Finding Suckout or Room Null

[jmathers]

Is there some specific test that you can perform to find a suckout or room null? I have heard that you can use something like the Stereophile test CD to do this. How?

Jeff

 

[whoaru99]

How serious do you want to get?

For somewhere between $50 and $200 you could get set up with Room EQ Wizard and do it pretty much all automatically. The cost window depends on what you have already and what you need. If your soundcard is full duplex, which most are these days AFAIK, a few cables and a measurement mic like Behringer ECM8000 (~$50) is the biggest expense. A SPL meter is handy too if you want to get more absolute measurements but not necessary for relative measurements. Some SPL meters, like some of the Radio Shack ones, can also be used for the measurement mic.

 

[jmathers]

How serious do you want to get?

For somewhere between $50 and $200 you could get set up with Room EQ Wizard and do it pretty much all automatically. The cost window depends on what you have already and what you need. If your soundcard is full duplex, which most are these days AFAIK, a few cables and a measurement mic like Behringer ECM8000 (~$50) is the biggest expense. A SPL meter is handy too if you want to get more absolute measurements but not necessary for relative measurements. Some SPL meters, like some of the Radio Shack ones, can also be used for the measurement mic.

So, not possible with just some test tones playing and walking around the room?

Jeff

 

[whoaru99]

Possible? Perhaps. Effective? Dunno.

A full frequency sweep takes about 5 second in REQW and you have a clear visual result.

 

[ebacon]

Jeff,

There is a bunch of free audio generator software on the web. All of them generate test tones. Simply search "audio generator software".

REQW might even be a free one. You don't need to use a microphone with it. Using your ears is fair game.

I tried playing the microphone measurment game and it was a bigger PITA than it was worth. I was never sure of the curve I got. Behind the scenes are a PC calibration curve, the microphone curve, the stereo curve, and the room curve. What I learned is this:

I whistle at about 2 KHz.

 

[whoaru99]

It's not that hard...

REQW is free. Sound card cal is a one time deal (unless you change cards). Mic cal is a file you download and just save in a directory, or don't even need it, really, for relative measurements. Room curve is what you measure with one push of a button, house curve is what you lay out that you're trying to adjust to - again, not even necessary for relative measure.

Here's an example of a sweep I did at some point in the past...

These are one in the same measurement, just that the first one has smoothing applied, second one not smoothed.

 

[toxcrusadr]

There are also calculators available on the web that allow you to input the dimensions of your room and they will spit out the expected problem frequencies. Which will not solve all your problems but it may help inform the rest of your investigation.

 

[ebacon]

Sound card cal was the PITA that stopped me dead in my tracks. For some reason on the loopback test my system would not generate the requisite signal amplitude. Outside of loopback everthing seemed to work fine. I simply wasn't certain about the curves I was getting.

There's also chatter that the microphone calibration curves from Dayton are garbage. I dunno the truth, but I do know that rumor adds another layer to my uncertainty cake.

Jeff is free to borrow my gear and play with it if he wants.

In the meantime I have an audio generator with a knob. Spinning the knob makes the stereo go wheeyoow wheeyoow and that's enough for me ATM.

 

[jmathers]

I like the idea of going all out with REW but don't have the stuff right now and wanted something quick and dirty.

Based on room dimension measurements the modes (or is it nodes?) are at 43Hz, 21Hz, and 70Hz.

My question is what the heck do I do with those numbers? Does it point me to anything specific to address? From what I understand these would be the frequencies that the room hangs on to - so these frequencies should be louder than everything else. Have I got that right? Or is it opposite?

Looking at the graphs is easier too - I can clearly see that at 40Hz whoaru's room is down by 10db or so. This tells me that frequency is softer - a suck out perhaps?

Sorry if I've bungled the logic here......

Jeff

 

[ebacon]

That's a good question. I would bump up 40 Hz on the eq. The 400 Hz dip is a harmonic of 40 so I would expect it to follow the treatment at 40 but with less effect.

 

[jmathers]

That's a good question. I would bump up 40 Hz on the eq. The 400 Hz dip is a harmonic of 40 so I would expect it to follow the treatment at 40 but with less effect.

And if you don't have an EQ? What then?


Jeff

 

[whoaru99]

That's a good question. I would bump up 40 Hz on the eq. The 400 Hz dip is a harmonic of 40 so I would expect it to follow the treatment at 40 but with less effect.

That doesn't really help though with nulls. More or less, if it's cancellation/null, it's like trying to fill a black hole, it just keeps sucking up whatever you put in. This is where room treatments and some forms of electronic room correction (beyond simple equalization), and multiple subs (in case of bass issues) come in.

As to what it is in that particular graph, I don't recall if I was dialing in my sub (phase issues around crossover to mains, etc.) or what exactly was going on there.

 

[ebacon]

And if you don't have an EQ? What then?


Jeff

Hold on a minute J.R. Whoareu is fixin' ta tell us.

 

[jmathers]

I'm still not sure what to do with these numbers:

Based on room dimension measurements the modes (or is it nodes?) are at 43Hz, 21Hz, and 70Hz.

The problem in the room is a considerable lack of mid bass frequencies. Do those numbers suggest that mid bass is a problem? Again, I thought modes would be frequencies that the room wants to accentuate.....adding the harmonics would suggest that mid bass should be no problem whatsoever if modes enhance those frequencies. So, confused still.

Jeff

 

[RevMen]

Based on room dimension measurements the modes (or is it nodes?) are at 43Hz, 21Hz, and 70Hz.

My question is what the heck do I do with those numbers? Does it point me to anything specific to address? From what I understand these would be the frequencies that the room hangs on to - so these frequencies should be louder than everything else. Have I got that right? Or is it opposite?

It's both.

A sound wave is an energy exchange between pressure and particle velocity. Where the air particles are at their highest pressure (what you can hear), they're at their lowest velocity. And when they're at their highest velocity (what you can't hear), they're at their lowest pressure.

A room mode is a frequency where the room is resonant. It happens because the length of the room is a multiple of the wavelength of the frequency.

A node is a location where pressure peaks of waves overlap, causing an even higher sound pressure (louder). The overlapping waves can be difference pieces of a wave reflecting off of walls. That is why you get nodes at room mode frequencies, because the dimensions of the room, with respect to the speed of sound, cause the sound to reflect back into itself at just the right time.

An anti-node is the same thing, except where pressures cancel out.

You can figure out where a node should be based on the wavelength of a room mode.

Going back to a sound wave being a period exchange between pressure and velocity...

On the surface of a wall, the air particle velocity must be 0. The air can't move into the wall, and it can't move away from it ( or else it leaves a vacuum). So at that location, you have a pressure zone, where the pressure goes up and down but the velocity never changes. (This is why you usually hear a bass boost when you put your ear near a wall or corner.) In a free field the pressure fields move along and don't hang out in a single location like they do at a wall.

If you move away from the wall a distance of 1/4-wavelength, you arrive at the zone of maximum velocity, where sound pressure is at a minimum. The air particles are moving back and forth towards and away from the wall here, but they never develop pressure. You can't hear velocity, only pressure, so at this location the sound drops out. I guess this is what you mean by suck out.

If you move away from the wall 1/2-wavelength, you arrive at the next pressure zone that is always out of phase with the pressure at the wall. This is where the sound will be its loudest.

Now if your room dimensions are such that pressure and velocity zones from sound waves bouncing off of opposite walls line up, you get spots where the sound is very loud and/or spots where the sound is very quiet. These are the nodes and anti-nodes.

The way to solve this is to keep coherent waves of the same frequency from bouncing into each other. This is the reason for building a room with non-parallel walls. It's also why you add diffusers. Diffusers break up the sound waves as they bounce, scattering little pieces of them in all directions.

 

[RevMen]

I'm still not sure what to do with these numbers:

Based on room dimension measurements the modes (or is it nodes?) are at 43Hz, 21Hz, and 70Hz.

The problem in the room is a considerable lack of mid bass frequencies. Do those numbers suggest that mid bass is a problem?

Those are pretty low frequencies, but multiples of those frequencies will also be modes. So you may be hearing one of the higher order modes.

Again, I thought modes would be frequencies that the room wants to accentuate

Amplify in some spots, attenuate in others. You get both effects at the same time at different distances along the path of the sound wave.

 

[toxcrusadr]

All of which means, if you have a listening position, do your tests there. If you don't, i.e. you move around a lot, you'll probably never get the whole room to sound spectacular but will degerate into gibbering insanity in the process.

Kidding! Mostly.

 

[sKiZo]

This is the reason for building a room with non-parallel walls.

Or in my case, mostly because I'm a Polish carpenter and couldn't build a square corner if my life depended on it. Come to find out that's a good thing? Kewl ...

Mid bass is a very common problem, and it's usually mostly to do with how your speakers couple to the floor. Takes some experimenting to fix that naturally - either isolation or height adjustments or absorption pads or ... whatever, depending on your environment. Lots easier to fix electronically.

I went with the TrueRTA software on my laptop, an external sound card (because most any built in is shite, and that's being generous), a breakout box, a Dayton mike, a Panasonic 31 band pro eq, a McIntosh passive eq, a hope, and yes a prayer. Killed some serious resonance by adding pads under the big boxes, then attacked the big lumps in the bottom with the passive eq, then fine tuned with the 31 band graphic. Click the go button on the last pass and run it at the 1/3rd octave level that matches my eq switches, get the printout and match the sliders to that.

I remember when it took people in lab coats with uber loads of sophisticated equipment to do that kind of stuff. Now it's pretty much a DIY no brainer - hardest part is getting all the toys to work together properly. Most of the software packages will even allow you to overlay a "room curve" to match your predominant music preferences. Which is a good thing, as a perfectly flat room just ain't right either. One extreme to the other kind of thing. All told, it cost me around $200 to get what I needed that I didn't already have. I've spent a lot more for a lot less.

If all this sounds like too much work, there's always headphones ...

 

[ebacon]

Get my underwear at K-Mart.
Judge Wopner comes on at two. Yeah. Judge Wopner.
Counting cards. Counting cards.
A whistle is two kilohertz.

 

[toxcrusadr]

Get my underwear at K-Mart.
Judge Wopner comes on at two. Yeah. Judge Wopner.
Counting cards. Counting cards.
A whistle is two kilohertz.

There, there. Relax and take your Thorazine.