Room Treatment

[charic]

1 hour ago, Jus Lukin said:

Always worth pointing out that room treatment and soundproofing are very different jobs!

There are many principles which cross over between the two, but if you're going for a simple approach, you will achieve nothing as far as containing the sound.

To make a cartoonishly simple generalisation, with room treatment, every little helps- with soundproofing nothing helps until you have everything!

Oh yeah absolutely, but I already have a pretty good idea on the soundproofing aspect

The room treatment is to make it so that you can actually practice in the room with it being a wall of mush!

The garage is currently made of breeze blocks within bricks on 3 walls and it's a detached.  So bricking up the inside of the garage door will cover the majority of noise issues I expect! (although ventilation will need factoring)

Adding some damping around the room\ceiling will help alleviate reflections and I expect a few basstraps will be required, the floor is likely to be a bit more problematic I think

[charic]

34 minutes ago, WinterMute said:

Been absent from this thread for a while, been building studios for other people, just popped in to say it is entirely possible to increase isolation in any room, but it's likely to be reasonably expensive and not especially easy to do unless you've a good builder or a VERY handy DIY'er.

I did mine with the help of a very decent local builder, but I had to lay out exactly what I wanted in terms of the construction and materials, it cost about £8K to fully proof the garage to the point where I can work at proper levels into the early hours without disturbing the good lady wife or the boys.

As pointed out above, isolation is not acoustic treatment, and you'll need to do both if you want the room to be useful with other people in the vicinity.

 

Skol: your room looks grand mate.

I'm not looking to make this "completely" soundproof, electronic drums into a PA is the kind of volume level I have in mind

[WinterMute]

22 minutes ago, charic said:

I'm not looking to make this "completely" soundproof, electronic drums into a PA is the kind of volume level I have in mind

Technically, that's called "loud"...

Getting a good set of seals on the doors, including a raised sill, and any windows will serve you well, but walling up the door won't do much unless you can seal the roof in some way. 

Mass is definitely your friend however, as is distance.

[charic]

36 minutes ago, WinterMute said:

Technically, that's called "loud"...

Getting a good set of seals on the doors, including a raised sill, and any windows will serve you well, but walling up the door won't do much unless you can seal the roof in some way. 

Mass is definitely your friend however, as is distance.

There are no windows and the roof is under the house

[WinterMute]

9 minutes ago, charic said:

There are no windows and the roof is under the house

Is the roof connected to the house? You mentioned the building is detached. The door opens to the outside?

Audio transference particularly in the LF is a big problem where there is physical connection, my room shares a common wall with the house, so isolating the interior wall was essential.

 

[SubsonicSimpleton]

As far as garage conversions go, this band built an isolation room using inside out walls to tame the internal acoustics, definately worth a watch, although it is worth noting that their use of high mass vinyl was probably an unnecessary expense as an extra layer of plasterboard would have been cheaper. The interesting thing about this approach is that although it isn't meaningfully portable, because it is built in a modular fashion, it could be disassembled and moved at a later date should the need arise.

There is a real world room within a room build series that starts here,

and another garage conversion

There are plenty of other examples, but many of them fail to follow any sort of best practice and end up being exercises in wasting time/money because they overlook one or more important factors.

 

[Skol303]

11 hours ago, WinterMute said:

Skol: your room looks grand mate.

Cheers!

It’s been a labour of love; very educational and strangely enjoyable. I’m still chasing some minor improvements... but the diminishing returns are now so steep that it’s becoming difficult to measure let alone hear the benefits. That’s what I like about acoustic treatment: there’s an end point in that you either run out of money or space! It’s a finite form of GAS  

Good points above about sound proofing vis-a-vis treatment. I haven’t attempted a soundproofing project myself so can’t comment from experience, but I know it’s an endeavour that doesn’t reward half measures.

@charic don’t let that put you off though mate! If you’re a handy DIY-er then you can make a very decent job of it with sufficient elbow grease (and budget). If you’re interested I highly recommend this book:

Home Recording Studio: Build It Like the Pros

Quite technical in places, but nothing you can’t get your head around! Covers everything you need to know about building a home studio from scratch. The author, Rob Gervais, is also quite active on gearslutz.com and commonly responds to questions. I found it overkill for my purposes - mine wasn’t a build project as I wasn’t aiming for soundproofing, but it was a good read nonetheless. If I can find my copy you’re welcome to have it... I’ll have a root around the bookshelf and see if it turns up.

Edited August 15, 2018 by Skol303

[charic]

On 15/08/2018 at 12:23, WinterMute said:

Is the roof connected to the house? You mentioned the building is detached. The door opens to the outside?

Audio transference particularly in the LF is a big problem where there is physical connection, my room shares a common wall with the house, so isolating the interior wall was essential.

 

Sorry, badly phrased.. the garage is part of the house.  The roof is solid brick underneath some of the upstairs of the house (the house is a detached and the side with the garage doesn't have any neighbours)

[charic]

Thanks for all the links and material!  Once I can get into the garage I'll start a thread

[Skol303]

Ok, time for some analysis of Pete’s results, posted earlier.

Firstly, here’s a rough plan of Pete’s room, which I’m guessing is around 3m square:

This is a classic ‘small room’ set up that’s very typical of most home studios; so it makes a great case study for anyone else starting out with acoustics. As such, I’ve tried to include some general advice along with the detail, because much of what’s being discussed is applicable to any small space… so thanks to Pete for providing the data and being our guinea pig.

It’s a long read, but it’s a big subject and as the kids would say: “go hard or go home”  

Right, that’s enough preamble. Let’s get to work…

[Skol303]

Frequency Response

Left speaker red; right speaker green.

• Very typical of a small, untreated room… but not too bad at all. I’ve seen a helluva lot worse! So that’s a good start.
• There’s an obvious difference between the left and right speakers around 90-100Hz, notably a big null (dip) in the left speaker and a peak in the right speaker. This is almost certainly caused by the left speaker being close to the adjoining side wall. Our ears - or more correctly our brains - tend to ‘sum’ low frequencies to mono below around 80Hz (which is why many subwoofers crossover at this frequency). Above 80Hz we’re able to more accurately pinpoint the source of low frequency sounds, so I’m guessing that the inconsistency in Pete’s room around 90-100Hz is audible, albeit perhaps subtly so. Either way, we definitely need to fix it.
• In the low end below 200Hz, the frequency response varies by around +/-14 dB in the left speaker (i.e. the difference between the loudest peak and lowest dip is 28 dB) and +/-11 dB in the right speaker. A variance of +/-10 dB in the low end at 1/24 smoothing is generally considered ‘acceptable’ for a home studio, so we’re already close to a workable result. Like I said a good start, especially if we can get the left speaker response looking more like the right speaker.
• Having small speakers is a real advantage in a room like this. The frequency response here drops off around 50Hz, which isn’t ideal when it comes mixing the low end, but it does greatly reduce some of the common low frequency problems in small rooms. Using headphones as a second reference is a way of helping to compensate for the lack of ‘oomph’.
• There’s a weird dip in the high end commencing around 8-9kHz. Never seen that before, but my guess is that it has to be a Boundary Interference Reflection (BIR) - i.e. sound waves bouncing off a hard surface and cancelling out where peaks and dips of the same magnitude and frequency meet each other. You can see evidence of these boundary reflections, also known as comb filtering, throughout the frequency measurements - the jagged little peaks and troughs in the mid and high frequencies (usually more pronounced in the highs, because high frequencies are more susceptible to reflecting off hard surfaces). The fact that this weird null appears very similar for both speakers means that it’s not just the adjacent wall that’s causing the problem. So it may be something to do with the irregular shape of the roof space.

Edited August 17, 2018 by Skol303

[Skol303]

Impulse Response

Impulse response is easy to overlook when you’re starting out learning acoustics, but it’s a very useful measurement. The peaks in the response represent reflections and can be used to work out the location of boundaries that need acoustic treatment. For example, in the screenshot here, the first big spike is caused by a reflection off a surface that is located [4.38 ft + the distance from the speaker] away from the microphone. Hours of fun to be had with that! (which I’ll explain another time…).

In a ‘good’ response, you’d want to see no or very few peaks above 20dB (that is confusingly, between 0 and -20dB on the graph) within the first 20ms. On Pete’s measurement, we see a lot of reflections in this region… which is perfectly normal for an untreated room. And even in a well treated room, it’s common to find spikes caused by furniture and other surfaces.

For the time being, there’s no point getting into the detail here until we’ve found the optimum listening position. So I’ll park it and move on.

[Skol303]

Reverberation Time

The classic ‘RT’ measurements (RT60/30/20) are designed for concert halls and don’t apply with any accuracy to small rooms. However, REW has its own clever interpretation (called Topt) which is designed to give a better representation of reverberation in small spaces.

Industry standards (such as EBU Tech 3276) suggest a reverb measurement of 0.2s rising to 0.4s in the lower frequencies, so Pete’s room is already performing well in terms of reverb time. Happy days. As mentioned above, this is almost certainly helped by using smaller speakers.

[Skol303]

Waterfall Graph (decay)

LEFT SPEAKER:

RIGHT SPEAKER:

This shows how different frequencies decay over time; and it’s just as important as the frequency response. At this stage we’re really just concerned with the low end below 200Hz… and what we’re looking for is ‘modal ringing’ - those elongated spurs that stick out from the main body of the graph, showing frequencies that take longer to decay than others. These are caused by room modes - aka standing waves - which are determined by the size and shape of the room. Modal ringing is inevitable in every room, but too much of it can cause big problems when mixing (e.g. a ‘boomy’ low end that makes it difficult to discern between different bass notes). Generally speaking, you want the decay times to be as even as possible across the frequency range, mindful that some irregularities in the low end are inevitable.

In Pete’s graphs, we can see clear evidence of modal ringing around 30Hz and 60Hz (seen in the measurements from both speakers), as well as other frequencies. The 30Hz ringing is especially gnarly, given that the speakers don’t seem to be producing much output at the very low end. So we definitely have a strong room mode acting at that frequency… and I'm guessing it's due to the room dimensions being of equal lengths (square rooms suffer from the worst modal problems). If so, this is going to be difficult to remedy without installing truckloads of acoustic treatment. But... 30Hz is a frequency that’s more ‘felt’ than heard, so it’s nowhere near as problematic as modal ringing at low end frequencies above this range.

Other than that, both graphs look good in terms of decay times. No cause for concern, based on what I’m seeing here. The ‘Grand Canyon’ in the frequency response of the left speaker is very evident however. Definitely need to fix that, speaking of which…

Edited August 17, 2018 by Skol303

[Skol303]

Recommendations

Here’s what I’d suggest as first steps:

1. Make sure that all future acoustic measurements are taken at a higher dB level. The current batch are around 56-57dB which is way too low… bearing in mind that the typical noise floor of a quiet room is around 30-35dB. So we’re probably losing some of the detail in these measurements. Shoot for around 70-80dB from now on. Wear earplugs if necessary (or better still, vacate the room!).
2. Before we start planning any acoustic treatment, we need to find the optimum listening position for this room. That is: the best possible location for the speakers, desk and chair. Acoustics favour symmetry and so the current set up (with the listening position in a corner) is not ideal. My suggestion would be to swap the position of the desk with that of the sofa: i.e. move the speakers/desk to the rear wall (where sofa currently is) and move the sofa into corner where your listening position currently is. The line to take with your partner is that the sofa will be nearer the window which is better for fresh air and feng shui 
3. Position the desk at the centre point of the rear wall, push your speakers as close as possible to the wall (within a few millimetres) and take another set of REW measurements. Also take a set of measurements with the door open - my guess is that may help tackle the modal ringing around 30Hz (free bass trapping!).
4. Keep repositioning the speakers and re-measuring in REW until you find what seems to be the ‘sweet spot’. Given that your reverberation and decay times are already very workable, what you need to focus on here is finding symmetry between the frequency response of your left and right speakers. Keep moving and measuring until the two graphs appear the same (broadly speaking).

Finding the right listening position is half the battle with acoustics… if not most of the battle. Once you’ve got your listening position sorted, post the new REW results here (and email me with a copy of the .mdat file). We can then look into what acoustic treatment is going to be most appropriate and practical, mindful that it’s a shared space.

[funkle]

@Skol303 absolutely stellar!!

For everyone else’s info I’m still to measure the room fully and send Paul the measurements. But the detail of the analysis is just incredible. And even a possible solution in regards to listening spot etc!

The high end dip is I assume because of the extremely odd shape of the room. It’s  more rectangular than it is square, but I actually only have one full height wall; all the others have various degrees of attic roof sloping in. So I’ll go with the analysis that the shape of the room is doing something weird to the high end. 

Reverb time is probably good because it’s fully carpeted. 

The monitors I bought knowing they had zip from 50Hz on down. It seemed a good idea because I had heard bass mixing was problematic in small rooms. I got lucky though because I was pretty ignorant as to the details which are becoming clearer over time. I bought the headphones I have specifically to be able to monitor the bass; the Equator D5s I thought would be clearer in the mid and high end. 

I’ll have to see if I can negotiate moving stuff around a little with the good lady. Then I can re-measure, and aim for a higher SPL to help out with better analysis. 

[Skol303]

Update: I’ve now received a sketch of the room showing its dimensions and the sloping areas of the roof space. Very useful! It’s slightly larger than I’d guessed, which is a good thing.

Room Modes
Here’s the pattern of room modes (standing waves) in Pete’s room, based on its dimensions (length/width/height). Tall, red lines indicate particularly strong room modes; and we can see there are strong modal frequencies at 36Hz and 55Hz, which coincide with the slower decay times seen on waterfall graphs above. So at least we know where the biggest low end problems can be found.

Both of these low frequency room modes are treatable, but the one at 36Hz is going to be difficult to reduce without installing truckloads of acoustic fibre or pressure-based bass traps (expensive to buy and tricky to DIY). Although from what I’m seeing, I don’t think we need to worry about it.

Curiously, the room dimensions fall within the ‘Bolt Area’ (see screenshot bottom-right), which means that the room modes should be quite evenly distributed in this space. That’s a good thing; but bear in mind that all of these estimates assume the room has vertical walls… so the angles in the ceiling might be sending some of these predications skew-whiff.

Moving the Listening Position
As mentioned above, I think the priority here should be fixing the lack of symmetry between the left and right speaker responses (as the rest of the room seems generally ok as a starting point). And to achieve that, I think we need to find better symmetry in where the listening position is located in the room  (i.e. the desk and chair). I’m fairly sure that having it in the corner is what’s causing the big dip in the left speaker around 100Hz (i.e. a cancellation null due to the proximity of the nearby wall).

Theory tells us that it’s usually best for speakers to fire down the longest dimension of the room, which would mean moving the listening position to one of the shorter walls (the one with the "wife's desk", or the one with the bookcase). But theory isn’t always right ...and I’m wondering whether the acoustics might actually fair better with the listening position located where the sofa is currently - especially if the door is left open.

These decisions are ultimately Pete’s to make and boil down to his own priorities for the room. All things considered, my recommendation remains unchanged from above: i.e. try moving the listing position to the opposite wall (where the sofa  currently is) and measure here twice using REW: once with the door closed and again with it open to allow comparison. My guess is that will help to fix the symmetry issues between the two speaker responses… and with the door open I'd expect to see an improvement in that 30Hz mode (e.g. faster decay time). But moving the listening position here may also cause a whole new set of problems! Acoustics tends to be full of frustrating compromises 🙂

So yeah. Try that new listening position, take some new measurements and we’ll then revaluate before looking at options for acoustic treatment.

Edited August 19, 2018 by Skol303

[funkle]

I’m on it. I can swap the DAW desk and my sofa around and re-measure. 

There is a huge amount of bass that I can audibly hear being bumped up in the corner where my wife’s desk is. If I set there when my stuff is playing from my PC, it just has this huuuuuuge low end. One of the reasons I avoided having my desk there in the first place. Although I now understand it is probably better acoustically to have the desk there (along the shorter wall than the longer wall), I think in this room it might actually be a worse option. 

I’ll report back when I have swapped stuff around. 

Thansk Paul @Skol303 for all the hard work!!!

[Skol303]

^ You're welcome.

I'm very much an amateur when it comes to acoustics, but I do get an odd sense of enjoyment out of it. Sort of my equivalent of doing a good crossword 😉

[funkle]

Some graphs coming.

I swapped the studio desk to where the sofa was. I got the monitors within 7-8 cm of the centre points of being equally positioned from either wall. I couldn't get them exactly where I wanted because I'd end up squashing into the space my wife is using in front of her desk to sit. So....

Here's the updated frequency response graph.

L speaker is red, R is in green.

I did measurements with the study door open and closed. The ones with the door open are slightly better in the bottom end frequency response. 

This looks a lot more workable. Amazing what simply moving position in the room can do. My bass response down to 50Hz is much better and there's much less difference between L and R speakers. I still have some nulls affecting the L speaker more than the R around 200Hz though...and the reduction in the high end above 9k. NOt sure what I can do about those. I'll have to look at room treatments next I guess. Depends on the good old law of diminishing returns...

 

[funkle]

 

Here's reverb time. I'll probably have to leave @Skol303 to comment....not sure this looks hugely different from before.

Edited August 21, 2018 by funkle

[funkle]

Here's the waterfall plots in the new position.

L side:

R side:

 

I'd say I still have a healthy amount of low end ringing going on; it extends even up into the mids and up to 1k. Ish

That looks like I need a healthy amount of bass trapping. I think. Paul can comment....

 

 

[funkle]

I'm going to need Paul to do the 'Impulse Response' graph. I can't quite figure out how to get it looking right, or indeed how to interpret it. Apart from the fact that apparently it tells us where room reflections might be.

[Skol303]

Well done Pete, a great improvement!

Frequency response is looking good. Very good for an untreated room. There’s still the weird ‘scoop’ starting at 9kHz which has me puzzled… but not overly concerned (some thoughts on that in my email).

Reverb time looks fine. Notice the sudden jump in the reverb time around 70Hz in comparison to your previous measurements. That’s almost certainly due to the new measurements being taken at a higher SPL level, which always shows up details you don’t see with quieter measurements. Again, no cause for concern here. Acoustic treatment will help.

Decay times are actually fine - it’s just that your waterfall graphs are showing more of the ‘Y’ axis than they need to! Bear in mind that the background noise in the room is probably around 30dB, so much of what you’re seeing in your graphs that looks horrendous is actually just inaudible guff. Here’s a more realistic interpretation… note that the ringing at 50Hz is perfectly normal; lower frequencies naturally have longer decay times and what we're seeing here isn't problematic.

I’ve also checked the impulse response and it’s very similar to the original position - i.e. showing lots of early reflections that acoustic treatment will improve on. Again, nothing that leaps out as being troubling.

Off to a great start! Keep us posted only your next steps… 👍

[Skol303]

A quick shot showing the frequency response of Pete’s Room in blue (currently untreated) and my own room in red (heavily treated… like a padded cell) for the sake of comparison.

What I’m trying to illustrate here is that although there are obvious differences, you can achieve very workable acoustics just by getting your listening position right. It’s by far the bulk of the battle.

The region where you really need acoustic treatment to make any kind of difference is from 100Hz and down; and it's an exponential curve with steep diminishing returns. The lower you go, the more time/ money/ effort is required to make any kind of difference (those very low frequencies are burly and brutal!). Getting things under control down to 30Hz is certainly where I spent the most toil.

Throughout the mids, however, both rooms respond very similarly. This is why thin layers of acoustic foam and hanging carpet on the walls, etc, are next to useless: they just dampen the mid and high frequencies (that typically require least treatment, as we can see here) and do nothing to tackle the low end (where the biggest peaks and dips are always found).

Edited August 22, 2018 by Skol303