the Rockwool) and make them myself.

There are supposed to be some more principled methods, e.g. the string technique [Getting Started with REW, pp. 82ff], but I could not make that work. The way they convert time to distance from the ETC did not seem to work on a Mac.

Marc

 

[radix]

There are supposed to be some more principled methods, e.g. the string technique [Getting Started with REW, pp. 82ff], but I could not make that work. The way they convert time to distance from the ETC did not seem to work on a Mac.

Marc

OK, with some more work I figured out the instructions on p. 81 about how to measure the distance of an ETC peak.

To determine the distance to a peak, place the cursor on the vertical axis at the height of the peak you want to measure, hold the Control Key down with your left hand, depress the right mouse button with your other hand, and drag the cursor horizontally to the top of the peak.

So, it turns out you need to start with the cursor just to the right of the vertical axis, so it's over the white background. Then, when you CTRL-Right click you will see a red cursor appear and can then drag it to the measurement point ( you are dragging over the distance you want to measure). It is a bit finicky about where the cursor starts and it is hard to get it exactly at the vertical axis.

 

[sigbergaudio]

Another tip is to measure not just in ear height, and unless you only care about one sweet spot, you also need to measure in several positions. In the lower frequencies you actually "listen" with your body as well, and you can end up with something that doesn't seem right if you only equalize or treat based on ear height measurements.

Here's an example of three measurements in the sweet spot. Green is ear height, yellow is at the stomach, and red is butt/knee height. The thick blue is average of the three.

Even though the green (ear height) seems low from 130-300hz, the stomach is roughly correct, and knee height actually has too much energy. This all averages out to just about right, and that's also how it sounds. Increasing the 130-300hz level based on the ear height measurement is not a good idea in this example. Notice how a dip in one position may be a peak in a slightly different position / elevation, for instance at 110hz.

Worth experimenting with to at least understand how things measure differently just a feet or two apart, both horizontally and vertically

 

[radix]

Another tip is to measure not just in ear height, and unless you only care about one sweet spot, you also need to measure in several positions. In the lower frequencies you actually "listen" with your body as well, and you can end up with something that doesn't seem right if you only equalize or treat based on ear height measurements.

Here's an example of three measurements in the sweet spot. Green is ear height, yellow is at the stomach, and red is butt/knee height. The thick blue is average of the three.

Even though the green (ear height) seems low from 130-300hz, the stomach is roughly correct, and knee height actually has too much energy. This all averages out to just about right, and that's also how it sounds. Increasing the 130-300hz level based on the ear height measurement is not a good idea in this example. Notice how a dip in one position may be a peak in a slightly different position / elevation, for instance at 110hz.

Worth experimenting with to at least understand how things measure differently just a feet or two apart, both horizontally and vertically

View attachment 174363

Thank you, that is a very interesting observation for the lower frequencies. In my stereo room (the living room), I really appreciate the F228be as I really can feel it, even in a largish room. I'll need to check this out in the HT room with the R3s and subs.

 

[Trdat]

@Trdat, I figured there's likely lots of people in my boat about how to get started with room treatments and how to use the wonderful tool of REW to "do the right thing." I'm trying to find a process to identify the biggest band-for-the-buck treatments rather than just shotgun absorbers or baffles around the room.

Here's the process I've used so far my HT room. I do a REW SPL measurement and look for significant dips (e.g. 6dB or more below what looks more-or-less like the average). In my case, I have one around 330 Hz. I then setup the REW generator for that frequency and then use the RTA (real-time analysis) to measure the response. There should be a single peak right under where the SPL measurement was. I then take a thick blanket or pillow and move around the room and check large flat spots to see if that's contributing to the null. I was able to find a couple spots that seem to make a difference. Adding treatments also changes things, so in some cases it partially fixed one thing, then seemed to shift other problems around.

Right now, I only have some inexpensive 1/2" panels ($31 for 12), but the are good enough that when I put them up with double-sided tape I can see an SPL difference if I re-measure with REW. I just ordered a 2' x 3' x 2" ATS panel ($52), which I think will be easier to walk around with and place on a wall ceiling area to test for reflections. ATS Acoustics sells a bunch of DIY materials so you could make your own panel for less than buying it. I think if I need several panels, I might order materials from them (at least the Rockwool) and make them myself.

There are supposed to be some more principled methods, e.g. the string technique [Getting Started with REW, pp. 82ff], but I could not make that work. The way they convert time to distance from the ETC did not seem to work on a Mac.

Marc

My mistake was to treat with a broadband approach but back then I was an amatuer. Now, the treated room looks nice and chopping and changing it just a hassle. I used 90kgm3 and it seems my ceiling coud with a 30cm gap isn't doing much.

I have SBIR Issues that I have addressed with absorbtion in its exact respective nearby wall but for some reason the attenuation is very little.

I haven't yet solved room modes, even though I have followed Abdo's mentioned opposite polarity subwoofer technique to the tether but I presume there are nuances which are learnt with experience so I need more time to get it right.

RT60 is good. Taken from the RT60 measurement or the from the filtered response which we are meant to take it from for small rooms
Filtered Impulse response is average, slopping downward then popping back up with a consistent drop off this is acording to Kvalvoll another top dog in interpreting measurements. Reflections are again average, nothing that stands out and enough smaller reflections that aid wanted reverb tail.
Decay graph is consistent with average specular reflection within each slice interval. Of course this is my guess from interpreting the measurements.
But the decay times from waterfall is herendous as stated by Sigerberg, typical room issues.
And I am not sure about the impulse response and time alignment, I just can't get the graph to look normal in REW and interpreting them is also difficult.

My take on the situation is as always horrible roommodes that need to be worked on. Everything else isn't too bad.

 

[Trdat]

OK, with some more work I figured out the instructions on p. 81 about how to measure the distance of an ETC peak.


So, it turns out you need to start with the cursor just to the right of the vertical axis, so it's over the white background. Then, when you CTRL-Right click you will see a red cursor appear and can then drag it to the measurement point ( you are dragging over the distance you want to measure). It is a bit finicky about where the cursor starts and it is hard to get it exactly at the vertical axis.

I remember reading about this technique in gearspace, its how you identify from where the reflection is coming from, thanks for reminding me.

 

[abdo123]

Another tip is to measure not just in ear height, and unless you only care about one sweet spot, you also need to measure in several positions. In the lower frequencies you actually "listen" with your body as well, and you can end up with something that doesn't seem right if you only equalize or treat based on ear height measurements.

Here's an example of three measurements in the sweet spot. Green is ear height, yellow is at the stomach, and red is butt/knee height. The thick blue is average of the three.

Even though the green (ear height) seems low from 130-300hz, the stomach is roughly correct, and knee height actually has too much energy. This all averages out to just about right, and that's also how it sounds. Increasing the 130-300hz level based on the ear height measurement is not a good idea in this example. Notice how a dip in one position may be a peak in a slightly different position / elevation, for instance at 110hz.

Worth experimenting with to at least understand how things measure differently just a feet or two apart, both horizontally and vertically

View attachment 174363

I don't think the bodily impact is relevant above 100Hz (as evident by Harman's research on headphones) but it's a very interesting approach.

 

[abdo123]

I haven't yet solved room modes, even though I have followed Abdo's mentioned opposite polarity subwoofer technique to the tether but I presume there are nuances which are learnt with experience so I need more time to get it right.

Other than the dip at 60Hz your sub-bass response looks phenomenal! Were you expecting something even smoother?

Btw my approach is about minimizing seat to seat variation in the sub-bass within a small room, you will still need EQ to get a smooth response. My approach (it's actually called Welti's approach, the man who did the research) makes it so you have a really uniform response once EQ is applied. Only the overall volume would signficantly change as you move along in the room, the response will not (theoretically).

Also the dips will be manageable and less deep in amplitude, for example that 60Hz dip is barely -10dB, you can definitely boost it up.

If you deal with the third mode, and use global EQ, that's pretty much as good as it gets in a small room. All other options are compromises in one way or another. There is not really much else about it.

I have SBIR Issues that I have addressed with absorbtion in its exact respective nearby wall but for some reason the attenuation is very little.

Did you simulate the absorber? and what percentage of the surface area behind the speaker did you cover?

Can you share more details on the situation? (the frequency, before and after, the absorber).

 

[sigbergaudio]

I don't think the bodily impact is relevant above 100Hz (as evident by Harman's research on headphones) but it's a very interesting approach.

Opinions differ.

 

[hege]

I remember reading about this technique in gearspace, its how you identify from where the reflection is coming from, thanks for reminding me.

I find it easier to just brute force. Keep a panel near the mic on some side and observe what happens on ETC. Keep backing away and moving around until you find the spot on the wall that removes a large peak. Found many reflections on ceiling/wall borders, concave crown molding seem to be great reflector..

 

[Trdat]

Other than the dip at 60Hz your sub-bass response looks phenomenal! Were you expecting something even smoother?

Im double checking all my response graphs and yes there all somewhat similar with no tmuch deviation, I do have a dip right under 40hz which can be cut off or rolled off. The 60hz is persistent with the two added subwoofers but my prior measurements dont have it in any case your correct EQ can solve that. But I have a SBIR issue between 120 and 140 and around 170 on the right side that is in all measurements in some form. Yes, from a frequency response standpont it isn't bad, as I don't know what constitutes as good I always think its bad but I get your point. But from a decay perspective the waterfall graph does show some long decay times across the board. Must be the 4m front to back room. I aslo have a massive 120 litre cabinet for my front left rights so it can't get any closer to the front wall. Lol.

Btw my approach is about minimizing seat to seat variation in the sub-bass within a small room, you will still need EQ to get a smooth response. My approach (it's actually called Welti's approach, the man who did the research) makes it so you have a really uniform response once EQ is applied. Only the overall volume would signficantly change as you move along in the room, the response will not (theoretically).

Yes, it is Welti's approach that why I added Abdo's "mentioned" technique but you were able to explain it in a way that was understandable. So when you say uniform response the technique doesn't attenuate room modes such as the dips or peaks and only provides uniform repsonse...? I thought it should.

Also the dips will be manageable and less deep in amplitude, for example that 60Hz dip is barely -10dB, you can definitely boost it up.

I will do another test and can manually EQ with Equalizer APO some dips are still showing after Audiolense DSP. But what about the dip around 120hz can I deal with that with EQ as well?

If you deal with the third mode, and use global EQ, that's pretty much as good as it gets in a small room. All other options are compromises in one way or another. There is not really much else about it.

I guess your right, I mean how much can I reduce decay even with targeted bass trapping probably not much attenuation.

I am adding 2D Binary template(RPG bad panels) to the front of a few bass traps, about 2 metres from listening spot should be okay. If you got any opinions on this it would be appreciated but I am already getting them cut out and unfortunately I only found a nice looking MDF laminate in 16mm hopefully that is not a problem, I know they are in 9mm or 12mm.

Did you simulate the absorber? and what percentage of the surface area behind the speaker did you cover?

Can you share more details on the situation? (the frequency, before and after, the absorber).

I am analysing the measurements and I can't tell if there has been any attenuation from adding a panel to the back of the speaker. The left measurement looks worse but the right looks somwhat attenuated after adding the panel.

I also probably have to double and triple check the issue is front wall SBIR and that I placed the panel in the right spot and retake measurements with and without panel and get back to you. But below I have added what I got.

No, I did not simulate the absorber. I only added another approx .5 to 1m2 to the front wall which already had about 2m2 but if you include the corner panels then there is about 4m2 plus the added panel making it about 5m2.

Basically if you let me know what kind of attenuation I am supposed to get, I will redo measurement taking that in mind and ensure a accurate measurement. I am not saying what I got is not accurate but there is always anomalies and once you spot them you kind of know where your going wrong.

First two is left side im looking at the 120hz before in green and after in the red.

These two are right side I am looking at around 170hz purple is before and green is after.

 

[abdo123]

Im double checking all my response graphs and yes there all somewhat similar with no tmuch deviation, I do have a dip right under 40hz which can be cut off or rolled off. The 60hz is persistent with the two added subwoofers but my prior measurements dont have it in any case your correct EQ can solve that. But I have a SBIR issue between 120 and 140 and around 170 on the right side that is in all measurements in some form. Yes, from a frequency response standpont it isn't bad, as I don't know what constitutes as good I always think its bad but I get your point. But from a decay perspective the waterfall graph does show some long decay times across the board. Must be the 4m front to back room. I aslo have a massive 120 litre cabinet for my front left rights so it can't get any closer to the front wall. Lol.

Yes, it is Welti's approach that why I added Abdo's "mentioned" technique but you were able to explain it in a way that was understandable. So when you say uniform response the technique doesn't attenuate room modes such as the dips or peaks and only provides uniform repsonse...? I thought it should.

I will do another test and can manually EQ with Equalizer APO some dips are still showing after Audiolense DSP. But what about the dip around 120hz can I deal with that with EQ as well?

I guess your right, I mean how much can I reduce decay even with targeted bass trapping probably not much attenuation.

I am adding 2D Binary template(RPG bad panels) to the front of a few bass traps, about 2 metres from listening spot should be okay. If you got any opinions on this it would be appreciated but I am already getting them cut out and unfortunately I only found a nice looking MDF laminate in 16mm hopefully that is not a problem, I know they are in 9mm or 12mm.

I am analysing the measurements and I can't tell if there has been any attenuation from adding a panel to the back of the speaker. The left measurement looks worse but the right looks somwhat attenuated after adding the panel.

I also probably have to double and triple check the issue is front wall SBIR and that I placed the panel in the right spot and retake measurements with and without panel and get back to you. But below I have added what I got.

No, I did not simulate the absorber. I only added another approx .5 to 1m2 to the front wall which already had about 2m2 but if you include the corner panels then there is about 4m2 plus the added panel making it about 5m2.

Basically if you let me know what kind of attenuation I am supposed to get, I will redo measurement taking that in mind and ensure a accurate measurement. I am not saying what I got is not accurate but there is always anomalies and once you spot them you kind of know where your going wrong.

First two is left side im looking at the 120hz before in green and after in the red.

View attachment 174461
View attachment 174462

These two are right side I am looking at around 170hz purple is before and green is after.

View attachment 174463View attachment 174464

Seems like you’re on the right track with regards to room modes, ~10 dB of boost is okay. I would not attempt anything higher though.

The Welti approach minimises overall variation and swings, but it will not get rid of room modes, these are physical qualities of the room. You will have smoother overall response, and less variation between seats. But you will always have more energy in some frequencies than others, it will just be much less offensive. Luckily with the Welti approach EQ is incredibly effective, and you have 12 dB of (anechoic) headroom extra that you wouldn’t have with just one subwoofer.

For absorption below 200Hz for the SBIR you need at least 30 cm thick panels, and they have to be not so dense otherwise sound will not penetrate all the way through.

That’s why I suggested originally to let the subwoofers deal with the first order SBIR.