New member designing a mixing/mastering studio

Hi guys! First, thanks for keeping a forum like that running, I’ve been searching that for quite some time! As I found a house to move in this summer, I now have some space to build a (small but nice) studio with 2 spare room. I am at drawing the plans. I’ve read a lot on this site, on John’s site, on Ethan Winer’s site and many places on the net about acoustic and studio building. But as it is a science that need experience (which you don’t get by reading…), I decided to ask here for advice on the construction of my room. I’ve included a jpeg up to scale of the studio (not the rest of the house…) by itself and one of my first design idea. (I’m new with sketchup…) Room utilization: -Producing my own stuff (recording guitar and vocals only) -Mixing -Mastering -Watching movie (for fun!) -The main goal is a good sounding room! Description: -It is on the second floor -The house is pretty old (1903 or something) so floors and wall are not very straight! -I can make walls if I want -Outside walls are insulated but I really don’t know how. (Since I don’t need insulation for outside noise, I hope this doesn’t matter to much…) -The floor is not insulated (strictly beams, old board and another layer of presswood. I plan to put industrial carpet in the studio). -The ceiling is insulated typically by blown wool (I hope this is the good term, I’m no English) Other facts: -As this house is on the country side without neighbor, outside noise is not a problem worth the price for isolation and construction. -I really don’t like loud music. Really! -I work on computer so there is a room outside the main room to eliminate the van noise. -I am about to buy the screen (typical 55inch flat screen) -My monitors are dynaudio BM5As and a M-audio sbx10 sub. -I didn’t draw the desk yet but you can easily see it is beneath the keyboard. About budget: You have to know I’m pretty handy at finding old stuff that is still good and that I have a great family working around materials. So I can get around much cheaper than buying new, I’d say 60% full-price. That said, I can spend about 1000$ or so if needed. What I need to know: -Are my dimensions good, bad or can it be better? -Recommendations for placement of absorption, reflexion. As the room is pretty small, I was thinking 1foot absorption everywhere. -Should I build an angled ceiling? Is there a way to get around it? -Air ventilation: how should I go about that? (The house is heated by propane in the basement and there is a dock just beside the studio door) If I forgot any details needed for advice, please ask, I’ll be glad to answer! Tommy

"-It is on the second floor" Get it one the first floor. A mastering room, and that is different from a mixing room, will not survive an elevated area.

why can't a mastering room be on a 2nd floor?

I forgot to mention the ceiling is 8 foot. If I don't care about outside noise, I don't see why i couldn't do that room on the second floor... Plus there is no room on the first floor... Thanks! Tommy

TommyDaveHarrisson wrote:

I forgot to mention the ceiling is 8 foot. If I don't care about outside noise, I don't see why i couldn't do that room on the second floor... Plus there is no room on the first floor... Thanks! Tommy

You not caring about outside noise is not a requirement for a mastering studio. Let's think that there is a constant noise level of a certain frequency on the exterior of this structure that is constantly making it's way into the environment. You as a potential mastering engineer will over/under compensate for this frequency that keeps getting into your ears. The roof area being the weakest link...thinking planes/rain and pine cones hitting the roof. And this IS a wooden structure....you have not even met the enemy yet. It's all for the same reason that an acoustical mixing/control room shouldn't be on an elevated level. You can do what you want...but just because you do not care, does not mean that sound and the ability for it to infiltrate the environment with flanking, does care.

Thanks for the advice! It is not that I don't care about outside noise, It is more that there really isn't (is far from the city, really nothing there!) Standing in the room at this moment, without isolation whatsoever, there is almost nothing! And when there is something, its not constant (a car every hour maybe...) I know treating the room for acoustic will provide sufficient isolation from the outside noise considering this very low level of noise. Also, may I remember you another project discussed in this forum about a professionnal mastering studio built on a second floor(designed by John himself): viewtopic.php?f=11&t=13069&start=0 Don't get me wrong, nothing rude here! I just want to clear this so I can start building. Thanks!

on the design - make the soffits wider and use a saw tooth set of treatments instead of long slats.

Thank you so much Gullfo! Is that better? I will analyse the room before calculating slots/slats dimensions. There will be a 4inch cloud above mixing point. the rear wall is bookshell/bass traps, small window and door Tommy

i think that works better - you'll have more space in the room.

Hello again! The studio is now almost done!!!!! It already sounds awesome! I'll post some pics soon! I need advice to finish, I searched for answers but couldn't find Saw tooth resonator on side walls: -Does a resonator attenuate a peak OR fill a dip? (haven't found an answer anywhere...) -Given the dimension below and my room test with a white noise, how do I calculate the slats? (given that the fronts of those walls are angled and the backs parallel to the old walls?)

How did you make that graph? It is somewhat surprising! If I'm reading it right, your room response is flat to within +/- 1dB, across the entire spectrum, from 30 Hz to 20 kHz! That's a pretty amazing accomplishment for a home studio! Most folks are happy to get +/- 6 dB, with a lot of luck and effort, and +/- 10dB is probably closer to the norm. So I'd be really interested in the procedure and software you used to get that measurement! Very., very seldom so you see almost totally flat response in a room, especially one with those dimensions and shape! I strongly suspect that something isn't right with the way that graph was produced...

-Does a resonator attenuate a peak OR fill a dip? (haven't found an answer anywhere...)

Both. Neither. It all depends... Are you talking about dealing with room modes, or something else? In any event, of that graph is correct, do NOTHING ELSE to your room. It is perfect! There is nothing at all you can do to improve a room that has flat response to within 1 dB across the entire spectrum. But assuming the graph is wrong, your "sawtooth wall" might be doing one or more of many things, but we can't tell you unless you provide the details of how it is built. From the diagram, it looks like vaguely like an attempt to create a CID layout, except that the reflective surfaces are in the wrong place. On the other hand, those could be broadband panel traps on the sides, since they just look like very large flat surfaces over sealed cavities. Or they might even be slot walls, and you just didn't show the slots! Then again, the caption on the photo says "Filled with rockwool, covered with cloth", implying that they are not tuned at all, and are only absorption! So it's really hard to say what the devices are doing. Maybe if you could post more details, or even photos of how you built them, plus the complete plans of the entire build, then we might be able to be a bit more use to you! BTW, I also just noticed that you conducted the test with WHITE noise, not with pink noise, which is the way it should be done. So there is actually something major wrong, either with the room or with the measurement setup... That graph does not show the energy distribution of white noise.... - Stuart -

Thanks for the answers, I'll try to add more details:

How did you make that graph?

- I simply recorded a white noise at 88db at listening position. (speaker Dynaudio BM5A, M-audio SBX10, rode NT2a in omnidirectional, fireface 400) -Then put the recorded sound in HarBal (It show an Eq graph of the sound)

That graph does not show the energy distribution of white noise....

- I know it doesn't correspond to the white noise graph, there is mostly more bass due to a personal preference (I did set the subwoofer level only with my ears to personal taste...)

I strongly suspect that something isn't right with the way that graph was produced..

-Maybe, I don't know... you can tell me...I thought I most importantly just need to know the frequency response at listening position...and did wanted to invest in expensive testing product...so a comparison should be fine...

implying that they are not tuned at all, and are only absorption!

-Until now yes they are. I planned to tune them only after testing the actual room (seemed logical to me...and simpler than calculating it before and then get a different ''in life'' result anyway...) -I simply took a resonator sketchup file in the construction section and made it sawtooth as suggested. Its basically a sealed wood box filled with rockwool, then tuned by lumber on the front...right? There is some pics of them being built( don't want to spoil the fun of having all the pictures of the entire studio being built...but wth!) Thanks! Having fun here! BTW it really sounds nice in here...clarity, tightness, details, ambiances, everything in the mixes!...can't believe I had to work without this!

- I simply recorded a white noise at 88db at listening position. (speaker Dynaudio BM5A, M-audio SBX10, rode NT2a in omnidirectional, fireface 400) -Then put the recorded sound in HarBal (It show an Eq graph of the sound)

Then I'd suggest that this "HarBal" thing is your problem! Take a look at the two graphs you posted, and see if you can see the huge discrepancy: The bottom one that you just posted shows the correct energy distribution for white noise, rising constantly at 3dB per octave across the entire spectrum. Yet what you are getting back in again, is a FLAT graph, with NO such increase evident! :shock: That's what I meant when I said that your response curve does not show white noise. So basically you are putting out a signal that through your speakers that gets 3dB louder for each octave, yet when you measure that signal with a mic, it does NOT get louder by 3dB per octave! How do you explain that? :?: The ONLY answer is that there is something wrong. Something in your measurement chain is not telling the truth. But even worse, what the graph you posted yesterday DOES show is virtually flat response, which is just as unlikely for a room of that shape and size. The numbers on the left edge of that graph show that there is less than one dB variation around the center line! That is unlikely in the extreme. So either you are the luckiest guy ever on this forum, and your room turned out absolutely perfect without even trying (in which case you should probably run right out and by a few dozen Lotto tickets, before your amazing luck runs out! :) ), or there is something wrong with your measurements. I'd suggest that you should download REW, and try testing with that. It is known to work well, and tell the truth. If REW shows the same response, I'll be really surprised!

-Until now yes they are. I planned to tune them only after testing the actual room (seemed logical to me...and simpler than calculating it before and then get a different ''in life'' result anyway...)

Well, that basic concept is excellent, but the implementation is not going to work out so well. Here's why:

-I simply took a resonator sketchup file in the construction section and made it sawtooth as suggested. Its basically a sealed wood box filled with rockwool, then tuned by lumber on the front...right?

Right, but wrong. You are mixing two entirely different and diametrically opposed concepts! You have taken a design for a slot wall, and are trying to make it into a panel trap. That simply wont work. A slot wall is a Helmholtz resonator device, and can indeed by made broadband by angling it, since each small section of slot only "sees" cavity immediately behind it. So angling the wall is virtually the same as building hundreds of tiny Helmholtz resonators next to each other, each with a slightly different depth. But a panel trap can't be built that way. The entire panel must resonate over the same depth of cavity. It is tuned to only one frequency, not broadband. The surface density of the panel and the cavity depth set the resonant frequency for the entire panel: You cannot make each part of it vibrate at different frequencies by angling it away form the wall: it acts as one entire unit, and all of it needs to be over the same depth. So, since yo already built the angled boxes, your only choice now is to scrap the idea of a panel trap, and build a slot wall instead. (Either that, or take of the angled boxes, and rebuild them as fixed depth boxes.) - Stuart -

Thanks! I'll give a try with REW! It seem like a pretty complete program. Which test should I run? (I'm completely lost in REW) So basically, could you explain the construction and utility differences between a panel trap and a resonator? It sounds like I mixed them... does a resonator have many ''boxes'' inside? if that is the case, the plan I downloaded was misnamed... Thanks again for your time!

Done the frequency response test with REW! -many similarities I must admit... I must add that the first graph you saw (in Harbal) was in 1/3 of an octave resolution. I can make a jpeg of the same file in 1/12 if your curious... here is my results in REW -1/3 octave -no smoothing I can now see I like bass a lot!

Sorry to interrupt here, but I'd like to clarify something Stuart infers. This should also interest the OP and any others!

Soundman2020 wrote:

A slot wall is a Helmholtz resonator device, and can indeed by made broadband by angling it, since each small section of slot only "sees" cavity immediately behind it. So angling the wall is virtually the same as building hundreds of tiny Helmholtz resonators next to each other, each with a slightly different depth. But a panel trap can't be built that way. The entire panel must resonate over the same depth of cavity. It is tuned to only one frequency, not broadband. The surface density of the panel and the cavity depth set the resonant frequency for the entire panel: You cannot make each part of it vibrate at different frequencies by angling it away form the wall: it acts as one entire unit, and all of it needs to be over the same depth.

This is something I haven't come to study in any great detail as yet (because I'm not quite at that point in my build yet) but as I also have slot wedges planned this throws me a curve. I understand and concur with your explanation and would like to point out that this highlights a certain ambiguity in John's Recording Manual which I perused for my design. The Manual is not wrong, as it also points this out in general, but fails to flag the important difference, leading the unwary to conclude that they can apply the given formula generally to any slat wall. The implications are that it would seem pointless to apply the given formula for slat specification on an angled slat wall and any variation of slat/slot width on an angled slat wall is purely aesthetic because of the broadband nature of the construction. The formula could only be calculated using an averaged depth, which is, in overview, perhaps pointless. Is this conclusion correct Stuart?

many similarities I must admit...

None at all, actually! :) Now THAT looks more like the response of a typical untreated small room! You can see the huge excursions, even on the smoothed graph: You have variations of +/- 15 dB over the entire range (peaking to -35 in places), and at least +/-10 dB even over most of the mid range! And that bass rise below 300 Hz is also typical of small rooms, with some obvious modal issues going on there. There also seem to be many reflections, and even some comb filtering going on. Huge differences, as compared to what the other graph was showing. Totally different. OK, but that's just the frequency domain: we also need to look at the time domain, to find out where your modal issues are, so that you can tune your slot wall and/or panel traps. So please also post the waterfall plot from REW. That will reveal the true details of how your room is reacting.

I must add that the first graph you saw (in Harbal) was in 1/3 of an octave resolution. I can make a jpeg of the same file in 1/12 if your curious...

It's not worth it: there is something seriously flawed with that setup. Nothing it tells you can be believed. Just go with what REW says.

I can now see I like bass a lot!

Yup! That has to be fixed, and urgently! Some of what you are seeing on that REW graph is modal issues with the room itself, but some of it seems to be a very poorly set crossover for your sub, or maybe very poor placement of the sub in the room. You need to set your crossover correctly. How is it set right now? What frequency, and how much boost do you have dialed in for it? The response of your speakers and room, as a system working together, has to be as flat as possible. Yes, most people like to hear more bass (human auditory system, human nature), but you cannot do that in a studio: the response must be flat, not distorted. The room and speakers must tell you the truth. If you have bass boost dialed in like that, your mixes will not translate well: they will be WEAK in the bass when played elsewhere. And with the rest of that REW curve, your mixes will also be rather "honky" when played elsewhere, since your room seems to be sucking out part of the mids. So first set up your system correctly for the flattest response you can get from the speakers alone (test at low levels, measure very close to the speakers), then re-do the REW test. EDITED TO ADD: A thought just occurred to me: your main monitors are soffit mounted: Did you roll of the bass response for them correctly? If so, at what frequency and how much did you roll it off)? That could be a big part of the issue.... ---

So basically, could you explain the construction and utility differences between a panel trap and a resonator? It sounds like I mixed them...

The panel trap (also called "membrane trap") is built just like the name implies: a single large panel (usually wood, such as plywood, MDF etc.) over a sealed cavity. It is a simple resonant system: mass (the panel) on a spring (the cavity). It resonates at a specific frequency. The resonant frequency of the device is set by the surface density (weight per square foot) of the panel, and the depth of the cavity. Adding absorption inside (as long as it does not touch the panel!) causes the device to work over a slightly wider range of frequencies, but at a slightly lower level. You can use a calculator or look up values on tables to find out what type of wood you need, how thick it should be, and how deep the cavity should be. You normally use a panel resonator where you room has a sharp peak in the response at a specific frequency, probably due to a modal issue. In that case, you also need to position the panel at the location where the pressure peak occurs for that specific mode. For example, if the problem is caused by the 2,0,0 mode, then putting the panel on the side walls would not be very useful! On the other hand, a slot resonator (or "slat resonator", if you prefer) is normally used as general broad-band treatment for a room. It is based on the principle of a Helmholtz resonator, and consists of a series of wooden slats over a sealed cavity. There are small gaps between the slats, and these GAPS are the active elements in the device (not the slats). Just like the panel trap, this is a resonant system: a mass on spring, vibrating. But different from a panel trap, here the "mass" is the air that is trapped in the "slots" between the wood "slats". That bunch of air, often referred to as a "slug" of air, is what vibrates. This is the same concept as blowing over the top of a Coke bottle to make it hum: the slug of air trapped in the neck of the bottle acts like a mass, and the air trapped inside the main body of the bottle acts like a spring: Just like the panel trap, together they resonate (vibrate) at a frequency set by the mass of the slug and the depth of the cavity. The difference here is that you can tune each slot to a different frequency! There is plenty of space on a slot wall for at least a dozen slots, and each one can be tuned to a different problematic frequency in the room, or they can be set to cover a general range of frequencies. You tune each slot by changing the dimensions of the gap and the dimensions of the slat on each side. But you can also angle the slats, so that one end is further away from the wall than the other, thus making the cavity deeper on one end and shallower on the other end: This also makes the device work across a wide range of frequencies, where the highest frequency is set by the small gap, and the lowest is set by the large gap. So you have TWO ways tuning a slot wall: change the dimensions of the slots (by changing the width and thickness of the slats, and the width of the gap between them), and also by angling the entire wall, to change the cavity depth. You don't have to angle it if you don't want to: If you have it parallel to the wall, then the entire length of each slot will be tuned to one single frequency, which might be exactly what you need if you have to treat a set of clearly defined frequencies in your room. Or if your problems are more general, then angle it to cover a broad range. So, you can angel the slot wall, but you cannot angle the panel trap. The entire panel acts as one single unit. You can't make different parts vibrate at different frequencies: the whole thing will only vibrate at one frequency, so angling it will do nothing useful, as it won't be tuned at all. In other words, it wont work. It will just be a pretty wooden decoration on the wall! Even worse, it will be a highly reflective pretty wooden decoration on the wall! The slot wall also has the advantage that it causes some diffusion, diffraction and refraction of the sound field, due to all those angled surfaces broken up by the gaps. And it also is filled with absorption, to broaden the response even further, but that also acts as pure absorption for frequencies to which none of the slots are tuned. So it has multiple things going on at once.

does a resonator have many ''boxes'' inside? if that is the case, the plan I downloaded was misnamed...

No. In both cases, there is just one large, sealed cavity behind the surface. That is what acts as the "spring". Both devices must be sealed airtight, or they won't work. The panel trap works the way it looks: one single big box. But the slot wall works differently: even though it is built as one single box, it acts as though it were actually made up of hundreds of smaller boxes, each one tuned to a slightly different frequency. If you really wanted to, you could go to the trouble of actually building all of those boxes, for a tiny increase in efficiency, but in reality you don't need to do that. Some people do suggest going part way, and putting a thin "wall" along the center line (lengthwise) of each slat, to seal it off from the next slat, and once again you can do that if you want, for a slight increase in efficiency. But once again, it's a lot of work for small returns. What you could do, is divide the wall into sections, separated from each other. For example, put all the slots tuned to low frequencies at the top and bottom of the wall, and the ones tune to mids in the middle (or vice versa), then put a couple of dividers between those sections. Again, you don't have to, but it might be worthwhile if you really want to. --- From looking at your graph, I'd definitely say that you need slot walls, and also a lot of bass trapping, and also treatment on whatever is causing those multiple major reflections. - Stuart -

The formula could only be calculated using an averaged depth, which is, in overview, perhaps pointless. Is this conclusion correct Stuart?

In general, yes. The ENTIRE WALL acts as a single broadband unit, overall, covering a wide range of frequencies. And you are right: it doesn't make sense to try to figure out the exact resonant frequency for each point along an angled slot, since the depth is infinitely variable. But what you can do, is to calculate the frequency for each end of the slot, and assume that the slot, as a whole, works for all frequencies in that range. Of course, it isn't quite that simple, as you point out! There is some interaction going on inside the cavity, and it is a very complex device, since it does a lot of things at once! But that gives you a general idea of what the wall will do: each slot covers a range of frequencies set by the dimensions of the air gap and by the deepest and shallowest depth, then the wall as a whole covers the entire range of frequencies handled by each slot. So if you were to try to graph it, it would give you a very broad Q overall. What got me for a long time here, is the concept that the only ting that sets the frequency for a given slot, is the cavity DEPTH behind it. It seems reasonable, logical and intuitive that it should depend on the entire VOLUME of the device, but as with lots of things in acoustics, it just ain't so! If you look at the simplified versions of the equations, what really mattes is only the depth of the cavity. That's why you don't really need dividers between slots, or along the slots. It would work more efficiently if you did, since the interactions between cells would be stopped, by in reality it just isn't necessary. So, conceptually, you can think of each slot being made up of a large number of individual Helmholtz resonators, each of which has a different depth and resonates at a different frequency. So when each part of the slot "sees" the frequency to which it is tuned, that part resonates and the rest doesn't. Kind of hard to get your head around that! It brings more questions than it answers! How much resonates? What happens at the point in the slug where the "resonating" part meets the "not resonating" part? What happens when there are multiple frequencies present, at adjacent parts of the slug? Etc. Mind-boggling! :!: :? You need to get into calculus and start integrating and differentiating to figure that stuff out, I reckon, and calculus never as my favorite subject! So I just try to look at the wall as a forest, arranged in neat roes and columns of different height, and forget about all those bothersome individual trees! :) - Stuart -

For those interested, here's the equation: f = 2160 x sqrt ( r / (( d x C x D ) * ( r + w ))) Where: f = resonant frequency (Hz) r = slot width w = slat width d = slat thickness C = mouth correction factor (1.2, normally) D = depth of the cavity. Notice something? No mention of volume of the cavity! The only factor in play there is the DEPTH of the cavity! :shock: surprising, but true (OK, for the purist, yes the volume does matter, and is involved, but not directly, as the equation shows...) ( By the way, the equation on the SAE web site has the WRONG version of the equation, which is often quoted and still being perpetuated around the internet! Do NOT use that version. see if you can spot the difference... :) ) Then again, instead of varying the slot dimensions and depth along each slot, you can get creative and vary the depth of each slot individually, by doing something like this: That is still a slot wall, and still works on the same principle! :) (Not sure where that came from originally, but I found if over at Gearslutz, here: http://www.gearslutz.com/board/attachme ... bild-1.jpg) - Stuart -

Thanks for help!

How is it set right now? What frequency, and how much boost do you have dialed in for it?

-my crossover is set to 80h with a boost of about -20db. -the sub is placed dead center on the ground between my speakers (usual placement for a sub) -I didn't roll of the bass of my BM5As since there are plugged directly in the sub for crossover. -I didn't boost/Eq on the BM5A in any ways, everything is set to flat on the back. Do you think I should Eq them? I'd like to sum up the panel trap/resonator please from what you explained: (thanks by the way) Panel trap: -A big sealed wood box on the wall with a complete sheet of wood on the surface. -Tune to a single frequency -With added absorption inside, the frequency range is broaden but efficiency lowered. -Cannot be angled Slat wall: -A big sealed wood box on the wall with many different sized slats on the surface -Tune to many frequencies depending on the slats and gaps size. -With added absorption, the frequency range is broaden, with added absorption for frequency not tuned by slots. -Can be angled So my question here: Is that right? Because if so, that correspond exactly what I built here (slat wall) so why can't I tune it? (I think you thought I wanted to built a panel trap because of my sketchup image, which does show single sheets of wood on the boxes...sorry, it was to long to draw many slats in sketchup so I didn't... I really planned to make a slat resonator.) Here is my complete results in REW P.S the is A LOT of bass trapping in this room...A LOT, maybe 10% of the room volume... Thanks for help! Tommy

So my question here: Is that right?

Yup!

Because if so, that correspond exactly what I built here (slat wall) so why can't I tune it?

Because that's not what you showed in the diagram (details matter!) and also is not what you described! "Its basically a sealed wood box filled with rockwool, then tuned by lumber on the front...right?" sounds like a panel trap to me! And the SketchUp certainly looks like one too. :)

sorry, it was to long to draw many slats in sketchup so I didn't... I really planned to make a slat resonator.)

:shock: I guess it would have been nice to make that clear on the diagram!

Here is my complete results in REW

Ummm... Your "waterfall before slats" and "spectrogram before slats" and "scope before slats" and "decay before slats" all seem to be rather blank! :shock: Are you sure you set those up correctly and generated the graphs correctly? They all look about right for total silence.... :) - Stuart -

Sorry about the panel trap thing, should have mentioned. I'm still new to REW..didn't see that generate button here are the other results. (I can't seem to get ''scope'' generate anything...)

for the really low stuff (below say 60hz) a VPR (or CBA / BCA) (Compound Baffle Absorber / Broadband Compact Absorber) panel would be a good choice. it is basically a 2mm steel sheet glued on a 100mm melamine foam (like Basotect G) backing. depending on the configuration (leaving the foam open on the edges increases the bandwidth) you can get good absorption down to 30hz on a unit which is only a bit over 100mm thick. mind you, a 1.2m x 2.4m unit will be close to 65Kg (100lbs) but it is pretty efficient and could be coupled with other absorption layers. placement across corners and mode points is the typically approach. some good info here: http://www.landtop.com/landtop/shfw/ztt ... 880847.pdf I'm working through some of the math and may have some SU components folks could use in the next few weeks.

Thanks Glenn, How do you calculate this thing? The calculator in this forum with the average depth of the boxes? Or is there much more to it? Tommy