Analysis/measurement of Room modes

**studiohog** · **10-20-09, 07:33 PM**

Hello there

I have been "lurking" on these forums for several weeks now and have been absorbing alot of the interesting and informative acoustics related posts that have been made.

If its ok i would like to ask a question or two myself?

I am working on the analysis and room correction of a smallish room (approx 150 cubic feet )which has a ceiling which complicates matters somewhat in terms of mathematically predicting room modes. The ceiling slopes down away from the listening position at an angle of 15 degrees. As any equations/room mode calculators i have seen assume parallel surfaces this poses a significant problem in terms of predicting modal resonances which may cause problems,well, the floor to ceiling axial modes at least.

Much of the research and opinions i have been exposed to indicates that predicting mathematical ideals is of questionable value anyway, so i feel its time to get on with some meat and potatoes analysis of the room using software and a measurement mic. However i can find little literature on the actual procedures and methods which acousticians use to analyse for room modes. Unfortunately the project i am undertaking is outside of my University lecturers expertise (my degree is only partly based in acoustics and our only acoustician has recently retired) so i am pretty much on my own on this one.

Which brings me on to my question. Would anyone like to share some knowledge with me and give me some pointers as to how i can go about identifying troublesome room modes by way of actual measurements and data interpretation?

I have REW and a couple of other well known pieces of acoustics software and access to an earthworks M50 measurement mic,clean preamps etc.

In addition to having my head perpetually stuck in acoustics books and research papers i have searched these forums with not much joy so any help or pointers would be greatly appreciated.

Kind regards and many thanks

Chris

**brucek** · **10-20-09, 08:02 PM**

Quote:

Which brings me on to my question. Would anyone like to share some knowledge with me and give me some pointers as to how i can go about identifying troublesome room modes by way of actual measurements and data interpretation?

Not sure I understand the question. Do you need help with using REW? It will certainly identify room modes in the form of peaks in the frequency response measurements. REW can identify those peaks and recommend filters to use in an sub equalizer.

brucek

**bpape** · **10-20-09, 08:08 PM**

I think the point is that not all peaks and dips have to do with room modes. Some are related to SBIR, some are related to cancellations based on seating distance, some don't show at all due to a peak and dip being in the same place at the same frequency.

The best you can do is to predict the modes for the length and width and mark those in your measurements. If you want to do research on SBIR, there's a lot of it out there and you can see how that works. An easier way is to simply move the speakers forward and backward and see what changes in the plots. Same with seating position.

Bryan

**SierraMikeBravo** · **10-21-09, 03:42 AM**

If you really want to know what the actual modes are...place a subwoofer in one corner of the room and the mic at the listening position. Placing the sub in the corner will excite all the modes for which the subwoofer is capable. However...your analysis will depend on the position of the mic. The way to eliminate this problem and get an actual mapping of the room is to use a multiplexer or spatially averaging several locations (it's a lot easier to move the mic rather than speakers). Calculating room modes is very accurate. All the testing will do will be to verify your calculations. Modes depend on the dimensions of room...nothing else. Remember though, there are more than just axial modes. There are tangential and oblique as well, and you'll need to calculate those to determine whether or not what you are seeing is an actual mode, or if the mode has enough modal support by surrounding modes (whether they be axial, tangential or oblique) if they are not as prominent in the analysis. In addition, if you are using this for actual research, I would suggest using a program a little more robust such as EASERA.

**studiohog** · **10-21-09, 06:35 AM**

Thanks for all of your replies. Apologies are in order as on reading back my post this morning (i'm across the pond in the UK) it was'nt worded very well. Was late last night when i posted :/ sorry for the ambiguity.

To clarify, the question was more to do with the actual technique in terms of the actual methodology involved, i.e. where to place the sound source and where to place the microphone etc. So i take it that as all modes whether Axial, Tangential or oblique terminate in the corners of a room then by placing the subwoofer in a dihedral corner your ensuring the sub is exciting as many LF modes as possible?

Do you mind if i ask Roughly how many times should i iterate the measurement in terms of changing the measurement mics position to get a good bunch of results? And what is the easiest or perhaps i should say preferred way to collate the data for interpretation, for example is it easier to superimpose the multiple measurements on a waterfall plot for comparison?

Again i feel a bit cheeky asking all these questions but i really am having trouble finding analysis methods that are written down apart from measuring RT60's,basic wave propogations or speech intelligibility tests.

If anyone can point me to anywhere where room acoustics measuring procedures/methods are available to read then that would be ideal. It appears to be a bit of a 'black art' or i've just been looking in the wrong places.

Again thanks a million for your posts.

Chris

P.s i have access to EASERA via the 30 day demo of the full version which is meant to be fully functional.Yet to try it however. Still reading the manual

**SierraMikeBravo** · **10-21-09, 11:08 AM**

Tangential and Oblique modes may or may not terminate at the point source. You'll have to do some reading as to what they are before you decide. Regarding spatial averages, do what you feel is necessary. Good statistical averages use 30+. That might be a bit excessive but you are doing this for research are you not? I would do at least 4. I am not sure if REW will do spatial averaging. I don't use it. One program that does spatially average multiple locations is True RTA. If you can't afford EASERA, I would look into this program, and get the highest resolution you can ($99 for 1/24 octave). I use a standard SPL/freq plot when doing this type of analysis. The only use I have found for waterfall plots (other than it looks cool) is to determine modal ringing. However, that is often a function of amplitude anyway. You'll also want to make sure you have the calibration file for that mic. Mics tend to drift, and this is often more prominent in the lower frequencies.

**glaufman** · **10-21-09, 01:48 PM**

I've found the best way to learn is to dive in... if you want to get your feet wet quickly...
Set up REW, take scans in 9 different positions in the room, overlay them all on the same plot, and post it here, see what people come up with... (I picked 9 because that's how many REW will let you overlay in one shot)...

**studiohog** · **10-21-09, 04:35 PM**

Hello again.

Brilliant stuff, thanks ! Got plenty to get my teeth into over the coming months! I've dove right in at the deep end so should be a memorable and hopefully enlightening journey

30 + measurements is doable, all in the name of science
after all. The more measurements the more valid the averaging data i guess. Better not forget my earplugs next week eh?

Room acoustics is such a massive subject i've just about read F.Alton.Everests master handbook of acoustics from cover to cover over the past 5 weeks and i've got several AES papers on order.

The one downfall with reading books is that you cant ask a book questions. Its great to find a forum like this where theres plenty of folks who are willing to share knowledge.

I'l be visiting here again and next time i'l have some test data to show. I plan to carry out the frequency response test (spatially averaged), impulse response and RT60 on the room in question next week. One last question before i go. Does anyone know of a mathematical equation which accounts for a roof slope in predicting modal resonances. Past using very expensive acoustic modelling software such as LMS sysnoise or Virtual labs for BEM i've yet to find a solution. It would be nice to be able to compare mathematical predictions with the actual test data.

Once again thankyou so much for your posts, was stuck in a bit of a rut but i think i can claw my way out now.

Cheers

Chris

**SierraMikeBravo** · **10-21-09, 09:58 PM**

Unfortunately, no, there is no shortcut to determining modal mapping in a room. All simplistic models depend on the shape of the room being rectangular. You can use CARA to determine, with pretty reasonable accuracy, odd shaped rooms. But be forewarned, in order to determine this properly, each surface needs to be counted since it uses the ray tracing method. The more surfaces, the longer the calculation. With 6 surfaces, assuming nothing else in the room and not accounting for flexible materials, it will take on the order of a month for it to calculate an accurate response...although you may want to reduce the frequency range and just account for the subwoofer which should reduce calculation time. BEM method is pretty good, but you have to program it in MATLAB usually, and you do sacrifice some accuracy for calcualtion efficiency since the BEM method is based on definite integration. FEM is the most accurate and preferred professional method, but the second order differential equation takes as long as the ray tracing method. There are other ways that have been figured out, and that do save tremendous amounts of calculation time, but those are closely guarded secrets as that is what we get paid for.

In addition...you want spatial averaging, not position overlay for various locations. Also, if you are going to do that much averaging, better use EASERA.

BTW, what exactly are you doing research on?

**studiohog** · **10-22-09, 04:58 PM**

A month ??!! wow, thats some serious number crunching there, still, plenty of time to make a coffee while it whirrs away.

To answer your question i'm an undergrad in my final year of BSc in Audio Music Tech. Which is partly acoustics based, partly recording engineering based with some audio electronics thrown in for good measure.

In the final year we are let off the leash a bit and are allowed to carry out a final year project of our own choice which constitutes a fair lump of our final grade. I've elected to do something acoustics based as its an area i've always found fascinating, kind of a natural progression for me from studio work. The science of sound

My current research is doing the ground work for my final year project, my proposal is to analyse the room in question, identify any acoustic problems using subjective and objective methods, devise a remedial room treatment plan, carry out the room treatment and then analyse again. The room is to be used as a reference critical listening room. Not strictly a home theater but it is in someones house

I've got my work cut out but i've got 22 weeks to do it in. Just coming up to the end of week 5 so the intensive research stage is coming to a close now and the meat and potatoes of it has got to begin. Our acoustics lecturer has recently retired (nice timing) . So i'm on my own for this one, best way to learn sometimes. Just a drag when you hit a rut. Cheers for giving me me some help. Much appreciated. And if you ever feel like sharing some of those tricks of the trade,especially where sloping ceilings are concerned, i promise i wont come across the Atlantic and steal your work

Nice one and take care.

Chris the budding acoustician. Heheh

**studiohog** · **10-23-09, 02:52 PM**

Back with another thought.

Been chewing over spatial averaging.

With Spatial averaging being the mean average of values in multiple points of space, and in this context I'm taking an educated guess that as we're dealing with 3 dimensions within a room the average we're concerned with is a average within a volume, possibly not though.

Then if i iterate the FR test say 30 times for arguments sake then enter the values into MATLAB (for plotting a nice graph) or a spreadsheet then tell MATLAB to run a mean average based on those 30 FR values then the resulting outputted graph plot will be the spatially averaged FR of the room.

Does that sound feasible if you dont have software that will do the donkey work for you ?

Regards

Chris

**SierraMikeBravo** · **10-24-09, 09:07 PM**

I guess you could do it that way provided you are adding and averaging the SPL values for each frequency say up to 300Hz. You could also do it in excel. This again depends on your resolution. The higher the resolution of the RTA the more accurate your modal resolution will be. It's easier though if the program can do it for you. Otherwise...if you go to two decimals on the frequency, that will be a lot of data to enter manually.

Advertisement