Same here. Rockwool suspended a bit off the floor topped by much audio poly was best but not good enough. In the end I just bought a small ladder and measure speakers at mid height of room, taking "vertical" measures with the speaker on its side, moving the mic horizontally.
I also use diffraction modeling tools to determine how close I can get with < 0.25 db error in the diffraction contribution (depends on baffle size). Also use the excel sheet I posted to find mic and speaker placement that maximizes window size.
Yeah, that transition region is a tough nut to crack.
Interesting about the angle of incidence, makes sense. I would have expected the tests to be done on the surface normal. But, yeah, other angles would matter a lot, it seems. I would expect that a grazing angle might make it ineffective, but then again, if you can make the sound pass through rather than bounce off the top, it would have more material to go through. I guess there's probably more to it than meets the eye.
But still, all that is academic as the fact is it would have to be really thick or set away from the floor to do any good at 100-150Hz.
Good feedback, thanks.
Oh, wow. How interesting! Thanks for the bit of history.
Even more history, thanks. Cool stuff!
I'm really jealous you were a part of that. I wish so much I could have heard some of those conversations, had a few smoke breaks with you guys. You must have some sentimental memories of those days. In some sense, the early 1980s seem like just yesterday to me, and in other ways, it was a million years ago. Remember you could still smoke at your desk back then and everybody did? The only place you couldn't smoke was the raised floor cool rooms for the computer systems, the ones with tape drives and disks the size of washing machines...
...OK, sorry, getting silly sentimental.
Well, sure, that's true. Having the one woofer just run to 100Hz, and the other run alone up to the DI-matching frequency for crossover to the tweeter, it was a point source through the whole audio bandwidth. A great speaker with a lot of oomph. Excellent system.
My first really great woofers were JBL 2205s. I just fell in love with that woofer. And the 2235 was even better for the low end, a little tradeoff of efficiency for extension. Those JBL Pro woofers were paradigm shifters for me. Used to get them through Ford Audio. My first constant directivity cornerhorns used JBL Professional Series drivers throughout. Incredible, I loved them.
On the 4435, the one thing I might have changed if I had my druthers, would have been the orientation of the woofers. Please forgive the armchair quarterbacking. No fair since it's 30 years later. And it's a teeny-tiny deal, not worth mentioning. But I do like to chat.
I think that would take the 4435 model from the 98th-percentile to a 99-percenter. Arrange the cabinet as a tower, which then would require finding the positions for the woofers that decreased the effects of internal standing wave modes on the one that would be run up through the midrange. So put the woofers where they needed to be to satisfy that, and still be aligned vertically.
Leave the crossover from midwoofer-to-tweeter as it is, and of course maintain the physical relationship between the two, so the forward lobe stayed right. But blend the two woofers up an octave or so higher, to help smooth that vertical mode like we've been talking about.
And finally, I'd turn the lower woofer around, face the magnet to the listener and recess it within the cabinet about 9 to 12 inches or so, enough that a grille could be put over it to hide the magnet. Run it so both cones moved forward (toward the listener) when positive voltage was applied. This would reduce distortion below 100Hz.
The shorting ring works great above 100Hz, make those 22xx series woofers best-of-class drivers, in my opinion. The ring reduces distortion by about 15dB above 100Hz. But since you have both woofers, push/pull drive could be used to improve distortion down low too, where the shorting rings lose effectiveness. The fore-aft offset would probably even help smooth the reflection notch from the wall behind the speaker, much as the vertical offset smoothes the vertical mode.
I've thought about this woofer rearrangement when I've looked at the 4435, but then again, like I said, it's easy to armchair quarterback on an internet messageboard. I think it's an excellent speaker as it is, and the 4430 and 4435 were definitely world class speakers, in my opinion. It's kind of funny to me, because now days "waveguide" speakers are all the rage. Same stuff, all the important stuff was right there, 30 years ago. Some classic designs truly stand the test of time.
I took a few minutes this past weekend to measure a couple speakers outdoors (and briefly indoors), specifically looking at the floor bounce notch formed when sound source and microphone are both placed some distance above the ground.
My only goal was to determine whether or not the notch frequency changed as distance changed, which would tell me if it was caused by destructive interference from a path length delta, from self-interference from the boundary when spaced 1/4λ away, or perhaps a combination of both.
Kind of a no-brainer in hindsight. Up close, the notch is approximately equal to the 1/4λ frequency, the further away you go, the higher the notch drifts upwards. It also becomes smaller in amplitude. (So I guess you can always just sit really far away
) This means it is due to path length delta.
The response with microphone on the ground is much the same as the ground plane measurement with the microphone suspended.
So the 1/4λ thing outdoors is a special case where the speaker is close to the listener.
What confounds is the fact that indoors, there are so many interactions. The low frequency room modes are sort of fixed in position, in that they make well defined dips and peaks that don't change much with vertical movement. But the higher frequency modes do change. When the speaker is on the ground, for example, I measured no dip above 100Hz. On a stand, it formed. Kinda looked like that close-up 1/4λ notch. Whatever the cause, when blending flanking subs low-passed high enough to reach, the notch went away again.
Likewise, when using a vertical array outdoors (i.e. blended midwoofers or flanking subs), the close-up notch goes away too. No floor bounce or vertical modes when using a vertical array.
I had intended to do more testing, perhaps make a little demonstration video. But I have a 4-year-old running around and don't have as much time as I used to. This is part of why I waited to post until today - I had hoped to be able to put together a series of demonstrative response charts and notes.
I regularly find a midbass/midrange mode (probably the first and second floor/ceiling axial modes) in many situations where the speaker is raised off the ground a couple feet. It may be just room modes, or it may be combined with other forms of self-interference, like path length delta notches. Seems a common problem, as I've seen it in several measurements performed by many different people in many different rooms. Lots of data that shows it in this thread alone.
But after talking with some of you here, I've become more interested in the exact causes of this behavior at midbass/midrange frequencies. I used to think it was all path length deltas. Then I started seeing some odd things that made me thing maybe there was a mirror-interference thing going on. I guess it all kind of is that, but more specifically, I started thinking there was a notch at the frequency where the speaker was 1/4λ from any boundary, not just a path length delta. And then of course, there are room modes in play.
One thing I noticed fairly early on was that a vertical array (even a short one, with ground reflection, i.e. 2 drivers plus the 2 ground reflections) would mitigate the notch and smooth the midrange. It kind of didn't matter what the interference cause was, because the array fills in the hole. But it has intrigued me. I now wonder if the 100-150Hz notch is always the second axial mode from floor to ceiling, or maybe interference from the wall behind the speaker. I'm anxious to have time to do some more measurements, and hope some more of you will too.
My only goal was to determine whether or not the notch frequency changed as distance changed, which would tell me if it was caused by destructive interference from a path length delta, from self-interference from the boundary when spaced 1/4λ away, or perhaps a combination of both.
Kind of a no-brainer in hindsight. Up close, the notch is approximately equal to the 1/4λ frequency, the further away you go, the higher the notch drifts upwards. It also becomes smaller in amplitude. (So I guess you can always just sit really far away
) This means it is due to path length delta.The response with microphone on the ground is much the same as the ground plane measurement with the microphone suspended.
So the 1/4λ thing outdoors is a special case where the speaker is close to the listener.
What confounds is the fact that indoors, there are so many interactions. The low frequency room modes are sort of fixed in position, in that they make well defined dips and peaks that don't change much with vertical movement. But the higher frequency modes do change. When the speaker is on the ground, for example, I measured no dip above 100Hz. On a stand, it formed. Kinda looked like that close-up 1/4λ notch. Whatever the cause, when blending flanking subs low-passed high enough to reach, the notch went away again.
Likewise, when using a vertical array outdoors (i.e. blended midwoofers or flanking subs), the close-up notch goes away too. No floor bounce or vertical modes when using a vertical array.
I had intended to do more testing, perhaps make a little demonstration video. But I have a 4-year-old running around and don't have as much time as I used to. This is part of why I waited to post until today - I had hoped to be able to put together a series of demonstrative response charts and notes.
I regularly find a midbass/midrange mode (probably the first and second floor/ceiling axial modes) in many situations where the speaker is raised off the ground a couple feet. It may be just room modes, or it may be combined with other forms of self-interference, like path length delta notches. Seems a common problem, as I've seen it in several measurements performed by many different people in many different rooms. Lots of data that shows it in this thread alone.
But after talking with some of you here, I've become more interested in the exact causes of this behavior at midbass/midrange frequencies. I used to think it was all path length deltas. Then I started seeing some odd things that made me thing maybe there was a mirror-interference thing going on. I guess it all kind of is that, but more specifically, I started thinking there was a notch at the frequency where the speaker was 1/4λ from any boundary, not just a path length delta. And then of course, there are room modes in play.
One thing I noticed fairly early on was that a vertical array (even a short one, with ground reflection, i.e. 2 drivers plus the 2 ground reflections) would mitigate the notch and smooth the midrange. It kind of didn't matter what the interference cause was, because the array fills in the hole. But it has intrigued me. I now wonder if the 100-150Hz notch is always the second axial mode from floor to ceiling, or maybe interference from the wall behind the speaker. I'm anxious to have time to do some more measurements, and hope some more of you will too.
Last edited:
Yes, of course nothing of any reasonable thickness is going to deal with below 200 Hz, but then agin the ear is not very good at resolving things like this at those frequencies. In a small room the time window is getting pretty great and that tends to fill in any holes from single reflections. IOW, I'm not so sure that killing anything below 200 Hz is necessary.
FYI, the absorption thing can get very complex. Its like the classic optics problem of total internal reflection between two media. Things vary with angle, and there can be angles where the absorption increases and others where is decreases, but in general a decrease happens at larger angles. But this again has to be conditioned. A wave traveling at a grazing incidence along an absorption will see some significant absoption because of the time aspect of the travel.
Ino Ingard wrote an entire book on "Absorption" (or something like that) some years back (many many years back!). I reviewed it for the AES so I got to keep a copy. I have never seen it anywhere else, but it is a gem of information. The book was about 200 pages on nothing but what we are talking about here. Its a very complex subject.
Last edited:
But where's the dividing line between what's important and what's not? Is it 100Hz? 130Hz? 160Hz? What frequency range do you consider important enough to smooth with multisubs and where do you think the ear stops being "very good at resolving things like this"?
I mean, I think we'd agree that modes at 40Hz or 60Hz are audible. You can easily tell the difference between a note that's blaring at 50Hz but then a couple notes up in a progression and it's so soft you can't but just barely hear it. In a room with rigid walls, it can be like that.
This is the very point I was concerned with in your arguments about random placement, many years ago. You seem to be satisfied with smoothing the bass, but once you reach the point where localization is beginning to emerge, you seem to think modal smoothing is no longer important. That's roughly the midrange, what some of us call the transition region - from about 100Hz to 200Hz. Clearly, there can be a pretty strong notch or two there, sometimes separated by a peak. When the room/speaker layout forms this notch-peak-notch situation, it is clearly audible, sounding quite nasal.
I know from several discussions with you that you are all about room treatments. I would agree that is important. But room treatments don't work in the transition region. Absorbent materials have to be very thick, or better yet, spaced away from the boundary. It isn't convenient to space the absorbent material away from the floor. I'm not sure that's the boundary I'd want to damp anyway - I think it would be better to damp the ceiling. It would probably be best to use a suspended ceiling, a damping material separated from the more rigid actual ceiling by some distance. Again, not real practical for people with standard eight foot ceilings.
It seems to me that layout/placement is a practical solution. A form of multisub arrangement, but one that can be used through the upper end of the modal region into the transition region. This is one reason I liked the Welti configurations at first, because some of the subs could be placed near the mains and blended higher, through the transition band, then tapered off. I began to the call the subs placed closed to the mains - flanking subs - and proposed that they could be used with one or more distant subs, placed either symmetrically per Welti or asymmetrically as you prescribe.
The blended midbass sources can come in many flavors, from traditional line arrays, or truncated midbass arrays like 2.5-way speakers, low/mid blending or flanking subs. I personally prefer using low/mid blending (in a three-way configuration) or flanking subs (when the mains are two-ways). These approaches fit into the line I'm using. But any of those truncated array methods smooth the transition region and can be used with constant directivity mains and more distant multisubs.
Last edited:
I would very much agree with the above quoted statements in particular. It's also pretty much accepted amongst commercial retail channel manufacturers, who design for it (keeping in mind that there exists alternate philosophies that don't buy it, with speakers you can find "one example of anything" for sale somewhere), and reviewers tend to prefer speakers with good power response behaviour.
Last edited:
I think the most pratical and effective approach is to damp the whole front wall:
- attenuates very early reflections
- reduces adjacent boundary effects
- reduces modal effects
- attenuates reflections form surround speakers
Wouldn't it be the 2nd order vertical mode ? 1st order would put it around 70Hz for an 8 foot / 2.4 metre ceiling...
Simple in reference to
I would suspect grazing might be more effective for such configured sound absorption materials.
Here are more detailed specs. I used the 2" stuff years ago ('80s) in construction of a live end - dead end room. Very effective. I know it doesn't address the "floor bounce issue, but I just thought I'd throw it out there.
Yesterday I got my Z4 out and was top down cursin and after all these years I finally realized that the reason I always say music sounds best in my car is because I prefer outside concerts. Be that rock, jazz or classical. I remember before there were houses all around me I used to set up my Acoustat Model III's on my deck and sit out in the back yard and listen. Loud and proud!
Ok, you can go back to your regular discussion... 🙂
I had these exact bad boys in the early 90s: they all rotted! Snow storm everytime I moved one so they damped the inside of a garbage truck. 🙂
I have been considering placing a sub at the base of each main, to array around the floor bounce and have done some calculations that suggest it will bring things back to a manageable level. I have also thought about a D'Apolitto config. This would make the floor/ceiling close to symmetrical. Do you think this would be a good alternative?
I'm of the (listening) opinion that subs can fill the floor bounce. I presume to base this on looking at the low end as steady state, and therefore such a fix wouldn't apply in a large room. Do you agree, and have you considered any dual woofer configurations?
I guess that I would say that I remain unconvinced that floor bounce below 200 Hz is a problem. So I haven't looked for solutions. I stand by my comments about using a futon on the floor as being sufficiently effective. I have seen nothing to change my opinion.
The floor bounce is completely swamped by the room's modal behavior. So I would agree that the floor bounce itself is not a pronounced problem. The problem is peaks and dips from modes. Mounting the woofer close the the floor probably helps.
Absorption on the floor "just" helps attenuating higher frequency reflections.
I'll vote with the floor-bounce skeptics as something you can measure but not hear with music. While you can make a sanitary drawing of a right-angle sound ray neatly reflecting from the woofer to the floor to the mic, aren't there a million other rays arrayed around at different distances from the floor?
Clothes closets, large heavy drapes, Tectum over deep studs, wall-to-wall wool carpets with felt underlayer, or even wall-to-wall futons on the floor... a DIYer with imagination has a lot more scope for absorbing LFs that aren''t easy to implement commercially, esp. with an understanding SO or if not married.
Clothes closets, large heavy drapes, Tectum over deep studs, wall-to-wall wool carpets with felt underlayer, or even wall-to-wall futons on the floor... a DIYer with imagination has a lot more scope for absorbing LFs that aren''t easy to implement commercially, esp. with an understanding SO or if not married.
Again I ask my rhetorical question:
Where's the dividing line between what's important and what's not? Is it 100Hz? 130Hz? 160Hz? What frequency range do you consider important enough to smooth with multisubs and where do you think the ear stops being "very good at resolving things like this"?
I mean, I think most of us would agree that modes at 40Hz or 60Hz are audible. You can easily tell the difference between a note that's blaring at 50Hz but then a couple notes up in a progression and it's so soft you can't but just barely hear it. In a room with rigid walls, it can be like that.
But why then do some of you that would use multiple subs to smooth room modes below 100Hz, be completely happy to allow them to exist in the transition region between 100Hz and 150Hz or so? In most rooms, there is a pronounced notch in that region.
Modes in this range sometimes make male voices sound thin, sometimes a little too throaty. Add flanking subs and it immediately smooths out. I notice it the most in rigid hotel rooms (like at shows) but have measured it in other rooms too. Just about every measurement chart posted in this thread shows a deep notch between 100Hz and 150Hz, for example.
So to the "unconvinced" - what exactly are you unconvinced about? The measurements? Or the importance of the lower midrange, where fundamentals of so many instruments begin?
Where's the dividing line between what's important and what's not? Is it 100Hz? 130Hz? 160Hz? What frequency range do you consider important enough to smooth with multisubs and where do you think the ear stops being "very good at resolving things like this"?
I mean, I think most of us would agree that modes at 40Hz or 60Hz are audible. You can easily tell the difference between a note that's blaring at 50Hz but then a couple notes up in a progression and it's so soft you can't but just barely hear it. In a room with rigid walls, it can be like that.
But why then do some of you that would use multiple subs to smooth room modes below 100Hz, be completely happy to allow them to exist in the transition region between 100Hz and 150Hz or so? In most rooms, there is a pronounced notch in that region.
Modes in this range sometimes make male voices sound thin, sometimes a little too throaty. Add flanking subs and it immediately smooths out. I notice it the most in rigid hotel rooms (like at shows) but have measured it in other rooms too. Just about every measurement chart posted in this thread shows a deep notch between 100Hz and 150Hz, for example.
So to the "unconvinced" - what exactly are you unconvinced about? The measurements? Or the importance of the lower midrange, where fundamentals of so many instruments begin?
Last edited:
Dan's charts, for example:
Scanning the web with Google, finding indoor response charts show others with the same kind of thing in this range:
A pair of McIntosh speakers, measured indoors:
Charts from RealTraps.com, talking about room modes and boundary interference:
How can anyone be "unconvinced"? Modal smoothing to 100Hz isn't enough. The 100Hz to 200Hz range needs some love too. That's the beginning of the midrange, arguably even more important than the bass.
Scanning the web with Google, finding indoor response charts show others with the same kind of thing in this range:
A pair of McIntosh speakers, measured indoors:
An externally hosted image should be here but it was not working when we last tested it.
Charts from RealTraps.com, talking about room modes and boundary interference:
How can anyone be "unconvinced"? Modal smoothing to 100Hz isn't enough. The 100Hz to 200Hz range needs some love too. That's the beginning of the midrange, arguably even more important than the bass.
- Status
- Not open for further replies.