Speakers in a null - Why is this bad?

[DM1]

I get how to figure out where the worst nulls are in my room (using knightfly's technique,) and it makes sense why I wouldn't want my ears in one of these nulls ... But I don't understand why it's bad to place my speakers there.

The axial modes don't change based on where I put my speakers. And my speakers will emit the same sound regardless of where they're placed in the room. So why is it specifially bad to have them in a null? Sure if I sit in one of these nulls, certain frequencies will be attenuated. But what's happening with the standing waves at my speaker position should be irrelevant to what I hear in my sitting position. Shoudn't it? What am I missing?

[apl]

Originally Posted by DM1

I get how to figure out where the worst nulls are in my room (using knightfly's technique,) and it makes sense why I wouldn't want my ears in one of these nulls ... But I don't understand why it's bad to place my speakers there.

The axial modes don't change based on where I put my speakers. And my speakers will emit the same sound regardless of where they're placed in the room. So why is it specifially bad to have them in a null? Sure if I sit in one of these nulls, certain frequencies will be attenuated. But what's happening with the standing waves at my speaker position should be irrelevant to what I hear in my sitting position. Shoudn't it? What am I missing?

A null is a place where air won't vibrate as much as it should at the corresponding frequency. Your ears won't hear that frequency there, and conversely, your monitors will not be able to cause vibration at that frequency in those spots.

A baseball bat has a two nulls (or nodes) in its first bending mode of vibration. It has antinulls at either end and in the middle. This bending mode has a certain frequency associated with it. If you smack the bat at the end on concrete it will sting your hands because you've excited the first bending mode. If you smack a baseball on the null closest to the end of the bat your hands won't sting because you can't get the vibration of the first bending mode by exciting it at that mode's null.

[DM1]

Originally Posted by apl

A null is a place where air won't vibrate as much as it should at the corresponding frequency.

So even though the speaker is moving, the air at that point won't pick up the vibration?

[apl]

Originally Posted by DM1

So even though the speaker is moving, the air at that point won't pick up the vibration?

Exactly.




.......

[TexRoadkill]

Here is a good article

http://www.kettering.edu/~drussell/D...s/driving.html

[TexRoadkill]

I hadn't read Steve's method for speaker placement before but unless I'm getting confused it would run counter to the article I posted. My understanding of that acoustics article is that it's better to put your sources in an antinode (null) because then they will not excite the room modes.

Anybody know enough to clarify?

[mixsit]

Originally Posted by TexRoadkill

I hadn't read Steve's method for speaker placement before but unless I'm getting confused it would run counter to the article I posted. My understanding of that acoustics article is that it's better to put your sources in an antinode (null) because then they will not excite the room modes.

Anybody know enough to clarify?

I noted that his spreadsheet program had the desirable spots where there are neither peaks nor nulls which makes sense (for my very crude level of understanding ).. looking for the most neutral spots.

By the way, I didn't see in the Russell article where he was advocating one way or the other though.
Wayne

[DM1]

Originally Posted by the author of the page TexRoadkill linked

If the source is located at a pressure node for a given resonance frequency, the room will not respond no matter how loud the source is.

Wow ... I know it's basic physics, but this still blows my mind.

Originally Posted by TexRoadkill

My understanding of that acoustics article is that it's better to put your sources in an antinode (null) because then they will not excite the room modes.

Anybody know enough to clarify?

I definitely don't know enough But I think the quote above is relevant. When you put the speakers in a null, they can never emit the corresponding frequency. (And mixsit is right: Steve's appoach tries to find the "least bad" spots, which are somewhere between the nodes and antimodes of the main axial waves.)

[apl]

An untreated room can have nulls and anit-nulls only inches apart.

www.ethanwiner.com/acoustics.html

[SouthSIDE Glen]

Originally Posted by DM1

I get how to figure out where the worst nulls are in my room (using knightfly's technique,) and it makes sense why I wouldn't want my ears in one of these nulls ... But I don't understand why it's bad to place my speakers there.

It seems to me there are several important caveats to consider when reading Knightfly's thread. Somebody please correct me where I might be wrong here...

First, he's assuming an untreated, completly empty room with perfectly white reflections on all surfaces. Those are the only rooms to which those maps will accurately apply.

Second, those maps imply that in such rooms, 87.5% of the volume of the room can be considered as "bad placement" for your ears and for your monitors. This means that any given placement for your head or for one of your monitors has only a 1 out of 8 chance of being OK. That's only a 1 out of 512 (0.2%) chance that any given setup will get two monitors and one head all positioned well within a room at the same time. That seems awfully pessimistic and unrealistic.

Third, the Knightfly/Siefert method calculates modes and nulls based strictly on room dimensions, and therefore implies that speaker position is not relevant to the calculation. If this were true, then it wouldn't matter if the speaker were placed in a null or not, since the null locations are fixed by room dimension anyway. But we all know that speaker position relative to the room boundaries *does* matter, that there are bass buildups due to corner reflections if placed too close to a corner, that the speakers need room to "breathe" away from the rear wall to avoid certain cancellations for that first reflection, etc.

Regardless of the frequency, the length of it's wave has to be measured starting from it's source, not from the surface of an arbitrarily selected wall. As such, the position of the speakers themselves should affect the position of modes and antimodes within the room. Add to that the phase issues resulting from the two speakers/sources affecting each other, and you have a room that'll be *far* different then the Knightfly/Siefert map will show.

And we haven't even talked about populating the room with furnitire and treatment yet.

G.

[DM1]

I'm definitely not an expert Glen, but I have some thoughts:

Originally Posted by SouthSIDE Glen

he's assuming an untreated, completly empty room with perfectly white reflections on all surfaces. Those are the only rooms to which those maps will accurately apply.

The charts deal with the first 4 axial standing waves, which in most rooms are bass frequencies .. So I think his assumption is fair, since (again, in most rooms) these frequencies are least affected by wall hangings and the like.

Originally Posted by SouthSIDE Glen

That's only a 1 out of 512 (0.2%) chance that any given setup will get two monitors and one head all positioned well within a room at the same time. That seems awfully pessimistic and unrealistic.

Knightfly kind of addressed this in the linked thread, in the post where he says "This is what I would call a damned if you do, damned if you don't situation."

It's obviously not a perfect system, but every little bit helps.

Originally Posted by SouthSIDE Glen

Third, the Knightfly/Siefert method calculates modes and nulls based strictly on room dimensions, and therefore implies that speaker position is not relevant to the calculation.

As far as the modes are concerned, I'm pretty sure speaker position ISN'T important. Modes are a physical property, independent of the presence of sound waves.

I think the point of Knighfly's chart is that if you place a speaker into one of the nulls, it'll be incapable of emitting the corresponding frequency, so we should avoid the nulls. But the converse needn't be true. That is, even if it's possible to move the speaker and have that frequency appear again, you still need to avoid the antinode. So maybe his chart would be clearer if he coloured it red in all the bad places, and said "here's where NOT to place speakers."

Originally Posted by SouthSIDE Glen

Regardless of the frequency, the length of it's wave has to be measured starting from it's source, not from the surface of an arbitrarily selected wall.

I can't get my head around the physics just now, but here's a thought: Standing waves arise from the reflection of a single wavelength between 2 parallel surfaces. As you move the source towards one of these surfaces, you move it away from the other in the same amount ... So wouldn't the null stay in the same place?

-Doug

[apl]

Originally Posted by SouthSIDE Glen

As such, the position of the speakers themselves should affect the position of modes and antimodes within the room.

Nodes, nulls, and modes! Oh, my!

Anyway, those are properties of the room. The changes to those properties as a result of the speakers being in the room are way negligible.

[Ethan Winer]

Tex,

> My understanding of that acoustics article is that it's better to put your sources in an antinode (null) because then they will not excite the room modes. <

It may not excite room modes there, but if it's a true null you won't hear anything at all!

--Ethan

[mixsit]

Originally Posted by SouthSIDE Glen

That's only a 1 out of 512 (0.2%) chance that any given setup will get two monitors and one head all positioned well within a room at the same time. That seems awfully pessimistic and unrealistic.

Not if you consider that a) We're pretty tolerant of these problems to begin with. b) All the other variables that would stack up to average things out.
I doubt these nulls are infinitely deep. Looks to me like he's simply providing the spots to consider and avoid.

Originally Posted by DM1

...Knightfly kind of addressed this in the linked thread, in the post where he says "This is what I would call a damned if you do, damned if you don't situation."

What he was referring to there was regards the question of dead-center width wise being in a bad listening spot, but you wouldn't want to not be centered for L/R imaging. Again I think an example where you'd be aware, but still have good results in spite of it being 'poor' in that regard.

[mixsit]

Originally Posted by SouthSIDE Glen

It seems to me there are several important caveats to consider when reading Knightfly's thread. Somebody please correct me where I might be wrong here...

First, he's assuming an untreated, completly empty room with perfectly white reflections on all surfaces. Those are the only rooms to which those maps will accurately apply..

I thought Ethan might have caught this one. Unless I misunderstand, treatment doesn't move the peaks and dips (in the low end of course) or change the frequencies, but only makes them less deep or tall yes?

[SouthSIDE Glen]

Originally Posted by DM1

The charts deal with the first 4 axial standing waves, which in most rooms are bass frequencies .. So I think his assumption is fair, since (again, in most rooms) these frequencies are least affected by wall hangings and the like.
...
Standing waves arise from the reflection of a single wavelength between 2 parallel surfaces. As you move the source towards one of these surfaces, you move it away from the other in the same amount ... So wouldn't the null stay in the same place?

Doug (and APL),

Thanks for the responses. For the record, I am not an expert on this subject either; I'm just thinking "out loud" and coming up with honest questions.

And you're right about the standing wave locations as being a property of the room independant of speaker location. I was thinking the wrong way in my original post about that. I understand that part of it now.

Also agreed on stuff like wall hangings. Larger items like chairs, couches, bookcases, people, etc., however, are a different story. Read on...

Originally Posted by DM1

It's obviously not a perfect system, but every little bit helps.

I guess the thing that comcerns me the most is that people are going to take that formula and the resulting charts as fast and precise answers to the question of room acoustics and placement. As soon as one complicates the situation with things like the furniture described above (including the mixing desk itself) and people (walking bass traps), it *does* change the properties of the room, even at low frequencies.

As APL stated earlier, and as we have seen time and time again on this forum, inches can make a world of difference. When the properties of the room are changed by everyday things of real mass (not just curtains and foam) that most of us do have in our home studios, these inches can move around quite a bit.

There was an interesting article in Mix (I think; it could have been SOS) in either September or October about a guy in a Big Boy studio control room that was trying to tame the bass to his liking. He found - and showed with detailed lab-quality RTA measurements - how just removing a tape deck and some empty guitar cases belonging to the band from one side wall signifigantly changed his bass response map; he also demonstrated the large difference it made when he opened up the space under the front of a mixing desk instead of having a front "kickboard" there.

Heck, we all preach every day how with just a couple of rolls of fiberglass we can treat rooms to affect the bass. We also preach how speaker placement - while maybe not affecting the properties of the room itself - does have an effect on the perceived response in other ways; things like the need for symmetry, for moving speakers off the wall, how choosing properly between the long wall and the short wall, etc. all make a large difference in the bass response characteristics of the monitoring chain, and that are not all necessarily reflected (pun intended) in the Knightfly maps. OK, the keeping off the wall is inferred in those maps, but there is nothing in those maps that indicates any need for symmetry, for staying away from corners, or for choosing long vs short walls, let alone how the shape of the mixing desk or the location of a couple of guitar cases can have a signifigant effect.

As long as you are in the green blocks you are OK. And as long as the room volume (not just the walls) is empty. The first assumption is wrong if you choose asymmetric green block positions, and the second one just does not happen in real life.

Every little bit helps, yes. I just feel the need to point out that hs method helps in describing and understanding the theoretical underpinings, but is probably not going to actually provide accurate maps for the average home recording control roomm, and ignores other important factors in room design and placement above and beyond the relationship of standing wave frequencies to wall dimension. Those looking to that spreadsheet and chart method as an easy button solution to finding ideal speaker and head locations is just likely to be disappointed as pleased, I would think. That's I think the important point here.

Again, just trying to hash it all out in my head.

G.

[SouthSIDE Glen]

Originally Posted by mixsit

I thought Ethan might have caught this one. Unless I misunderstand, treatment doesn't move the peaks and dips (in the low end of course) or change the frequencies, but only makes them less deep or tall yes?

As I understand it, that's true for treatment like bass trapping and such.

But what happens when you have items of signifigant mass and size (desks, couches, bookcases, etc) that can be said to signifigantly change the dimensions of the room; i.e. to change the numbers of, distances and locations of parallel reflecting surfaces? I would think that is actually changing the properties of the room itself, and therefore changing more than just the amplitude of the mode map.

G.

[mixsit]

Originally Posted by SouthSIDE Glen

As I understand it, that's true for treatment like bass trapping and such.

But what happens when you have items of signifigant mass and size (desks, couches, bookcases, etc) that can be said to signifigantly change the dimensions of the room; i.e. to change the numbers of, distances and locations of parallel reflecting surfaces? I would think that is actually changing the properties of the room itself, and therefore changing more than just the amplitude of the mode map.

G.

My guess (since I'm sorting and gathering here too BTY. .. it (they) add new reflections and resonances -(both complicating things but also just as likely helping to average things out?) but the primary wall effects remain.
Wayne

[SouthSIDE Glen]

Let me muddy it up even more (sorry!) ...

To bring things back to the original thread question, if the mode map of the room remains fixed regardless of the speaker location (I think of it rather like a "field potential" of the room), then what difference does it make whether the speaker is placed in a null or not?

It seems to me that the speaker is still getting the air moving, still creating a sound that will fill the field in relative amplitudes according to the potential mode map. The fact that wave cancellations are happening at the location of the speaker should not cancel out the sound elsewhere in the room, I would not think. I'd think that a speaker placed in, say, a 60Hz (just an arbitrary number here) null would still produce a loud 60Hz sound at the location of of a potential peak mode in the room.

Put in a more Zen way, does a mode map have any meaning if there is no sound? A sound frequency has to exist before it can cancel itself out.

If it exists, it will follow (more or less) the mode map. If the speaker is placed in a null location, it's still emitting that frequency, and therefore still following the mode map. It has to, because it if didn't, there'd be no standing wave and therefore no cancellation and no actual null location as predicted by the map.

Right, Master Po?

G.

[apl]

Originally Posted by SouthSIDE Glen

Let me muddy it up even more (sorry!) ...

To bring things back to the original thread question, if the mode map of the room remains fixed regardless of the speaker location (I think of it rather like a "field potential" of the room), then what difference does it make whether the speaker is placed in a null or not?

It seems to me that the speaker is still getting the air moving, still creating a sound that will fill the field in relative amplitudes according to the potential mode map. The fact that wave cancellations are happening at the location of the speaker should not cancel out the sound elsewhere in the room, I would not think. I'd think that a speaker placed in, say, a 60Hz (just an arbitrary number here) null would still produce a loud 60Hz sound at the location of of a potential peak mode in the room.

Put in a more Zen way, does a mode map have any meaning if there is no sound? A sound frequency has to exist before it can cancel itself out.

If it exists, it will follow (more or less) the mode map. If the speaker is placed in a null location, it's still emitting that frequency, and therefore still following the mode map. It has to, because it if didn't, there'd be no standing wave and therefore no cancellation and no actual null location as predicted by the map.

Right, Master Po?

G.

A guitar string is like a room with many modes. A pick is a speaker, putting energy of a broad spectrum of frequencies into the string. But the string can only respond to certain frequency, ie, the note that it's tuned to and its harmonics. Furthermore, take an open string on the guitar and pick it at the 12th fret. This is the null of the second bending mode of the string. Look at the signal through an analyzer and you'll see almost zero of two times the open note frequency. Pick it again at the fifth fret and you'll see boatloads of the 2x but no 4x because that's the anti-null for the second and null for the fourth bending mode. That volume of air in the room has modes, too, but not as clearly definded.

A perfect reflection free room, an anechoic chamber, has no nulls or antinulls, but sounds very dull. We want to have interesting reflections, like the floor so our brain can figure out how far away a source is. When we add treatment to a room we do a couple of things. First, we add absorption so that sound gets to leave the room. We also add damping. This has the effect of reducing the effect of a null or anti-null, but there's no free lunch, so it has the opposite effect on nearby frequencies, changing the Q of the room at a given frequency to borrow from electronics.

This might be a crappy explanation.

[SouthSIDE Glen]

Originally Posted by apl

A guitar string is like a room with many modes. A pick is a speaker, putting energy of a broad spectrum of frequencies into the string.
...
This might be a crappy explanation.

No, actually that is - for me, anyway - and excellent analogy and explanation which helps me out quite a bit. Thanks!

G.

[apl]

I put this afterthought into the wrong thread...

You can't make an open A string sound a B even if you try to mechanically excite it with an electromagnetic driver.

[Ethan Winer]

Originally Posted by mixsit

treatment doesn't move the peaks and dips (in the low end of course) or change the frequencies, but only makes them less deep or tall yes?

Yes.

Also, the text below from a recent post I made elsewhere is relevant here.

--Ethan

There are three separate factors that determine the response at any given location in a room.

First is the room's natural resonances, or modes. These are independent of speaker and listener placement, and the frequencies are determined entirely by the wall-wall and floor-ceiling spacings. If you stand in an empty stairwell and clap your hands, you'll hear a "boing" sound whose pitch is usually determined by the spacing between the walls. Whether you stand near to one wall or the other, or crouch down low or clap your hands high above your head, the pitch of the "boing" remains the same. That is the mode frequency.

The other two factors are SBIR and LBIR - Speaker-Boundary Interference Response and Listener-Boundary Interference Response. These are the classic 1/4 wavelength related peaks and nulls whose frequencies are based on the distance between each speaker and each boundary, and from the listener and each boundary.

The response at any given cubic inch in a room is the sum of all three of these factors, and that's a lot of variables given two speakers, one listener, and six boundaries. There are programs like RPG's RoomSizer that attempt to calculate the response based on all of these variables, but I've never tried them. People I trust say they're somewhat useful as a starting point, but they never give the same results as actually measuring with a microphone and meter etc.

[SouthSIDE Glen]

Originally Posted by Ethan Winer

The other two factors are SBIR and LBIR - Speaker-Boundary Interference Response and Listener-Boundary Interference Response. These are the classic 1/4 wavelength related peaks and nulls whose frequencies are based on the distance between each speaker and each boundary, and from the listener and each boundary.

The response at any given cubic inch in a room is the sum of all three of these factors, and that's a lot of variables given two speakers, one listener, and six boundaries. There are programs like RPG's RoomSizer that attempt to calculate the response based on all of these variables, but I've never tried them. People I trust say they're somewhat useful as a starting point, but they never give the same results as actually measuring with a microphone and meter etc.

THANK YOU! I was hoping you'd chime in on this Ethan, you usually manage to get ideas through my thick skull pretty well

While I knew I did not - until halfway through this thread - have the right mental handle on mode resonances, I knew there was more to the situation it than just that; that placement did make a difference (except just not with the simple mode map, which is where I made one of my mistakes), and that maps like what are shown in the Nightfly thread were far too simplified and idealized to bear much real realtion to what's happening in a real room. Which was my main concern.

G.