Acoustic treatment control room

Started by Ratty_7 on 18 April 2008. 47 replies, 2008–2016. In the Library under Acoustic treatment.

Originally posted at johnlsayers.com, topic 10495.

Set the limits from 15Hz to 20k, although they said 25K but i'm sticking to the range of my DynAudios
WHY??? Which DynAudio speakers do you have? And why do you not want to follow instructions from the experts? There's a reason why we tell you to run the test out to 22k (me) or 25k (REW website). It has nothing at all to do with the point where your speaker manufacturer says the response of their speakers is down by 3dB... What reason could you possibly have for not running the test to 25 kHz?
Calibrated the system again to 86db
Not according to your graph! take a look at it: In reality, you ran your test at a level of around 70 dB, probably a bit less: You can clearly see that there is nothing on that graph even vaguely close to 80 dB, let alone 86 dB! Also, you did only one single test there, using both speakers at once. What I had actually asked you to do is: "This time, [i]run three tests with the mic positioned precisely in the middle of your head, or rather, in the middle of where your head will be when you are mixing, exactly on the room center line, with the tip at exactly 1.2m above the floor, and the mic angled upwards at about 60°. Do one test with just the left speaker, one with just the right speaker, and one with both speakers. Do not adjust anything between tests! Don't even touch the system at all, except to turn of and on the speakers.[/i] So first please re-calibrate everything using the REW pink noise full-range test signal at 80 dB for each individual speaker, and using a good quality hand-held sound-level meter to do that, confirming that REW is set to the same level as what your meter is showing. You can find that test signal on the "preferences" menu. Under "Levels" select the option "Use main speakers to Check/Set levels", then hit the "Check Levels" button. Adjust the speaker volume only until you see 80 dB(C) on your sound level meter. Once you have calibrated, run those three tests I mentioned above, with REW set to sweep from 15 Hz. to 25 kHz. One test on left speaker, one on right speaker, one on both speakers. Label them. Then upload the resulting MDAT file to some file-sharing service, such as Dropbox, and post the link here so we can download the file and look at it properly. The MDAT file is too big to post on the forum, so you'll have to use a file sharing service. - Stuart -
Hi Stuart Speakers are DynAudio BM6A Mk2 I will do the tests again following exactly the experts advice 15-25K As you know I only have a ECM8000 which you don't like Any advice on a good quality sound level meter as I only have an older analogue Realistic/RadioShack one...? Mike
Speakers are DynAudio BM6A Mk2
Great! Nice speakers. And the manual says that at 21 kHz they are only 3 dB down, so they will easily hit 23 or 24 kHz. I'm betting your interface will run out of steam before the speakers do. Assuming your interface is set for 44.1 kHz sampling, they will hit the wall at about 22 kHz, according to Nyquist.
As you know I only have a ECM8000 which you don't like
It's not that I don't like it: I actually own one myself. It's just that it isn't very accurate, and I have come across several of those that have low-frequency "hum" problems. It seems yours does not have that problem, so it should be OK.
Any advice on a good quality sound level meter as I only have an older analogue Realistic/RadioShack one...?
Actually, those old RS meters are not too bad at all. Not perfect, but still not bad. However, if you are looking for a good meter, get an Extech. The 407732 is good, and the 407750 is even nicer (but pricey...) . Galaxy makes some good meters too. But be careful of the cheap Chinese junk that goes for US$ 50 and even cheaper. I use a Phonic PAA3 mostly, but that's probably overkill for what you need. It's way more than just a sound level meter (but it does that too!). - Stuart -
Thanks mate :) ps I do have one question. If you only had an older analogue Radio shack SPL meter and the ECM8000 mic which would you use...?
If you only had an older analogue Radio shack SPL meter and the ECM8000 mic which would you use...?
I would use the ECM8000 to do the actual measurements with REW, and I would use the Radio Shack meter to calibrate REW. Even though some radio Shack meters do have a "mic out", I would not use that for REW: the ECM8000 is much better for that job. The mic on the Radio Shack meter is fine for simple SPL measurements, but not so good for precision acoustic measurements. - Stuart -
Hi Stu Gotcha.... I have printed your instructions, highlighted all the points and I'll do my best. Mike :)
Hi Stuart Have re-done REW as per your latest instructions saved as Mdat file which is now in Dropbox For me to share files in D/B... I need an email address ? cheers Mike
For me to share files in D/B... I need an email address ?
Post the link to the DropBox file here on the forum. That way anyone else can download it too, if they feel like it (at least until it expires). - Stuart -
Soundman2020 wrote:
For me to share files in D/B... I need an email address ?
Post the link to the DropBox file here on the forum. That way anyone else can download it too, if they feel like it (at least until it expires). - Stuart -
Here is the link https://www.dropbox.com/s/4kb7exo41obo4 ... .mdat?dl=0 Cheers Mike :)
OK, I'm downloading it, and I'll take a look... - Stuart -
First, I noticed that you removed the images from a couple of your earlier posts, with the incorrect graph settings. Please put those back in again. One of the major purposes of the forum is so that other people can learn from your thread in the future, just like you have learned from the threads of other members who built their places before you. If you remove the images or text from your thread, then you take away that possibility from them. If they cannot see the images that we are talking about in the thread, then the thread is no use to them, and they won't be able to learn anything form it. So please put the missing images back again. Second, the REW data is much better but your levels are still too low:
Image not preserved: ratty-7-levels-low.jpg
As you can see, the levels for both your left and right speakers are around 70 dB(C), but we need them to be around 80 dB(C): Are you certain that you calibrated everything correctly? Are you sure you have both your RadioShack meter and REW set to "C" weighting and "Slow" response?. If the levels are too low, then your speakers won't be triggering all the relevant modes, so when you are actually mixing at normal levels, you'll be getting a different response from what was in the tests, since those modes will be triggered. Next, there is noise in your data. First, the waterfall plot for left speaker (at least, I'm assuming the green data is left, but I'm not certain as you did not label it...)
Image not preserved: ratty-7-wf-left-500-16k.jpg
You can clearly see the noise that happened in the room while the test was running, around 2 kHz. Now the plot for the right speaker:
Image not preserved: ratty-7-wf-right-500-16k.jpg
As you can see, there was no noise for that test. It is nice and clean. And finally both speakers:
Image not preserved: ratty-7-wf-lr-500-16k.jpg
Once again there is noise, but in a very different part of the spectrum, around 4kHz to 8kHz. So it's clear that you were still in the room when the test were running, and you were making a noise. You need to either remain absolutely silent for several seconds before and after the test runs, or you need to leave the room and use the "Start delay" feature in REW, which you can set to give you enough time to leave the room before the test runs, so that you are not affecting the tests in any way. The good news is that the overall frequency response in your room is already reasonably... except for the low end. Are you using a sub-woofer for these tests? What are the settings that you have on the rear-panel controls of the speakers? The bad new is that the room is already over-treated and too dead, but you still need some more bass trapping. So you will have to modify some of the treatment that you already did, to return some life to the room. What treatment do you have in there already? Please describe that in detail. Also, post photos of the room exactly as it is right now. - Stuart -
Hi Stuart Ok I am getting there, however 1. I am sorry for removing my images but unfortunately, they were deleted off the computer. I won't remove any graphs ever again, so that I and others can learn from my mistakes. 2. Yes I was in the room however I swear I never made a sound. I will do what you say in terms of the REW auto measurement taking to make sure all is good. I will take some pics or please go to http://www.mcsrecording.com.au for all to see my control room and tracking and other ISO's 3. I did have my SPL meter set to "C" weighting but "Fast" so I will change that REW was "C" weighted and "Slow" 4. When I tool the measurements of both speakers the measurement was higher than when I did Left and Right speakers. 5. Also, when the first measurement was taken I only had -1.3db headroom left. Previously to that when I took the measurements REW said there was clipping but I could see no distortion....so what do you suggest as my system in terms of mic input volume and speaker output is at maximum without clipping...? 6. My DynAudios are set to "flat" response, and 0 db 7. The whole studio including control room is John's inside outside design all sealed, except the door is not finished. Control room has John's soffit mounted speaker design with bass hangers underneath and all corners have bass hangers. Mike ps Again sorry for removing graphs....you are right and I did not realise the importance and so this will not happen again.
2. Yes I was in the room however I swear I never made a sound.
Sometimes just being in the room can make a difference, but not usually as big as is evident here. But using the "start delay" should get rid of that. Set it to a delay long enough for you to be able to get out of the room and close the door, plus another few seconds for things to calm down inside. Also set "Sweeps" to "2". That means that REW will run two sweeps for each test, then take the average. That helps to improve the signal-to-noise ratio as well.
please go to ... for all to see my control room and tracking
Mackie 32.8 with meter bridge! :thu: I love that console. I use a 24.8 regularly. Nice! Please also post a few up-to-date photos here on the forum, for the same reason as above: in the future, your web site might be at a different location, so the link you posted might not work, thus making it inaccessible to others in the future who might want to follow your thread. Also, I'd like to see photos of the sides, rear, and ceiling of your room, so I can get a better idea of how it is working right now.
3. I did have my SPL meter set to "C" weighting but "Fast" so I will change that REW was "C" weighted and "Slow"
REW is set to "C" and "Slow" by default, but it's worth checking, as you might have changed it by accident. But are you SURE you calibrated REW to your meter: You really did tell REW what the real level is when you did the calibration, right? You typed that into the box and hit "enter", then checked that the level on your meter matched what the REW calibrator window was telling you? If you don't tell REW what the real measured level was in the room, then it has no way of knowing...
4. When I tool the measurements of both speakers the measurement was higher than when I did Left and Right speakers.
Correct. The level should increase by 6 dB, from 80 for each individual speaker to 86 for both at once. That's why studios are calibrated at 86 dB... REW uses coherent sine sweeps for the test signal, so it sums to exactly twice the intensity, thus giving you an increase of exactly 6 dB.
5. Also, when the first measurement was taken I only had -1.3db headroom left. Previously to that when I took the measurements REW said there was clipping
.:: which makes me suspect that you have too much gain at some point in your signal chain, and not enough at other points. Are you going through your Mackie console? If so, make sure that the output level from your DAW really is at -12 dB on the channel meter on the console when the speakers are producing 80 dB each, and that the meter on the channel you are using for your ECM8000 peaks no higher than 0 dB when you are playing the calibration test signal (pink noise). Then set your sound card interface accordingly. If you set up your gain structure correctly, there should be plenty of head-room: at least 10 dB. Make sure you have the EQ button for those channels turned off on the channel strips, and that the output channels are panned hard left and hard right, while the mic input channel is panned dead center.
so what do you suggest as my system in terms of mic input volume and speaker output is at maximum without clipping...?
The goal is to have the entire signal chain calibrated to 0 dB when the test signal is playing, and adjust the speaker volume controls to get 80 dB(C) SPL from each speaker (which will automatically be 86 dB(C) SPL) with both speakers playing). The second goal is to have your ECM8000 input channel gain set such that the meter is showing 0 dB when you have both speakers producing 86 dB(C) SPL, and your interface gain controls set the same. If everything is showing 0 dB while the speakers are playing individually, then you will not have clipping.
6. My DynAudios are set to "flat" response, and 0 db
And you are not using a sub, correct?
7. The whole studio including control room is John's inside outside design
Ahhh! OK, that explains it. That's fine then. With that approach, the idea is to start with the dead space then add treatment to liven it up selectively. That's the opposite of what you do for conventional construction: starting with a very live reverberant space, then adding to treatment to deaden it. I prefer John's approach, and I do that whenever I can. So that's fine: At least I now understand where you are coming from, adn the reason for the deadness.
Control room has John's soffit mounted speaker design with bass hangers underneath and all corners have bass hangers.
Great! What I was thinking was a lack of bass trapping is actually a lack of mids- and highs-. The room is bigger than it seems from the graphs, and since I wasn't aware that it was built inside-out, I was making the wrong assumptions. What we need to do is to liven up the mids and highs. Highs are easy, mids not so much... The thing that is strange about your data is the lack of bass. I don't get why there's that major roll-off in bass levels below 120 Hz. Your BM6A's are capable of better than that, which is why I was wondering if you maybe had the controls set wrong. What else is there in the signal path that could be rolling off the low end? Do you have any equalization in the signal path? If so, set it flat. What I'm seeing on the graphs is not what I would expect from DynAudio BM6s. Those are great speakers, and they don't roll off at 120 Hz., so it must be something else in the signal path that is killing the low end. Or it might be some artifact of the way the room is set up. Please provide an accurate diagram of the room, especially the positions of the speakers, the mix position, and the major treatment, and showing all the dimensions. As well as the photos! - Stuart -
Hi Stuart The dimensions of the control room is length 4.550 metres, width 3.4m, height is 2.8m at highest point. The distance between speakers is 1.7meters distance from floor to centre of speakers is 1.250metres If you need any more measurements please let me know. There are bass hangers in all four corners and being johns inside/outside design, the wall studs are 70mm deep filled with acoustic insulation from memory around 42kg then the beige and red hessian covering I have checked the dyn audios again and both are set to High Pass Filter Flat db Level = 0 Eq LF, MF, HF =0 Here are the pictures 1. The rear of control room 2. The sides 3. The front 4. The ceiling (could not upload rear ceiling max 12 allowed, however rear has the cloud between air-con inlet and outlets same design as the front.) Also in terms of signal path, I will make sure all eq on Mackie channel is flat Mike :)
Image not preserved: Left Corner.jpg
Image not preserved: Right Corner.jpg
Image not preserved: Control Top Front.jpg
Image not preserved: Control Room Front .jpg
Image not preserved: Control Left Side .jpg
Image not preserved: Left Top.jpg
Image not preserved: Control Right Side.jpeg
Image not preserved: Top Right Side .jpg
Image not preserved: Top Left Rear.jpg
Image not preserved: Top Rear Side.jpg
Image not preserved: Control room rear.jpg
Image not preserved: Front ceiling.JPG
Hi Stuart Re measured room with book case, keyboard and all removed. 1. Set REW to 10 sec delay then left room and shut the door 2. Made sure my SPL meter held next to ECM8000 read 80-81 db and reflected REW (see screen shot) 3. Measurements in drop box link below https://www.dropbox.com/s/iey3j7c5hurhw ... .mdat?dl=0 Mike
Image not preserved: MCS Screenshot.jpg
Hi Stuart
"Those are great speakers, and they don't roll off at 120 Hz., so it must be something else in the signal path that is killing the low end."
Sorry, yes your right. I forgot to take the low cut off on the mic input, so now that's done, and the low end looks a little better, here are the latest measurements in drop box. https://www.dropbox.com/s/uaaxqh1w33f2p ... .mdat?dl=0 1. I am not using a sub 2. Did not go through my Mackie Console. 3. No eq in signal path everything flat, low cut "off" 4. Calibrated REW to my SPL meter I noticed that the bass roll off is now around 40hz but I still have some strange stuff going on from 2 to 7k even though I left the room and set REW to a 10 sec delay What do you think this maybe ? Mike 8)
I forgot to take the low cut off on the mic input,
:shock: :roll: :cop: :blah: 8) :!: Well, at least that mystery is cleared up! That makes a lot more sense now. And the frequency response is looking rather good, actually. But there's still a problem with the way you are doing the measurements: you are re-calibrating between tests! Never. You can't do that, as it makes the results invalid. The reason I know you are doing this is because the test for both speakers at once is at the same level as the tests for the individual speakers by themselves, which is impossible. That's like saying you had two 1 liter bottles of coke in front of you, both full, but when you poured them both into a bowl, the result was exactly one liter of coke.... And you are also doing your tests at 75 dB. It needs to be 80. So, once again: you need to calibrate your left speaker so that it produces 80 dBC all by itself, with the right speaker turned off completely, then you never, ever touch that setting ever again. Make certain you know exactly where you have all of you controls set to produce that level, and never change them. Ever. The do the same for the right speaker by itself with the left one turned off: it must also produce exactly 80 dB. Then you run REW once for just the left speaker, once for just the right speaker, and once with both speakers together. Do NOT change any settings at all between tests, and use those exact settings for all future tests.
4. Calibrated REW to my SPL meter
Then your meter is bad, or your interface is bad, ... or you changed the levels after calibrating.... :shock: :!: If you had calibrated correctly and then never changed anything, REW would show the individual graphs at 80 dB, not 75, and the graphs for both together would automatically end up at 86 dB.
but I still have some strange stuff going on from 2 to 7k even though I left the room and set REW to a 10 sec delay. What do you think this maybe ?
If that noise is coming out of the speakers, you would hear it, even when REW is not doing anything. It would be a loud hiss that dies away and drops down the scale to low rumble. That's what the graphs are showing. You would certainly hear it. Is that true? Do you hear that? - Stuart -
Here is the latest REW in dropbox, the test shows 80.6db
Nope! It's about 74 dB. Here:
Image not preserved: ratty-too-low.jpg
The horizontal line marks 74 dB. There is nothing at all even approaching 80 dB, and certainly not exceeding it. There is something wrong with the way you are calibrating REW. And once again, you adjusted your settings between doing the readings... There is no difference at all in levels between the tests for each speaker individually, and both of them together. Therefor, you changed the settings,thus invalidating the results. Third, the sizzling "mush" in the high mids and highs is still there:
Image not preserved: ratty-mid-mush.jpg
That is LOUD! You MUST be able to hear it. Either that, or there's something badly wrong with your gear, or something else in the room that is resonating. It is very weird: It drops down the scale in a linear run, but does not decay over time, as yuou can clearly see in this sepctrogram:
Image not preserved: ratty-sp-mush.jpg
You need to fix that. Try with a different computer, interface, and mic. From what I can see, the overall room response is not too bad, but I'm not at all confident that the readings are valid, due to all of the reasons above. - Stuart -
Hi Stuart Very disappointing result as I used an 1. 2010 apple mac with (OS Mountain Lion) 2. REW 5.2 (mac) 3. ECM8000 which is all I have in terms of reference mic (with calibration file) 4. Analogue decibel meter 5. Zoom Recorder interface (for input into apple mac) output via Mackie 8 buss analogue desk to Dynaudio BM6 MK2 I have tried to calibrate according to all the REW instructions and your posts. I set my apple mac sound inputs and outputs to the max and input on interface as far as possible (without distortion). please note that if I set the input level on my interface as high as possible I get a message from REW that the levels are to high and distortion will corrupt the readings. So to get the level of 80db, I will have to overlook the warnings from REW. Stuart I can promise you I did not change any settings while doing the readings of each speaker left then right and both as you made this perfectly clear on a previous post. So I am not sure what is going on. "That is LOUD! You MUST be able to hear it." Sorry, my ears heard nothing, however, I will do the tests again and really listen for the "mush" However, I concede that the readings are still not correct, so I will try my windows laptop (OS windows 10) and see what else I can use and see if the changes makes any difference. Any other suggestion greatly appreciated...!!! Mike
Hi Stuart Very disappointing result as I used an 1. 2010 apple mac with (OS Mountain Lion) 2. REW 5.2 (mac) 3. ECM8000 which is all I have in terms of reference mic (with calibration file) 4. Analogue decibel meter 5. Zoom Recorder interface (for input into apple mac) output via Mackie 8 buss analogue desk to Dynaudio BM6 MK2 I have tried to calibrate according to all the REW instructions and your posts. I set my apple mac sound inputs and outputs to the max and input on interface as far as possible (without distortion). please note that if I set the input level on my interface as high as possible I get a message from REW that the levels are to high and distortion will corrupt the readings. So to get the level of 80db, I will have to overlook the warnings from REW. I did NOT change any settings while doing the tests of each speaker left and right and then both as you made this perfectly clear on a previous post.
That is LOUD! You MUST be able to hear it.
Sorry, my ears heard nothing so I will do the tests again and really listen for the "mush" However, I concede that the readings are still not correct, so I will try my windows laptop (OS windows 10) and see what else I can use and see if the changes makes any difference. Any other suggestion greatly appreciated...!!! Mike
Very disappointing result
I'm wondering if there is something wrong with the software at some point: maybe a driver conflict, or a latency issue, or something. That "mush" might not actually be there if you can't hear it. It might be some type of artifact from software stuff... You would definitely hear it if it was real. But if it is an artifact, then that brings into question all of the data. so switching to another machine is the smart thing to do.
I did NOT change any settings while doing the tests of each speaker left and right and then both as you made this perfectly clear on a previous post.
That's very strange. Think of it this way: If the left speaker by itself is putting out 80 dB and the right speaker by itself is also putting out 80 dB, then it is impossible that the combined level of both is also 80 dB. Mathematically, it HAS to be 86 dB. It MUST be double. There's no alternative. That would be like having two bottles of water, each with exactly one liter in each bottle, but when you pour them both into a bucket together, the result is that you still have only one liter, total.... It just does not make sense. Therefore, the "both speakers" level MUST be 6 dB higher than the left and right channels. Are you SURE that other speaker is being turned off completely? When you do the test for the left speaker, are you physically turning off the actual power switch on the right speaker, and the light goes off, with no sound at all coming out of it? Are you SURE that there are no routing issues with your interface? No dynamics plug-in enabled that is increasing the level?
please note that if I set the input level on my interface as high as possible I get a message from REW that the levels are to high and distortion will corrupt the readings. So to get the level of 80db, I will have to overlook the warnings from REW.
Then there's definitely a problem with the setup somewhere, 86 dB should not be too high for your system to handle. If that were the case, you'd never be able to record anything! Many, many instruments put out way more than that. Are you sure that the input gain on the interface is set correctly? If the speakers are putting out 86 dB, then there's no reason why that should overload the interface, or the input to REW.
I will try my windows laptop (OS windows 10) and see what else I can use and see if the changes makes any difference.
That's an excellent idea. That will help to narrow down the possibilities. - Stuart -
Hi Stuart Thanks mate for your input and patience. I will try all alternatives to hopefully get all the problems sorted...!!! thanks again Mike :)