help optimizing acoustics in control room

[rez]

hello people at gearslutz,

after visiting this board as a guest for years now i decided to take the opportunity and post my first thread here.

a friend of mine has built a nice, little studio, but unfortunatly the bulid was not finished due to several circumstances. long story short: we now decided to finish the build which is primarily accoustic optimization.



when they bulit the room they choose a nice layout with angled walls and relatively good proportions. the room length of 4.7m is a bit close to the width at 4.6m.



all the yellow parts are the planned bass trapping areas. the front wall is less than ideal because of the two recessed windows and the heating. at the backwall there is a door in one corner.



the empty room sounds okay but there is some fluttering audible.

[rez]

oh, i hit the wrong button...

here are the waterfalls: the pink one is the empty room and the green one is the room with four packets of yellow isover accoustic glasswool just stacked in the front corners, plus four 50cm and 2m high plates of accoustic board vertically against the wall behind the sofa (to form a pseudo diffusor they are aligned like this: vvvv).









we were surprised how much of a difference just the stacked packets of glasswool made to the rt60 time.

what do you think of our planned placing and amount of basstraps? the big trap at the rooms rear end to roof is our biggest concern - is the placement good? this is the place in the room that can fit the biggest amout of trapping so i tought i might be a good design.
at the front corners i like to make the traps a little longer into the room up to the point the window to the recording booth begins (the two rectangular traps on each side will not be built).

can you help us a little in interpreting the graphs? if you need some other graphs or format please let me know.

we will document the build in pictures so you can see.

greetings from germany,

peace, rez

[takman]

the after...green waterfall looks pretty good. you can see that adding the absorption reduced the ringing at 75hz and 100hz.

With bass trapping, its pretty simple. get as much bass trapping as you can do.

I think you can knock down that mode at 75hz more, and generally smoothen things out with more bass trapping.

also, symmetry is not really a big deal with bass traps. try to keep it symmetrical as much as possible.....but if the placement is asymmetrical in places, thats ok. Generally though, if you can keep the bass trapping on the front wall symmetrical, that would be best.

And dont let that door stop you from putting bass trapping there, get creative!

We are not really concerned with the RT60 graph, as this is a small room.

More importantly we need the Envelope Time Curve graph. Set the time (ms) to 0-100.

[rez]

thank you for your fast response!

here are the etc graphs. the pink one is the empty room (with the console and the cloud in place!) and the green is with the stacked packets glasswool in the front corners:





hopefully i choose the right one as rew does not select this per default. normaly this graph shows the impulse response. i also noticed some up and down in this graph called "step response":





could that be a sign of flutter echo between the console and the cloud?
the cloud is actually a hard surface with many holes in it, but no absorbing material on top. and i argue that the angle of the cloud might not be step enough so it reflects to the lister position instead to the back wall. have to do a mirror test on that.

to me the phase graphs also looks strange:





as you see we have planned to place a basstrap directly onto the door to maintain symmetry with the other corner. we do not know if we should make these two basstraps just porous absorbers or a mass loaded design with a heavy plate at the front, because of the lack of depth (max. 20 cm!).



when building the corner chunks in the front and at the back wall / roof we have not yet decided if we should fill them completly with glasswool or leave an (small) airgap between the absorbing material and the walls?



the monitors used are ns10 and there is an jbl sub in the middle of the monitors on the floor at about 1m from the front wall firing directly to the back wall. maybe we also should experiment a little with sub placement?
the floor is carpet because the owner opted so despite other advice, but there is at least some hard surface on the floor where the chair is rolling.
overtrapping is not an issue because than i would build reflecting panels to the front of the big trap at the back wall / roof to regain some reflections.

please tell me if you need more info.

peace, rez

[localhost127]

please repost the ETC (one speaker at a time) with more resolution in the 0-100ms area. looks like you have a loud early reflection, likely off of the working surface.

this is why it is likely best (if not soffit mounting) to place the speakers on stands several feet in front of the working surface (towards the front wall) and angle the desk downward so that the reflection is not incident off of the desk.

[Glenn Kuras]

when building the corner chunks in the front and at the back wall / roof we have not yet decided if we should fill them completly with glasswool or leave an (small) airgap between the absorbing material and the walls?

If you can fill the whole area then go for it.

[rez]

thank you for the responses.

i will do the measurements with one speaker at a time as soon as possible. why do we need the response from only one speaker, when the room is symmetrical and we always monitor in stereo? wouldn't this be more realistic? or do we gain information that would be masked otherwise?

thank you for ensuring me to fill the cornertraps completly with glasswool. there were some designs on the internet that suggested more efficiency if there was an airgap between the insulation and the wall.
the big trap at the room end / roof might be a little more effort to build, but i like to do it, because i can fit more than the double amount of absorption material in there than i can put into the front traps.
can anybody say something regarding the two traps in the back corner and on the door? should i just build 20cm deep porous absorbers and mount them with a small airgap to the wall / door? or is it better to build something with a heavy plate at the front?

thank you for your help.

peace, rez

[takman]

Originally Posted by rez

thank you for the responses.

i will do the measurements with one speaker at a time as soon as possible. why do we need the response from only one speaker, when the room is symmetrical and we always monitor in stereo? wouldn't this be more realistic? or do we gain information that would be masked
otherwise?

The ETC needs to be done with one speaker, because we are looking for reflections......has nothing to do with the frequency response. You fire one speaker, see where the reflections are. Then you fire the other and see where the reflections. By running both at the same time..well, maybe someone with more knowledge than me can explain. But, yes, something along the lines of it masks the information, or alters the information in a way that makes it hard to decipher.

Regardless, the ETC is very important! You need to create an RFZ in the listening position, then introduce reflections after the ISD. Which is the whole reason you have QRD on the back wall.

thank you for ensuring me to fill the cornertraps completly with glasswool. there were some designs on the internet that suggested more efficiency if there was an airgap between the insulation and the wall.
the big trap at the room end / roof might be a little more effort to build, but i like to do it, because i can fit more than the double amount of absorption material in there than i can put into the front traps.
can anybody say something regarding the two traps in the back corner and on the door? should i just build 20cm deep porous absorbers and mount them with a small airgap to the wall / door? or is it better to build something with a heavy plate at the front?

The Gap between the wall and absorber needs to be large enough, a small gap wont do much. Why? Well, the sound wave is at 0 velocity when it strikes the wall. if you have absorption there, its not absorbing anything. As the wave leaves the wall and heads back into the room, its speeds up. We want to have the absorber right at the point where the wave is at maximum velocity. Its typically more cost effective just to add an air gap, rather than add more material. Thats why they say, a 4 inch absorber with a 4 inch gap is like having an 8 inch absorber. why put material there if its not doing "anything" anyways? Well, i shouldnt say its doing nothing...it is doing something. But its just not very efficient, until the wave is at maximum velocity.

thank you for your help.

peace, rez

master handbook of acoustics has a lot of good information if you want to read about the ISD, RFZ etc.

heres an example...

[johndykstra]

Hey Rez.

ETC and any measurement for that matter, are preferred with a single speaker firing so you know which speaker it is that is causing the anomaly. On the ETC for example... (and the scaling is off... set the initial sound (big spike) as the "zero", and really you only need like, 50msec there after...) the amount of msec between direct and reflection tells you how far the surface that creates the reflection is. speed of sound times amount of delay= extra travel time. Cut a length of string that corresponds with what you find. Hold one end of the string on the speaker, and the other on the mic. Pull the slack towards room surfaces and when it touches something, there is your reflection point.

ETC is an analysis for reflections to ONE POINT... meaning, you should move the mic around to all the places you will find your head to be while mixing, and hunt down all the reflections within a mixing zone. Now do all of that again with the other speaker. Many/most will be the same in a symmetrical setup... but best to check to be safe.

[johndykstra]

The cloud being reflective... very cool. I would put absorbent on top of it though.

The idea of a reflective cloud being to preserve some amount of reflected energy, by redirecting it rather than absorb it. What you want here is not only preventing the reflected energy from hitting you directly in your mix spot, but also aiming those reflections at your rear wall diffuser. This should make for a denser, more pronounced psycho-acoustic "haas effect".

Verifying it is missing the mix spot will be achieved with the ETC. Aiming it to the diffuser can be done this way too... cover the diffuser with absorbent, place the measure mic right in front of the diffusor, and verify that you are infact, getting reflections off the cloud using the string cut to distance again. Or, you can hold up a mirror on the cloud's surface, and hold a laser pointer on the speaker to represent sound waves. See if the laser hits the diffusor squarely. Adjust.

You can mimic this same concept existing with the cloud with side wall deflectors as well.

Generally speaking, it's common that one continuosly large angled deflector will not be adequate to successfully redirect reflections for a mixing zone, and it may be necessary to employ an array of smaller panels. Imagine a "saw tooth" orientation, where each panel is a separate "tooth" of a saw.

[rez]

thanks for your responses.

yes that cloud might be part of the problem. to me it seems that it was more design driven than acoustically as it is at the same angle as the roof. when that design decisions were made i was not yet involved in this studio build.
the cloud is a heavy construction made of wood and has many holes with appr. 2cm diameter. it is part of the light concept and there are also lights on top to make some indirekt light. but we are willing to change what has to be done to accieve better sound. to bring the whole cloud to a steeper angle can be done easily as it only involves changing the length of the four hangers the clowd is mounted to. bringing just the front half to a steeper angle is way more difficult, because then we must cut the cloud and turn it around 90° as it has a division in the middle that would then look better this way. as you can see here even with just guessed angles it shows that there might be reflections to the listening spot indeed.



some mirror testing has been done but i have to do some more to get the info myself. and thank you for the suggestion with the string that is really a fast way to check certain frequencys reflections.

the monitors are actually on stands, but the console is on top of a table that is 10cm too high at the moment. they have to bring the new table in before we can actually verify reflections from the console.


peace, rez

[ciro]

I´d focus on the "shy" low end, you have the ~75 peak and a roll of, almost nothing below 60hz. Move your mix position and monitors to see if helps. Sub woofer between monitors is not an easy thing to deal (there´s some great posts about it here).

Ciro

[rez]

yes, i see and hear the problem in the bass area. the 75hz spike and the loss of sound at 60hz and down can be due to a room mode, also at 90hz and 120hz.

when i get to the studio i will experiment with subwoofer placement. are there any suggestions on what i can avoid to test, because it will not lead to a better result? the method i will definetly try is to put the sub into the mix spot and then crawl around to look for the perfect place.

another thing i have a problem with: since this system is passive the ns10 are connected to the satelite terminals of the subwoofer. that means there is no way to take the subwoofer out of the picture without really disconnecting it. the sub has a cutoff at 160hz which is not adjustable. will the ns10 not sound different if they have do deliver 80-160hz when used alone as opposed to only 160hz and up when connected to the sub?
my personal preference would be to have the sub connected seperatly, but how can this be done with a passive sub? to my knowledge this need a termination on the satelite outs with 4 or 8 ohm. the amplifier used can power two independend sets of speakers. also what would happen to the overlapping fequencys if the filter of the sub is not used? will the freqs from 80-160hz then be played by both ns10 and the sub?

and how do i measure the full frequency spectrum now? first the left and the right speaker alone and than the subwoofer alone? that leads to the problem mentioned above: how can i use the passive sub without satelites?

thank you for the support.

peace, rez

[DanDan]

There are simpler ways to measure which also identify a solution.
Use the Sig Gen in REW turn on Frequency follows cursor. Place the cursor over the strong 75Hz mode. Move about the room, find the the nulls and hot spots. Mark them on a map or on masking tape on the floor. Do the same with a couple of other nulls and peaks. Now you have a map of where problems are.
You will also have a map of boundary areas with peaks. Bass trap there.
You can drive one speaker plus sub easily 1.1. Just disconnect the other speaker. There is a great methodical exercise in sub integration here
Demo Room
DD

[boggy]

It's a bit complicated than usual, and I'll try to help with measurements... if I can.

Firstly, I'll use stands for loudspeakers and subwoofer.
Then, I'll place subwoofer on a stand at same height like Yamaha's, but in center area, between satellites, and behind VGA (LCD!) display... at same distance from listening place, then I will measure complete system... and I'll move complete system to find a better position. You may use any cheap and strong enough stand that you may find in local store... later, you can build nicer, better and heavier stands... for now ... it's not important.

Originally Posted by rez

.......
and how do i measure the full frequency spectrum now? first the left and the right speaker alone and than the subwoofer alone?

I'll try to measure separately left and right channel (one in same time, you can mute other channel at line level, or disconnect cable from amplifier input) but with subwoofer. If this is a single subwoofer for both Yamaha's this is a bit complicated because at low end, subwoofer will be louder than one single NS10

Originally Posted by rez

that leads to the problem mentioned above: how can i use the passive sub without satelites?

Please, don't. Unpredictable response may occur...

Second, and probably best possibility, is to measure both channels in same time with sub (this would be "flat") but only for frequencies from... about 200Hz and down. You may easily ignore "events" and "artifacts" above 200Hz.

And, if possible, exporting wav file of measurements and attaching it here may enable others and more skilled to better help you.

Hope this helps...

Bogic

[localhost127]

Originally Posted by rez

they have to bring the new table in before we can actually verify reflections from the console.

why not place one of your broadband absorption panels on top of the working surface. if the reflection is attenuated further or disappears altogether, there's your answer.

or more correctly, use the string method as mentioned above. it'll be a good habit to get used to.

[boggy]

Originally Posted by localhost127

why not place one of your broadband absorption panels on top of the working surface. if the reflection is attenuated further or disappears altogether, there's your answer.

or more correctly, use the string method as mentioned above. it'll be a good habit to get used to.

It's easier and safer to move mixing desk to the floor, table might be removed from room... while measurements going, at least.

Cheers

[DanDan]

The destructive effect of reflections can be a lot easier to see than the little spike in the ETC. This 'magnified' view makes it very easy to see if you have fixed them. A couple of blocks of wood under the back legs of your desk would fix this. If it is a problem.
Desktop Reflection Revealed

Note the string thing, while ingenious and thanks to SAC for it, identifies reflections by length, not by angle. You might find a surface which fits the string but causes no reflection to the listener, due to angles.
DD

[takman]

Originally Posted by DanDan

Note the string thing, while ingenious and thanks to SAC for it, identifies reflections by length, not by angle. You might find a surface which fits the string but causes no reflection to the listener, due to angles.
DD

im not understanding what you mean.

the reason we know about the reflection is because the mic picked it up, and it was displayed in the ETC. Therefore, if the mic is in the listening position, that reflection, regardless of angle, reached the listening position ( listener).

correct me if i am wrong.

[johndykstra]

Originally Posted by DanDan

You might find a surface which fits the string but causes no reflection to the listener, due to angles.
DD

Sure. But I would hope it's apparent that string contact with a surface is only part of it. The angle at which the string arrives and departs from the surface should be "the same". If the angle is not the same, that is not your spike.

[DanDan]

Jeez, john, the fastest drawer on GS, you even talk graphically...
Also, to restate a Lupo gem. Move the mic towards the suspected reflection. Measure ETC again. If the spike has shifted on the time line, you have that reflection in your sights.
DD

[johndykstra]

For clarity, I'm not suggesting one need break out a protracter here. "eyeballing" it should tell you if you're on the right track or not.

[takman]

i think i see what DD is saying...

i haven't done the string ETC thing yet, but i plan on doing it soon...

i just thought when we calculated the length of the string, that it would only touch one place exactly...

also, when we find the point the of reflection with the string, what about running another string from that point to the mic?

[johndykstra]

I think you are missing something Tak.

The string is cut to represent the total travel time of the reflection. That will include the travel from surface to mic.

Let's say you(mic) are 4' from your speaker. Your ETC shows a reflection at 1msec.

speed of sound 1126 ft. per second

1msec. = roughly 1.13ft. (1126x.001)

Cut a string to 5.13ft.

holding one end on the speaker, and the other touching the mic, pull the slack from the string until it makes contact with a surface. With your finger contacting the surface, holding the string against it, if the angle on both sides of your finger looks the same, that's your surface causing the reflection. If one angle is rather obtuse and the other acute, the surface you are touching is not the correct surface.

unless you have three arms, or are comfortable taping string to speakers and mics, this is a two man operation. Dan via Lupo's suggestion of moving the mic and watching for msec movements, seems more reasonable. If for no other reason than you are going to be executing multiple ETCs around the mix zone regardless.

A tandom of both methods is likely the best option. Keeping in mind that in a symmetrical listening room, most of your work is being done with one speaker... the second one will likely follow the same treatments as the first... just look for anomolies.

[rez]

thank you for all the replies.

here are the graphs for the left speaker (red) and the right speaker (green):













mirror test showed reflection areas on the console and on the front part of the cloud. the cloud and console will be changed to a steeper angle to correct that.

there was not much time for testing today, but i will continue on sunday or monday. the front corner traps also will be built probably in the course of next week.
the lowest frequency the subwoofer can produce is 38hz according to the data sheet. should i start the test at 38hz instead of 20hz because there is nothing to measure if the system cannot deliver? and also can somebody explain the phase curve so i can understand if this is good or bad and what i can do about it. (i know the principles of phase and how a speaker sounds if it is out of phase and to me it looks as if the phase flipping has something to do with my spl problems. or is this simply the visual clue of what happens in a frequency null?)

regarding the back wall questions came up if we can avoid to build the big trap at the wall/roof if we make the two absorbers in the back corner and on the door more efficient? and then decide if the room needs more traping or if we can live with the results?

peace, rez

[takman]

john, i get the string thing now. thanks for clarifying.

although i dont get your math. where did you get 5ft from?

if 1 ms is 1.13....should the reflection be 1.13ft away at 1ms?

for example if a graph shows a reflection at 5ms.... 5x1.13= 5.65ft?

or am i missing something here?

edit: or i should say the total time of flight would be 5.65 ft, not just the time from speaker to the actual reflection.

[takman]

rez, can you post the REW file, so i can open it on my computer?

easier this way, rather than saying post this graph, that graph, make it show these values etc..

[rez]

here is the rew file. it has three measurements in it for the left speaker (red), the right speaker (green) and for the stereo config.

why is this forum censoring my file title? just because there is a german "umlaut" in it? and why does it censor in the middle of the file name? "testfünf" was the name and the forum censored "stfü"? ah, now i see it is censoring "shat tha fakk upp", that is funny... and here you can see that human creativity will always triumph over dumb censorship!

peace, rez

[John White]

Originally Posted by takman

john, i get the string thing now. thanks for clarifying.

although i dont get your math. where did you get 5ft from?

if 1 ms is 1.13....should the reflection be 1.13ft away at 1ms?

for example if a graph shows a reflection at 5ms.... 5x1.13= 5.65ft?

or am i missing something here?

edit: or i should say the total time of flight would be 5.65 ft, not just the time from speaker to the actual reflection.

I took it as 1.13 plus the 4' distance from speaker to mic.

[takman]

Originally Posted by John White

I took it as 1.13 plus the 4' distance from speaker to mic.

yea...

i think i see what hes doing....

the distance from mic to speaker, plus the time of the reflection = total length of string..

the reflection added time to the 4ft distance, because it had to take a longer path than the direct sound.