REW cntrl room test plus video tour. Advice?

Hi Everyone! I recently "completed" construction of a new studio and i am getting into the control room adjustments. I've made other rooms before and always only used my ears, never really REW before but i thought it would give me a better idea of what i am hearing. The first time i set up the speakers i was ofc coure.... super nervous.. afraid of getting hit in the face by uncontrollable low end inconsistencies... To my surprise the low end of things seemed to be pretty even compared to most other rooms that i've been in in the past. To my ears i could hear some strange 3.5 k build up that made vocals and other instruments more aggressive ( in a bad way ) . At the time i had attributed that to the missing cloud over the mix position, windows partially covered behind the speakers and the control room window to my right. I have attached a video link to a quick tour and explanation of how the room was build, the dimensions are in the video description as well. I do see dips and peaks in the lower end but honestly I did not hear them as much as the graph shows, my speakers 8" woofers are not bass heavy at all but i went through a lot of bass heavy material during the listening and never felt much uneven low end tones. On the higher end of things i do see some comb filtering that might be due to the mixing desk, just an opinion.. i am open to suggestions! I think the 3.5 k harshness is probably due to the missing cloud and the missing heavy drapes behind the speakers. I have attached screenshots of the REW tests. I would be very happy if anyone with more experience in testing and investigating problems could chime in. Thank you so much! here is the video link https://www.youtube.com/watch?v=4kGlmAPlCHc Ale

Forgot to apply smoothing to the higher freq screenshot!

I would need the actual MDAT file to be able to fully analyze the data, and I'd also need to know exactly how you did that test (location of the mic, speakers, etc.), as well as the dimensions of the room, and the position the acoustically important things in it, such as the desk, sofa, etc., and also details of the treatment in the room. Photos would help too!

never really REW before but i thought it would give me a better idea of what i am hearing.

REW will do a lot more than that! It will enable you to tune your room correctly, precisely, so it meets the specifications for control rooms, such as ITU BS.1116-2 and others. Here's an example of what how a well tuned room should look, in REW graphs: viewtopic.php?f=2&t=20471

To my surprise the low end of things seemed to be pretty even

It's not too bad, as far as I can see, but without knowing all the other details, it's hard to say. There's clearly a large modal problem at about 48 Hz, which is right in the range of bass guitars, large kick drums, the low end of keyboards, and even come electric guitars. I'm surprised you can't hear that. There are also dips at around 80 and 120 Hz, which might be SBIR, or might be something else (floor bounce, maybe, or ceiling bounce). The rest of the low end looks reasonable, but it does seem rather dry and dead. The decay times seem to be rather fast, although once again, it's hard to say for sure without fully analyzing the MDAt file. That's the low end. But the high end is a huge mess! You have very major swings of over 20 dB in the high end! That's very strange, and very hard to explain....

i could hear some strange 3.5 k build up

I'm not seeing that at all in the data. In fact, there's a clear dip between 2 k and 6 k, bottoming out around 3.5 k! I would suspect a phasing issue of some type, reflections, flutter echo, or something that just isn't evident in the limted set of basic graphs that you posted.

At the time i had attributed that to the missing cloud over the mix position, windows partially covered behind the speakers and the control room window to my right.

Possibly, but I'm not convinced. There are many other things that could be causing this, and the ones you mentioned would not be very high on my list of culprits! :)

I have attached a video link to a quick tour and explanation of how the room was build, the dimensions are in the video description as well

I'll take a look at that one day when I have time, but in the meantime it would be better to see an actual diagram of the room layout, a few photos, and at least the basic dimensions.

I do see dips and peaks in the lower end but honestly I did not hear them as much as the graph shows, my speakers 8" woofers are not bass heavy at all but i went through a lot of bass heavy material during the listening and never felt much uneven low end tones.

I'm surprised about that, because the graphs are showing clear issues that should be audible. Even simple sine sweeps should highlight those issues clearly.

On the higher end of things i do see some comb filtering that might be due to the mixing desk,

In a properly designed and laid out room, artifacts from the desk and console generally show up in the low end of the mid range, not in the highs. That would be very unusual. My number one suspect for your problems would be the way the room is laid out, in terms of room geometry: position of the speakers with respect to the walls, floor, ceiling, desk, mix position. My second suspect would be incorrect speaker mounting. Third would be incorrect treatment in the front end of the room, or objects in the front of the room. My fourth suspect would be the speakers themselves, or the signal chain: What speakers are they (make and model)? How do you have the controls set? What else is in the signal path? Did you perform the entire REW calibration procedure?

I think the 3.5 k harshness is probably due to the missing cloud

I doubt it! There's not much energy that far off axis at 3.5k with most speakers. At 3.5k, the Q is pretty much a beam, with not a lot of energy at high angles off-axis, where the cloud usually goes.

and the missing heavy drapes behind the speakers.

I'm not sure I understand: Why would you want drapes BEHIND the speakers? On the rear side of the speakers, most of the energy (and all of the problems!) are in the low end, not the high end. Drapes can help a bit for high frequency treatment, but do nothing at all for lows. SBIR coming off the front wall, for example, would not even notice drapes, no matter how heavy they are. Drapes would be no use at all on the front wall. What other treatment do you have on the front wall? Please provide the MDAT file, photos, diagrams, and dimensions. - Stuart -

Soundman2020 wrote:

I would need the actual MDAT file to be able to fully analyze the data, and I'd also need to know exactly how you did that test (location of the mic, speakers, etc.), as well as the dimensions of the room, and the position the acoustically important things in it, such as the desk, sofa, etc., and also details of the treatment in the room. Photos would help too!

never really REW before but i thought it would give me a better idea of what i am hearing.

REW will do a lot more than that! It will enable you to tune your room correctly, precisely, so it meets the specifications for control rooms, such as ITU BS.1116-2 and others. Here's an example of what how a well tuned room should look, in REW graphs: http://www.johnlsayers.com/phpBB2/viewt ... =2&t=20471

To my surprise the low end of things seemed to be pretty even

It's not too bad, as far as I can see, but without knowing all the other details, it's hard to say. There's clearly a large modal problem at about 48 Hz, which is right in the range of bass guitars, large kick drums, the low end of keyboards, and even come electric guitars. I'm surprised you can't hear that. There are also dips at around 80 and 120 Hz, which might be SBIR, or might be something else (floor bounce, maybe, or ceiling bounce). The rest of the low end looks reasonable, but it does seem rather dry and dead. The decay times seem to be rather fast, although once again, it's hard to say for sure without fully analyzing the MDAt file. That's the low end. But the high end is a huge mess! You have very major swings of over 20 dB in the high end! That's very strange, and very hard to explain....

i could hear some strange 3.5 k build up

I'm not seeing that at all in the data. In fact, there's a clear dip between 2 k and 6 k, bottoming out around 3.5 k! I would suspect a phasing issue of some type, reflections, flutter echo, or something that just isn't evident in the limted set of basic graphs that you posted.

At the time i had attributed that to the missing cloud over the mix position, windows partially covered behind the speakers and the control room window to my right.

Possibly, but I'm not convinced. There are many other things that could be causing this, and the ones you mentioned would not be very high on my list of culprits! :)

I have attached a video link to a quick tour and explanation of how the room was build, the dimensions are in the video description as well

I'll take a look at that one day when I have time, but in the meantime it would be better to see an actual diagram of the room layout, a few photos, and at least the basic dimensions.

I do see dips and peaks in the lower end but honestly I did not hear them as much as the graph shows, my speakers 8" woofers are not bass heavy at all but i went through a lot of bass heavy material during the listening and never felt much uneven low end tones.

I'm surprised about that, because the graphs are showing clear issues that should be audible. Even simple sine sweeps should highlight those issues clearly.

On the higher end of things i do see some comb filtering that might be due to the mixing desk,

In a properly designed and laid out room, artifacts from the desk and console generally show up in the low end of the mid range, not in the highs. That would be very unusual. My number one suspect for your problems would be the way the room is laid out, in terms of room geometry: position of the speakers with respect to the walls, floor, ceiling, desk, mix position. My second suspect would be incorrect speaker mounting. Third would be incorrect treatment in the front end of the room, or objects in the front of the room. My fourth suspect would be the speakers themselves, or the signal chain: What speakers are they (make and model)? How do you have the controls set? What else is in the signal path? Did you perform the entire REW calibration procedure?

I think the 3.5 k harshness is probably due to the missing cloud

I doubt it! There's not much energy that far off axis at 3.5k with most speakers. At 3.5k, the Q is pretty much a beam, with not a lot of energy at high angles off-axis, where the cloud usually goes.

and the missing heavy drapes behind the speakers.

I'm not sure I understand: Why would you want drapes BEHIND the speakers? On the rear side of the speakers, most of the energy (and all of the problems!) are in the low end, not the high end. Drapes can help a bit for high frequency treatment, but do nothing at all for lows. SBIR coming off the front wall, for example, would not even notice drapes, no matter how heavy they are. Drapes would be no use at all on the front wall. What other treatment do you have on the front wall? Please provide the MDAT file, photos, diagrams, and dimensions. - Stuart -

Hi Stuart! As usual, very careful and detailed... thank you so much! I've been tryint to upload the file but somehow i get an error, trying again now. Also uploading the rt60 screenshot. After the test i went back and listened using an oscillator and indeed confirmed the peaks and nulls. They are so much more noticable with a generator... It is also weird to me that i didn't here any masking in the low end, went back to listen to stuff and felt again like i had clarity in the low end... ( unbalanced clarity but clarity). So i discovered the rt60 graph and checked... I think the reason why i thought the low end was clean was the rt60 response, the low end seems to only raise by 0.4 after 100 hrz .. so i guess i was hearing a clean "upper low end". The speakers I am using now are Mackies MR8 the very first model https://www.recordingmag.com/productrev ... 03/25.html Not great speakers but i know them pretty well. I can't attach the MDAT it says: the extention MDAT is not allowed... I've measured from the listening position, where my head would be, the speakers are lower than they should be and will soon invest into supports... The high frequency mess is most likely the bare concrete ceiling ( with beams ) Floor, and the last 2 feet of the side walls being bare as of now. I can actually hear some slight flutter echo up there. I've explained treatment and dimensions in the video but i can summarize here. Dimensions 142 1/2 inches wide 145 " tall 164" is the long wall where the speakers are firing, big BUT! the back wall is totally irregular with a huge bass trap acting as a false wall, a small pantry behind it and a bathroom further down so in reality the long wall is much longer than that at different points... the 2 front corners are 2 superchunk bass traps with pink fluffy 12 tall and around 26 inches wide 10 of the 12 feet in height of the side walls are made by 3" roxul rockboard. Back wall has 2 bass traps columns approx 10ft tall 5 inches deep and over a foot wide. on top of those columns there is another bass trap on top approx 8 ft long . Its very hard to explain but basically part of the back wall is an irregular giant bass traps with 703 on the wider surfaces and pink fully in columns and wall to ceiling corners. I will take pictures tomorrow and attach them here. I also do think it could be sbir and need to play with placement ( speakers are ported in the back ) and right now i started with the speaker as close as possible to the back wall and eventually work my way back and re test. Also having width and height of the room very similar it might be more of a modal issue? I do have space on the upper side wall to ceiling and could tra more if needed, although trapping 40hz.. unless i use pressure traps and get the q in the right spot... As you can see i am a bit all over the place right now... I will update with pics, and may be i can send you an email with the mdat file? Thank you so much for you time! Alessio

Also Steward. Just saw the linked thread. What a fantastic job!!!!! I don't think i've ever seen a flatter low end response in a room!!!

Just saw the linked thread. What a fantastic job!!!!! I don't think i've ever seen a flatter low end response in a room!!!

Thanks! It wasn't easy to do, but it can be done, with careful design, careful testing, and careful tuning.

142 1/2 inches wide 145 " tall 164" is the long wall where

Well that explains the low end! :) The modal issue at around 48 is actually two at once: 0.0.1 axial and 0.1.0 axial, at pretty much the same frequency (46.8 and 47.6), because the profile is square! The cloud will help a little with that, provided that it is huge, hard-backed, and angled. The one "around 80" is your 2.0.0 axial at 82.7 Hz., and the one "around 120" is your 3.0.0 axial, at 124 Hz. In other words, you don't have enough bass trapping in the length and height axes, and width is also deficient.

Its very hard to explain but basically part of the back wall is an irregular giant bass traps

That's fine, but there still isn't enough in the low end.

I also do think it could be sbir and need to play with placement ( speakers are ported in the back ) and right now i started with the speaker as close as possible to the back wall and eventually work my way back and re test.

There's no need to play: Just set the up correctly! :)

Also having width and height of the room very similar it might be more of a modal issue?

Yep!

although trapping 40hz.. unless i use pressure traps and get the q in the right spot...

You'd be surprised... the wavelength at 40 Hz is 28.25 feet, so theoretically it is feasible to get usable trapping from 23" of porous absorber for normal-incident sound, and as little as 12" for randomly incident sound. Most people don't realize that porous absorption is effective down to rather low frequencies.

I can't attach the MDAT it says: the extention MDAT is not allowed...

Upload it to a file sharing service, such as Dropbox, then post the link here.

As you can see i am a bit all over the place right now...

First things first: A lot of your issues arecoming from the incorrect speaker placement. the very first thing you need to do is fix that. Your speakers need to go right up tight against the front wall, except for a gap of 4" where you will insert a panel of 4" 703, directly in between the rear corner of the speaker and the front wall. They also need to go on stands... Heavy! Massive! Lots of weight! Some people stack up concrete blocks or bricks, others make it from hollow steel profile and fill it with sand. But whatever you do, it has to be HEAVY. It also has to be built to the correct heights, such that the acoustic axis of the speakers ends up at 1.2m above the floor, which is 47-1/4". Or maybe a bit higher. Perhaps as much as 49", but not higher. You also need turn your speakers the right way up! You have them laying on their sides, which is terrible, and is probably part of the problem with the frequency response in the high end. And you also need to get the in the correct location with respect to the side walls: each speaker must go 39" from the side wall, in your room. Which means that they will be 64.5" apart. Once again, I'm talking about the acoustic axis of the speakers, NOT the top, bottom, or sides of the cabinet. Check the manual to find out where the acoustic axis is on your speakers. OK, once you have your speakers set up correctly, you'll need to set up your listening position. For that room, you need to set up your chair such that your ears will be 61" from the front wall when you are mixing normally. Then set up your desk comfortably in front of that. Now set up a vertical pole of some type (eg, mic stand) about 16" behind your head (in other words, 76" from the front wall), and carefully aim each speaker such that it points directly at the pole. Here too, I'm talking about the acoustic axis: The acoustic axis must point at the pole. Double check all those measurements. That's the theoretical ideal geometry. Now calibrate your gain structure, and REW, such that each speaker is producing exactly 80 dBC with REW putting out a -10dB signal, and everything in the signal chain pegged at either 0 dB (analog gear) or -20 dBFS (digital gear). You will need a hand-held sound level meter to do this. If you do that correctly, when each speaker produces 80 dBC by itself, both of them together will automatically produce 86 dBC. With that situation,m calibrate REW to tell it that the level is 86 dBC when the meter is showing 86 dBC. Use the full-spectrum test signal in REW, NOT the sub-woofer test signal! With that setup, run three REW tests: one for the left speaker alone (right one turned off), one with just the right speaker (left one turned off), and one with both. Post the MDAT file with those three tests, with the room set up optimally. - Stuart -

Soundman2020 wrote:

Just saw the linked thread. What a fantastic job!!!!! I don't think i've ever seen a flatter low end response in a room!!!

Thanks! It wasn't easy to do, but it can be done, with careful design, careful testing, and careful tuning.

142 1/2 inches wide 145 " tall 164" is the long wall where

Well that explains the low end! :) The modal issue at around 48 is actually two at once: 0.0.1 axial and 0.1.0 axial, at pretty much the same frequency (46.8 and 47.6), because the profile is square! The cloud will help a little with that, provided that it is huge, hard-backed, and angled. The one "around 80" is your 2.0.0 axial at 82.7 Hz., and the one "around 120" is your 3.0.0 axial, at 124 Hz. In other words, you don't have enough bass trapping in the length and height axes, and width is also deficient.

Its very hard to explain but basically part of the back wall is an irregular giant bass traps

That's fine, but there still isn't enough in the low end.

I also do think it could be sbir and need to play with placement ( speakers are ported in the back ) and right now i started with the speaker as close as possible to the back wall and eventually work my way back and re test.

There's no need to play: Just set the up correctly! :)

Also having width and height of the room very similar it might be more of a modal issue?

Yep!

although trapping 40hz.. unless i use pressure traps and get the q in the right spot...

You'd be surprised... the wavelength at 40 Hz is 28.25 feet, so theoretically it is feasible to get usable trapping from 23" of porous absorber for normal-incident sound, and as little as 12" for randomly incident sound. Most people don't realize that porous absorption is effective down to rather low frequencies.

I can't attach the MDAT it says: the extention MDAT is not allowed...

Upload it to a file sharing service, such as Dropbox, then post the link here.

As you can see i am a bit all over the place right now...

First things first: A lot of your issues arecoming from the incorrect speaker placement. the very first thing you need to do is fix that. Your speakers need to go right up tight against the front wall, except for a gap of 4" where you will insert a panel of 4" 703, directly in between the rear corner of the speaker and the front wall. They also need to go on stands... Heavy! Massive! Lots of weight! Some people stack up concrete blocks or bricks, others make it from hollow steel profile and fill it with sand. But whatever you do, it has to be HEAVY. It also has to be built to the correct heights, such that the acoustic axis of the speakers ends up at 1.2m above the floor, which is 47-1/4". Or maybe a bit higher. Perhaps as much as 49", but not higher. You also need turn your speakers the right way up! You have them laying on their sides, which is terrible, and is probably part of the problem with the frequency response in the high end. And you also need to get the in the correct location with respect to the side walls: each speaker must go 39" from the side wall, in your room. Which means that they will be 64.5" apart. Once again, I'm talking about the acoustic axis of the speakers, NOT the top, bottom, or sides of the cabinet. Check the manual to find out where the acoustic axis is on your speakers. OK, once you have your speakers set up correctly, you'll need to set up your listening position. For that room, you need to set up your chair such that your ears will be 61" from the front wall when you are mixing normally. Then set up your desk comfortably in front of that. Now set up a vertical pole of some type (eg, mic stand) about 16" behind your head (in other words, 76" from the front wall), and carefully aim each speaker such that it points directly at the pole. Here too, I'm talking about the acoustic axis: The acoustic axis must point at the pole. Double check all those measurements. That's the theoretical ideal geometry. Now calibrate your gain structure, and REW, such that each speaker is producing exactly 80 dBC with REW putting out a -10dB signal, and everything in the signal chain pegged at either 0 dB (analog gear) or -20 dBFS (digital gear). You will need a hand-held sound level meter to do this. If you do that correctly, when each speaker produces 80 dBC by itself, both of them together will automatically produce 86 dBC. With that situation,m calibrate REW to tell it that the level is 86 dBC when the meter is showing 86 dBC. Use the full-spectrum test signal in REW, NOT the sub-woofer test signal! With that setup, run three REW tests: one for the left speaker alone (right one turned off), one with just the right speaker (left one turned off), and one with both. Post the MDAT file with those three tests, with the room set up optimally. - Stuart -

Thanks Again Steward! Great guidelines for setting things up right ( including the speakers...) I assume you've seen the video, and those speaker stands are attached to the desk.. so i guess i would have to decouple the speakers from the supports and decouple the desk from the floor? Also i've always mixed with horizontal speaker placement but I will get them up and re test! i am not able to find a diagram with the acoustic axis speakers.... any other way to find that out? I am in the process of building the could, I was using 6 wooden grids (space couplers) with 4 inch 703 stacked above them, so there would be no hard back on the cloud. Could i close the cloud with anything like a 1/4 inch board what would that achieve? Also i have downlights planned to be inside the cloud... Since the ceiling is 12" tall i guess i could mount it at the 1/4 wavelenght to catch that mode? I am a little confused, when you say we can dissipate modes down to 40hz with porus absorbers. Wouldnt they have to be enormous? Unless using a perforated panel in front? Again, thanks for all the suggestions. I will set things up as suggested ASAP then re test and send a new .mdat file! Now, just for kicks... i attach a link to the first test, just so people can start yelling at me on the settings and then i eventually learn to measure correctly! https://studio42brooklyn.wetransfer.com ... 613/0306b6 Appreciation! Ale

I assume you've seen the video, and those speaker stands are attached to the desk.. so i guess i would have to decouple the speakers from the supports and decouple the desk from the floor?

You need entirely separate speaker stands, not touching anything. Not the desk, or the wall, or anything else. They need to be completely stand-alone. Remove everything from the desk that is not absolutely necessary: speaker stands, shelves, equipment, cables, video screens, etc. Just leave the absolute minimum equipment that you need. It's not necessary to decouple the desk from the floor, but it would be good t decouple the speaker from the stand, using Sorbothane pads. That's the absolute bests stuff out there.

Also i've always mixed with horizontal speaker placement

Here's why that's a bad idea: That's very probably part of the problem you are seeing in the mid/high. It peaks at the cross-over frequency for your speakers, and rolls off about an octave either side.

i am not able to find a diagram with the acoustic axis speakers.... any other way to find that out?

It will always be on the imaginary line that joins the middle of the woofer to the middle of the tweeter,and will be much closer to the tweeter. For your Mackie's, I would estimate that it would be about here: Somewhere around 130mm from the top edge. I know that's not the exact same model you have, but the axis wont be far from there. It's not ultra-critical to have it perfect either. That would be a good approximation.

I was using 6 wooden grids (space couplers) with 4 inch 703 stacked above them, so there would be no hard back on the cloud.

That's not really a cloud! Take a look at that link I gave you yesterday: The cloud there is huge. It's the maroon monster across the entire front end of the room. There are several modules, some hard-backed, some not, and hung at different angles. You will need something similar. Your ceiling is high, so you have plenty of space to do that, and you'll be able to angle them well. You also have modal issues in the vertical plane: space-couplers won't to much to help that! You don't want the space above and below the cloud acoustically coupled: You want the opposite: Acoustically UN-coupled. Two different acoustic spaces with a large impedance mismatch in between.

Could i close the cloud with anything like a 1/4 inch board what would that achieve?

Much thicker. I would recommend 3/4" MDF. You need mass on that thing, if you hope to have much effect on modes.

Also i have downlights planned to be inside the cloud...

:thu: No problem! Take a look at the cloud in the link... :)

Since the ceiling is 12" tall i guess i could mount it at the 1/4 wavelenght to catch that mode?

Not necessary, and not all that useful.... and kind of hard to do! The quarter wave for your 47 Hz mode is 6 feet, so you'd have to hang it half way between the ceiling and the floor... it would sort of be in the way there, every time you stood up from the desk! :)

I am a little confused, when you say we can dissipate modes down to 40hz with porus absorbers.

A porous absorber that is as thick as 7% of the wavelength is sufficient to have a useful effect on normal-incident waves, and 3.5% thickness is good for random incidence. Not highly effective, of course, but still usefully so. The thicker you make it, the better it gets, and the further you move it away from the wall, the better it gets. The perfectly optimum distance is, indeed, a quarter wavelength away from the wall, but even at much smaller distances it still has some effect. The only place that it has no effect at all is right up against the wall.

Wouldnt they have to be enormous? Unless using a perforated panel in front?

If it has a perf panel in front, then it isn't a bass trap! It's a tuned trap, which might or might not be good for bass, depending on the tuning. Porous absorbers work by taking away some of the acoustic energy of any sound wave that passes through them, and converting it into low-grade heat. Sound waves are basically just vibrating air molecules. They transfer energy to the fibers of the porous absorber by striking them, causing the to flex, which absorbs the kinetic energy of the wave and converts it to thermal energy. As long as the wave pass through a reasonable thickness of the absorber, there's a reasonable chance that some of the air molecules will strike some of the fibers. The greater the thickness of the panels, the better the chances, but even at 7% there's still a statistically decent chance that many molecules will strike many fibers. Also, and maybe this is splitting hairs, but you can't "dissipate" a mode: All you can do is "damp" it. The mode is still there, but it just doesn't "ring" any more, because it is heavily damped. - Stuart -

Thank Steward! I even learned how to quote!

You need entirely separate speaker stands, not touching anything. Not the desk, or the wall, or anything else. They need to be completely stand-alone.Remove everything from the desk that is not absolutely necessary: speaker stands, shelves, equipment, cables, video screens, etc. Just leave the absolute minimum equipment that you need.

I did most of the clean up, I will probably change stands as soon as some finances start coming back in... If it is a completely stupid idea then.. I will do my best to improvise!

That's very probably part of the problem you are seeing in the mid/high. It peaks at the cross-over frequency for your speakers, and rolls off about an octave either side.

Placed speakers upright and took all the measurements, the weird 3.5 k madness is gone to my ear!!

Somewhere around 130mm from the top edge. I know that's not the exact same model you have, but the axis wont be far from there. It's not ultra-critical to have it perfect either. That would be a good approximation.

Yes its what i thought so i took measurements from there

That's not really a cloud! Take a look at that link I gave you yesterday: The cloud there is huge. It's the maroon monster across the entire front end of the room. There are several modules, some hard-backed, some not, and hung at different angles. You will need something similar. Your ceiling is high, so you have plenty of space to do that, and you'll be able to angle them well. You also have modal issues in the vertical plane: space-couplers won't to much to help that! You don't want the space above and below the cloud acoustically coupled: You want the opposite: Acoustically UN-coupled. Two different acoustic spaces with a large impedance mismatch in between.

It makes sense, but if I am not mistaken it'll be super heavy right? I do not trust the concrete on the ceiling it was poured poorly and its thicn in some places and thick in others... May be i can run 4x4's across from sidewall to sidewall and use those as support? Another option/question: What if I use the 7% rule and use 24 inches of porus absorption on top of the couplers? May be hanging the absorption rapped in black fabric a few inches away from the couplers ? Just trying to see if there is a lighter option... but i know sometimes its about mass and mass, is mass...

If it has a perf panel in front, then it isn't a bass trap! It's a tuned trap, which might or might not be good for bass, depending on the tuning.

I guess my question was, wouldn't a resonator (tuned for 40hz) with insulation backing to widen the q be more efficient/smaller? I can build most things, just pointing it out to see if you think its a better option for the room.

Also, and maybe this is splitting hairs, but you can't "dissipate" a mode: All you can do is "damp" it. The mode is still there, but it just doesn't "ring" any more, because it is heavily damped.

Yes I used the word dissipate thinking of sound being transformed in heat, I am aware of the concept but thank you for making sure! The room is ready for a new test! Which will probably happen tomorrow late night! Can't thank you enough for this, learning a lot and making my room better are two wonderful things! Ale - Stuart -[/quote]

Soundman2020 wrote:

Just saw the linked thread. What a fantastic job!!!!! I don't think i've ever seen a flatter low end response in a room!!!

Thanks! It wasn't easy to do, but it can be done, with careful design, careful testing, and careful tuning.

142 1/2 inches wide 145 " tall 164" is the long wall where

Well that explains the low end! :) The modal issue at around 48 is actually two at once: 0.0.1 axial and 0.1.0 axial, at pretty much the same frequency (46.8 and 47.6), because the profile is square! The cloud will help a little with that, provided that it is huge, hard-backed, and angled. The one "around 80" is your 2.0.0 axial at 82.7 Hz., and the one "around 120" is your 3.0.0 axial, at 124 Hz. In other words, you don't have enough bass trapping in the length and height axes, and width is also deficient.

Its very hard to explain but basically part of the back wall is an irregular giant bass traps

That's fine, but there still isn't enough in the low end.

I also do think it could be sbir and need to play with placement ( speakers are ported in the back ) and right now i started with the speaker as close as possible to the back wall and eventually work my way back and re test.

There's no need to play: Just set the up correctly! :)

Also having width and height of the room very similar it might be more of a modal issue?

Yep!

although trapping 40hz.. unless i use pressure traps and get the q in the right spot...

You'd be surprised... the wavelength at 40 Hz is 28.25 feet, so theoretically it is feasible to get usable trapping from 23" of porous absorber for normal-incident sound, and as little as 12" for randomly incident sound. Most people don't realize that porous absorption is effective down to rather low frequencies.

I can't attach the MDAT it says: the extention MDAT is not allowed...

Upload it to a file sharing service, such as Dropbox, then post the link here.

As you can see i am a bit all over the place right now...

First things first: A lot of your issues arecoming from the incorrect speaker placement. the very first thing you need to do is fix that. Your speakers need to go right up tight against the front wall, except for a gap of 4" where you will insert a panel of 4" 703, directly in between the rear corner of the speaker and the front wall. They also need to go on stands... Heavy! Massive! Lots of weight! Some people stack up concrete blocks or bricks, others make it from hollow steel profile and fill it with sand. But whatever you do, it has to be HEAVY. It also has to be built to the correct heights, such that the acoustic axis of the speakers ends up at 1.2m above the floor, which is 47-1/4". Or maybe a bit higher. Perhaps as much as 49", but not higher. You also need turn your speakers the right way up! You have them laying on their sides, which is terrible, and is probably part of the problem with the frequency response in the high end. And you also need to get the in the correct location with respect to the side walls: each speaker must go 39" from the side wall, in your room. Which means that they will be 64.5" apart. Once again, I'm talking about the acoustic axis of the speakers, NOT the top, bottom, or sides of the cabinet. Check the manual to find out where the acoustic axis is on your speakers. OK, once you have your speakers set up correctly, you'll need to set up your listening position. For that room, you need to set up your chair such that your ears will be 61" from the front wall when you are mixing normally. Then set up your desk comfortably in front of that. Now set up a vertical pole of some type (eg, mic stand) about 16" behind your head (in other words, 76" from the front wall), and carefully aim each speaker such that it points directly at the pole. Here too, I'm talking about the acoustic axis: The acoustic axis must point at the pole. Double check all those measurements. That's the theoretical ideal geometry. Now calibrate your gain structure, and REW, such that each speaker is producing exactly 80 dBC with REW putting out a -10dB signal, and everything in the signal chain pegged at either 0 dB (analog gear) or -20 dBFS (digital gear). You will need a hand-held sound level meter to do this. If you do that correctly, when each speaker produces 80 dBC by itself, both of them together will automatically produce 86 dBC. With that situation,m calibrate REW to tell it that the level is 86 dBC when the meter is showing 86 dBC. Use the full-spectrum test signal in REW, NOT the sub-woofer test signal! With that setup, run three REW tests: one for the left speaker alone (right one turned off), one with just the right speaker (left one turned off), and one with both. Post the MDAT file with those three tests, with the room set up optimally. - Stuart -

Hi There! I've re done all the measurements! And for the sake of science I did also L/R measurements with and without the 703 back panels behind the speakers. I only had 3 inches... but given the results I WILL GET 4" for sure :). I hope these test will also help other people, i think they are incredibly useful!! Here we go : https://wetransfer.com/downloads/ce49fc ... 445/ebaa4c Steward i would Love to get your opinion, I've re positioned everything as you described using a tape measure got things as accurate as i could. Thank you! ps: very interesting the difference between left and right speakers... seems almost unreal? How can that be?

I downloaded your zip file, but it is invalid. When I try to extract it, I get dozens of error messages like these: Documents\new measurements studio 42 .zip: Attempting to correct the invalid file name Documents\new measurements studio 42 .zip: Renaming new measurements studio 42 \L spkr 703.mdat to new measurements studio 42_\L spkr 703.mdat Documents\new measurements studio 42 .zip: Cannot create new measurements studio 42 \L spkr no 703.mdat Also, it looks like you didn't follow the instructions! You saved each new measurement into an entirely separate MDAT file, when what I really need is to have them all in the same MDAT file, to facilitate comparison. Yes, it would be dead easy to put them together (if I could actually extract the files), but it would be an awful to easier if you were just to keep the measurements all in the same file originally! So please put all of the measurements in the same file, then put that on Dropbox WITHOUT compressing it in a zip file, and post the link. - Stuart -

Soundman2020 wrote:

I downloaded your zip file, but it is invalid. When I try to extract it, I get dozens of error messages like these: Documents\new measurements studio 42 .zip: Attempting to correct the invalid file name Documents\new measurements studio 42 .zip: Renaming new measurements studio 42 \L spkr 703.mdat to new measurements studio 42_\L spkr 703.mdat Documents\new measurements studio 42 .zip: Cannot create new measurements studio 42 \L spkr no 703.mdat Also, it looks like you didn't follow the instructions! You saved each new measurement into an entirely separate MDAT file, when what I really need is to have them all in the same MDAT file, to facilitate comparison. Yes, it would be dead easy to put them together (if I could actually extract the files), but it would be an awful to easier if you were just to keep the measurements all in the same file originally! So please put all of the measurements in the same file, then put that on Dropbox WITHOUT compressing it in a zip file, and post the link. - Stuart -

Ops.. Sorry about that! I will do that now!

Soundman2020 wrote:

I downloaded your zip file, but it is invalid. When I try to extract it, I get dozens of error messages like these: Documents\new measurements studio 42 .zip: Attempting to correct the invalid file name Documents\new measurements studio 42 .zip: Renaming new measurements studio 42 \L spkr 703.mdat to new measurements studio 42_\L spkr 703.mdat Documents\new measurements studio 42 .zip: Cannot create new measurements studio 42 \L spkr no 703.mdat Also, it looks like you didn't follow the instructions! You saved each new measurement into an entirely separate MDAT file, when what I really need is to have them all in the same MDAT file, to facilitate comparison. Yes, it would be dead easy to put them together (if I could actually extract the files), but it would be an awful to easier if you were just to keep the measurements all in the same file originally! So please put all of the measurements in the same file, then put that on Dropbox WITHOUT compressing it in a zip file, and post the link. - Stuart -

Here we go! This should work! https://studio42brooklyn.wetransfer.com ... 342/3eb0a3

OK, that one at least worked to download, but clearly you did not calibrate REW: it says you performed all of those tests at a level of 140 dBC! :ahh: That's a little hard to believe, since that's roughly the level you would hear if you stuck your head inside the tailpipe of a 747 jet engine just as it is pushed to maximum power for take-off... :) There's also no sound-card calibration data ... So please calibrate REW correctly, following the complete calibration procedure laid out in the manual, and using a proper hand-held sound level meter. The re-do the tests, and post the new results. - Stuart -

Soundman2020 wrote:

OK, that one at least worked to download, but clearly you did not calibrate REW: it says you performed all of those tests at a level of 140 dBC! :ahh: That's a little hard to believe, since that's roughly the level you would hear if you stuck your head inside the tailpipe of a 747 jet engine just as it is pushed to maximum power for take-off... :) There's also no sound-card calibration data ... So please calibrate REW correctly, following the complete calibration procedure laid out in the manual, and using a proper hand-held sound level meter. The re-do the tests, and post the new results. - Stuart -

Yes i know... i had to hide during the sweep hahahaha But REW wouldn't run the test saying it was too low.. I calibrated the level matching the db out of the speakers to the what the mic was hearing, so i guess the measurement even if loud should be valid? Sorry about that...I thought it was strange...that i had to endure that loud torture...hahaha

Also i was attempting the calibration of the sound card, I am using an Orion 32 ad/da with all db 25 connectors...so If i want to create a loop I have to get an extra db25 to loop around it... I figured its a pretty high quality and flat sound card..... Is it critical to test it? If it really throws everything off I will set it up tomorrow and re run all the test.. Sorry about that. ALe

I calibrated the level matching the db out of the speakers to the what the mic was hearing, so i guess the measurement even if loud should be valid?

And at what point did you use your hand held sound level meter to confirm all that? Unless you tell REW what the real, actual , physical sound pressure level is in the room, it has absolutely no way of knowing. What is the make and model of your hand-held sound level meter, and how did you set the controls when you used it to calibrate REW?

But REW wouldn't run the test saying it was too low..

that's because you did not calibrate REW.

so If i want to create a loop I have to get an extra db25 to loop around it...

Nope! You just have run an XLR cable from the point where you are feeding your actual signal out, to the point where your mic connection comes in, and turn down the input gain on the mic pre-amp by maybe 50 dB or so for the sound card calibration, then turn it back up again and reconnect the speaker and mic for the actual REW sound level calibration using your hand-held sound level meter. (You might also need to turn off phantom power on that mic input when you do the test).

I figured its a pretty high quality and flat sound card..... Is it critical to test it?

There will be a roll off in the low end, as well as some slight instability and a sudden drop in the high end, on pretty much any interface. It probably isn't a big deal at all in the vast majority of interfaces but it can be important in some cases, and it does show that things have been set up correctly. What mic are you using for this? (Make and model)

If it really throws everything off I will set it up tomorrow and re run all the test..

Considering that you should be seeing 86 dBC at the mic (real physical level) and REW is telling you that you are seeing a level about one million times too high, then yes, it might be an idea to calibrate and re-test! :) (Do the math: 146 dB is, indeed, one million times more intense than 86 dB... six orders of magnitude. That's roughly the difference between the weight of one brick, and the weight of your entire house, including the furniture and the cars in the garage...) - Stuart -

Soundman2020 wrote:

I calibrated the level matching the db out of the speakers to the what the mic was hearing, so i guess the measurement even if loud should be valid?

And at what point did you use your hand held sound level meter to confirm all that? Unless you tell REW what the real, actual , physical sound pressure level is in the room, it has absolutely no way of knowing. What is the make and model of your hand-held sound level meter, and how did you set the controls when you used it to calibrate REW?

But REW wouldn't run the test saying it was too low..

that's because you did not calibrate REW.

so If i want to create a loop I have to get an extra db25 to loop around it...

Nope! You just have run an XLR cable from the point where you are feeding your actual signal out, to the point where your mic connection comes in, and turn down the input gain on the mic pre-amp by maybe 50 dB or so for the sound card calibration, then turn it back up again and reconnect the speaker and mic for the actual REW sound level calibration using your hand-held sound level meter. (You might also need to turn off phantom power on that mic input when you do the test).

I figured its a pretty high quality and flat sound card..... Is it critical to test it?

There will be a roll off in the low end, as well as some slight instability and a sudden drop in the high end, on pretty much any interface. It probably isn't a big deal at all in the vast majority of interfaces but it can be important in some cases, and it does show that things have been set up correctly. What mic are you using for this? (Make and model)

If it really throws everything off I will set it up tomorrow and re run all the test..

Considering that you should be seeing 86 dBC at the mic (real physical level) and REW is telling you that you are seeing a level about one million times too high, then yes, it might be an idea to calibrate and re-test! :) (Do the math: 146 dB is, indeed, one million times more intense than 86 dB... six orders of magnitude. That's roughly the difference between the weight of one brick, and the weight of your entire house, including the furniture and the cars in the garage...) - Stuart -

I think i finally got everything right! Thanks for your patience! Here we go! https://wetransfer.com/downloads/94e980 ... 410/997bf2

Here link to the .cal file in case its not embedded in the previous one. https://wetransfer.com/downloads/ef4...3235017/aadaad

I think i finally got everything right!

Much better! :thu: That data is usable, and appears to be valid. And what a difference, as compared to the original situation! Look at tour full-spectrum frequency response, especially in the high end: Red is the original layout, green is the new layout. All of that ugly "hash" in the high end is gone. And look at your low end, waterfalls: Once again, red is the original layout, green is the new corrected layout. For some reason, you cut off the very low end by only starting your REW test sweeps at 30 Hz: it should be about 15 Hz. You also stopped the sweep at 20 kHz: you should allow it to go all the way to 22 kHz. So for future tests, make sure you correct that. Your decay times overall are very nice: Reasonably smooth and even across the board, at around 200 ms. A little on the dry side, but not too bad. The rise in the high-end is a little hard to explain: Are you SURE you have the controls set flat on the rear of your speakers? No EQ applied anywhere in the signal chain? I did notice that you included both the sound-card calibration file and also the mic calibration file in all of the readings: the sound-card calibrations looks about right, typical of most interfaces, but the mic calibration file looks VERY strange. I've never seen a mic calibration curve like that! In fact, it looks very much like a sound card calibration curve! Where did you get that file? Did it come with your mic, or did you download it from the internet somewhere? Also, what mic are you using for these tests?

I even learned how to quote!

:thu: :thu: Just don't forget how you did that!!! :)

I ope these test will also help other people, i think they are incredibly useful!!

Absolutely! Yes! Definitely! So often people want to argue with me about why they don't want to follow advice, and why they think it's a terrible thing to set things up the way I tell them.... but here's proof that it works. All you did was move things around to the optimum theoretical positions, and look at how things changed! So it is VERY useful: In the future, I can point those "stubborn" folks to your thread, then maybe they'll get the idea: just do it! it works.

very interesting the difference between left and right speakers... seems almost unreal? How can that be?

It's not uncommon to have differences between the channels initially, especially with low cost speakers in a basically untreated room. There might be differences between the speakers themselves (we can check...), and there might be differences in the treatment so far... and there might by something(s) in the room causing asymmetrical response. To check if there are differences between the speakers themselves, run a test with just L, another test with just R, then very carefully swap the speakers physically: put the left one where the right one is, and vice versa, but without changing anything else in the room! And make very sure you set up the speakers perfectly in this new location again, in the exact correct angle, position, height, etc. Then run another L-Oonly and R-Only test, and post the MDAT. (Before you start these tests, make sure you adjust REW to run the sweep form 15 Hz to 22 kHz!) Please post a photo of how the front end of the room looks now, and another photo of how it looked before, so the "stubborn" folks can see that all you need to do is to move things around to the right place to get rather nice changes... then we'll carry on from there, for the next steps... - Stuart -

And what a difference, as compared to the original situation! Look at tour full-spectrum frequency response, especially in the high end:

Not preserved: studio-42-original-vs-correct-layout.jpg

Yes Pretty crazy! Thank you!

For some reason, you cut off the very low end by only starting your REW test sweeps at 30 Hz: it should be about 15 Hz. You also stopped the sweep at 20 kHz: you should allow it to go all the way to 22 kHz. So for future tests, make sure you correct that.

Will do!

Your decay times overall are very nice:

Not preserved: studio-42-corrected-rt60--1..3.jpg

I am very happy about that! I believe the rise in the higher end is the flutter echo that i can still hear... after cloud it should be gone i hope!

The rise in the high-end is a little hard to explain: Are you SURE you have the controls set flat on the rear of your speakers?

Controls are all flat

I did notice that you included both the sound-card calibration file and also the mic calibration file in all of the readings: the sound-card calibrations looks about right, typical of most interfaces, but the mic calibration file looks VERY strange. I've never seen a mic calibration curve like that! In fact, it looks very much like a sound card calibration curve! Where did you get that file? Did it come with your mic, or did you download it from the internet somewhere? Also, what mic are you using for these tests?

I am using the Beringher ecm 8000 - i do not recall doing anything else than calibrating the spl for the mic, so the second graph might just be the interface calibration? I wasn't aware of that... is there any way to adjust it?

So often people want to argue with me about why they don't want to follow advice, and why they think it's a terrible thing to set things up the way I tell them.... but here's proof that it works. All you did was move things around to the optimum theoretical positions, and look at how things changed! So it is VERY useful: In the future, I can point those "stubborn" folks to your thread, then maybe they'll get the idea: just do it! it works.

Fantastic! At least i get to give back something!

It's not uncommon to have differences between the channels initially, especially with low cost speakers in a basically untreated room. There might be differences between the speakers themselves (we can check...), and there might be differences in the treatment so far... and there might by something(s) in the room causing asymmetrical response.

I will attach a panoramic of the studio, I do have the cntrl room window to my right and a little bit of flat drywall on my back left ( the couch should go there). Everything else is pretty symmetrical.

To check if there are differences between the speakers themselves, run a test with just L, another test with just R, then very carefully swap the speakers physically: put the left one where the right one is, and vice versa, but without changing anything else in the room! And make very sure you set up the speakers perfectly in this new location again, in the exact correct angle, position, height, etc. Then run another L-Oonly and R-Only test, and post the MDAT. (Before you start these tests, make sure you adjust REW to run the sweep form 15 Hz to 22 kHz!) Please post a photo of how the front end of the room looks now, and another photo of how it looked before, so the "stubborn" folks can see that all you need to do is to move things around to the right place to get rather nice changes... then we'll carry on from there, for the next steps...

Great trick! I'll try to find a pics of the older set up ! Also, how do you think i can improve the low end dips further? Moving speakers and re testing? trying to understand the nulls and peaks in the room? Now things seem to have changed which axis do you think its causing the unbalances? Thank you again!

I believe the rise in the higher end is the flutter echo that i can still hear... after cloud it should be gone i hope!

Nope! Flutter echo won't show up much on the frequency response graphs: you'll see it best on the spectrograms, probably.

i do not recall doing anything else than calibrating the spl for the mic, so the second graph might just be the interface calibration? I wasn't aware of that... is there any way to adjust it?

I figured it out: you had the "C weighted SPL meter" button pressed on all your measurements, except the original one. Here's the comparison with that turned off: Your high end is good now. About as flat as you could ask for. The very high end (above 15k) roll off might be due to excessive treatment, or it might just be the speakers.

I'll try to find a pics of the older set up !

Maybe take a still from your video?

Also, how do you think i can improve the low end dips further? Moving speakers and re testing? trying to understand the nulls and peaks in the room? Now things seem to have changed which axis do you think its causing the unbalances?

There's a test that I do to help figure that out, that I call the "walking mic" test, but it's rather time consuming for you to do, and for me to analyze. Unfortunately I'm out of time right now! I have deadlines that I need to meet with several of my paying customers, so unfortunately I just don't have the time to do that with you right now. It will have to wait a bit. Sorry! What you can do in the meantime is to increase the 703 between your speakers and the front wall to 4", and make sure those speakers are tight up against it, with the rear corner practically touching the front surface of the 703. And also do the cloud: Make it large, hard-backed (3/4" MDF or OSB on top, 4" of 703 below, another 4" on top) and tilted at least 10°, or more (lower at the front. over the speakers, higher over the desk). It will be heavy. Use chains to attach it to the ceiling or whatever framing you set up, NOT wire. Use only closed hardware: not open hooks that could come loose. Use mounting hardware that is rated for at least ten times the weight it will support. Follow all the normal safety procedures that are used by riggers and gaffers to safely hang things overhead. Do REW tests after each of those. - Stuart -

Here the Set up Before the changes. Speakers laying down and general geometry out of wack. Sorry terrible pics takes from the first video i posted in the thread.

Here the picture of the adjusted set up and consequent change in REW analysis. Also showing a panoramic ( missing couch on the left side ) cntrl room window on the right and back wall pictures.

Soundman2020 wrote:

I believe the rise in the higher end is the flutter echo that i can still hear... after cloud it should be gone i hope!

Nope! Flutter echo won't show up much on the frequency response graphs: you'll see it best on the spectrograms, probably.

i do not recall doing anything else than calibrating the spl for the mic, so the second graph might just be the interface calibration? I wasn't aware of that... is there any way to adjust it?

I figured it out: you had the "C weighted SPL meter" button pressed on all your measurements, except the original one. Here's the comparison with that turned off:

Not preserved: studio-42-original-vs-correct-layout-no-c.jpg

Your high end is good now. About as flat as you could ask for. The very high end (above 15k) roll off might be due to excessive treatment, or it might just be the speakers.

I'll try to find a pics of the older set up !

Maybe take a still from your video?

Also, how do you think i can improve the low end dips further? Moving speakers and re testing? trying to understand the nulls and peaks in the room? Now things seem to have changed which axis do you think its causing the unbalances?

There's a test that I do to help figure that out, that I call the "walking mic" test, but it's rather time consuming for you to do, and for me to analyze. Unfortunately I'm out of time right now! I have deadlines that I need to meet with several of my paying customers, so unfortunately I just don't have the time to do that with you right now. It will have to wait a bit. Sorry! What you can do in the meantime is to increase the 703 between your speakers and the front wall to 4", and make sure those speakers are tight up against it, with the rear corner practically touching the front surface of the 703. And also do the cloud: Make it large, hard-backed (3/4" MDF or OSB on top, 4" of 703 below, another 4" on top) and tilted at least 10°, or more (lower at the front. over the speakers, higher over the desk). It will be heavy. Use chains to attach it to the ceiling or whatever framing you set up, NOT wire. Use only closed hardware: not open hooks that could come loose. Use mounting hardware that is rated for at least ten times the weight it will support. Follow all the normal safety procedures that are used by riggers and gaffers to safely hang things overhead. Do REW tests after each of those. - Stuart -

Hey Stewart I completely understand ! Very grateful for your help with all of this ! with the holidays coming up I will also have to rush on project and go back home for a coiple of weeks! So i might bump the thread once i am back around january. Thanks again and have a great Christmas! Ale