Rational Acoustics



fiddledance
January 3rd, 2010, 04:47 PM
In Harry Brill's article on Mixonline on Smaart tips, he says the following: "You can find room resonances by using terminated pink noise while viewing the Spectrograph. Place your measurement mic in the room, and with the Spectrograph display active, play pink noise into the room and set your levels. Mute the pink noise and note which frequencies continue to ring."

Great tip, but could anyone give more specific advice as to where to place the measurement mic? My situations are almost always two stacks of tops on poles on top of subs, not very complicated.

Thanks.

Arthur Skudra
January 3rd, 2010, 06:24 PM
Well, there isn't any specific point in a room that will be representative of all room modes that exist in that space, so you should move the microphone randomly around and note all the peaks and nulls between the positions. Easier to do this with a multichannel measurement setup.

With the upcoming Smaart version 7, there will be some really cool tools in the impulse response mode that will show resonances at a mic position based on the impulse response that you captured. You can also save and reload impulse responses for later analysis! :cool:

If you're doing this live, you can use a swept log sine wave, and see room modes with greater detail and precision, depending on how slow the sweep rate. I also use warble tracks for this purpose, there are a couple you can download as part of the Bink Audio Test CD package that you can find here:
http://binkster.net/extras.shtml#cd

The warble tracks are my "go to" tool for subwoofer to LF system optimization, after I have done the necessary alignment using the phase trace.

Harry Brill Jr.
January 4th, 2010, 04:32 AM
I wasn't very specific for a good reason. Arthur's post shows his interest in using this with LF. I use it to look at reverb trails, and resonances in the vocal range. What I call "hanging notes", or sounds that keep on playing after the speaker has been muted. If you simply want to look at reverb trails, you can slam a case shut. None of this is particularly repeatable or accurate, but is instead a tool you can use immediately to learn more about the space.
This is a tool that is limited only by your imagination. As for mic placement, I'd put it in the position I wanted to measure. You are measuring something in the room, not the speaker system. Multiple mics is definitely useful particularly after we can view them all at once (when computer power catches up). This is one of the most useful tools for visualizing what you hear.

If you want accurate and repeatable the IR is more appropriate for documentation.

fiddledance
January 4th, 2010, 03:27 PM
Thanks, guys, for the replies. Yes, I don't want to measure the speaker system, since it's the same rig for each gig and has been measured outside. I just want to understand better how the room is affecting the system, since the room changes, not the system. Version 7 is definitely on my "get" list.

James Woods
January 14th, 2010, 02:58 PM
This must be my record for doing two posts in one session.:p

I do things in a different manner. I usually run low frequency sine waves through the system (40, 63, 80, 100, 125) and then walk around the room to see where the peaks and nulls are. They are almost never in the same place for the different frequencies. I usually do just one side of the PA at a time, and then both together.

Since what we are hearing are room modes, they will change with temperature and humidity. They also do not respond to EQ well either.

You will also get some different results depending on where the subs are in the room. Most of the old texts want you to put the signal into a corner.... This is not always possible.

So after I have seen (heard) the room modes, I usually pay a LOT of attention to where the mixing console and specifically where the mix engineer is, so that I can tell him before hand the peaks or dips he may hear at his mix position.

I will try Harry´s and Arthurs suggestions as soon as I can.

PrestonSoper
February 6th, 2010, 02:16 PM
Well, there isn't any specific point in a room that will be representative of all room modes that exist in that space, so you should move the microphone randomly around and note all the peaks and nulls between the positions. Easier to do this with a multichannel measurement setup.

With the upcoming Smaart version 7, there will be some really cool tools in the impulse response mode that will show resonances at a mic position based on the impulse response that you captured. You can also save and reload impulse responses for later analysis! :cool:

If you're doing this live, you can use a swept log sine wave, and see room modes with greater detail and precision, depending on how slow the sweep rate. I also use warble tracks for this purpose, there are a couple you can download as part of the Bink Audio Test CD package that you can find here:
http://binkster.net/extras.shtml#cd

The warble tracks are my "go to" tool for subwoofer to LF system optimization, after I have done the necessary alignment using the phase trace.
Could you elaborate a bit more on your process of using warble tones after doing your phase alignment of subs/lows

Arthur Skudra
February 6th, 2010, 11:40 PM
Could you elaborate a bit more on your process of using warble tones after doing your phase alignment of subs/lowsKinda hard without pictures, I'll try to make a few screen snapshots the next time I'm in a room with subs.

Essentially I'll play the warble track or low frequency log swept sine wave into the subwoofer system, and in spectrograph, I'll see the peaks and nulls in the subwoofer's response with the room at that particular microphone position. It does involve tweaking the input level of the microphone a bit so that these peaks and nulls stand out with the color/intensity scheme. You will also notice some "smearing" in the spectrograph much like a feedback squeal at certain resonant notes. Be sure you repeat this measurement at several mic locations so you can see the common frequencies where problems may be experienced. Don't carve too much into the sub's response, usually I find 1-3 well placed parametric eq's do the trick. Your goal is to get the swept sine to be as smooth in intensity through the pass band as possible (apart from any shaped response you may desire to emphasize the kick drum or bass guitar).

Calvert Dayton
February 11th, 2010, 05:15 PM
You guys just reinvented reverb time analysis (by frequency) using both terminated pink noise and impulse stimulii. (Slamming a case lid shut = exciting the reverberant field with an impulsive stimulus, recording what happens when you do that gives you an impulse response). Nicely done. In practice both techniques should give you identical reverb times, the difference is that the impulse response can also give you a little extra info in addition to RT, such as early decay time (if you look a little closer).

It sounded like fiddledance was maybe doing corporate gigs so I wanted to mention that for rectangular rooms, you can also calculate the room modes (resonant frequencies). Specific Reverberation times are still going to depend on such factors as reflective/absorptive characteristics of major reflective surfaces and refraction from objects in the room, but you can still get an idea of where to look just based on the dimensions of the space. Somewhere I should have a spreadsheet for that. In more complex, non-rectangular spaces the resonant frequencies are likely to vary by location, so generally speaking you basically want to measure in places where you care how it sounds.