A 3" thick futon covered by a rug is hardly just a rug on the floor, and doesn't quite match your initial claim that "just using a thick rug at just the right place" will do the trick 😉
How much of the lower frequency absorption is from the rug, and how much from the mechanical losses of the futon itself ? Not to mention that spacing the rug 3" away from the floor would greatly improve it's absorption at lower (low midrange) frequencies than the rug lying flat on the floor, as it puts the rug closer to velocity nodes at the frequencies of interest instead of being at a pressure boundary. Standard acoustic absorption theory.
From a practical point of view almost nobody is going to throw a rug over a 3" thick futon in front of both speakers, so it's a bit misleading to suggest that throwing a rug on the floor somehow magically fixes floor bounce cancellation when that's not what you're actually doing.
I have to wonder why you're finding it necessary to take such extreme measures ? Perhaps a speaker that was designed with free field measurements in mind inevitably having it's response altered by being near the floor ?
Don't get me wrong - I think having rugs in front of the speakers (or at the very least dense carpet) is a good idea - it's quite effective at higher frequencies, but I don't expect it to be effective at the lowest floor bounce cancellation frequencies, which is what's being discussed here.
Until you trip over it 😉
I doubt whether it's that effective at all at 100-200Hz, since the notch there is occurring from a vertical bounce rather than one angled towards the listener, as well as the limited effect of the materials at that frequency.
The notch in this range is the most severe one of those being discussed, (in my particular room it's 8-10 dB) and it doesn't really address that at all, whereas a low mounted properly crossed over woofer will fix it entirely with no elaborate modifications to the room.
Just because the vertical location of the driver operating in the floor bounce frequency range (100-300Hz) doesn't affect vertical image localization doesn't make cancellation in this region unimportant such that it can be ignored. It still affects depth imaging as well as tonal balance. It's just as important but for other reasons.
So what is the effect of a port placed near the floor? Will it fill this notch, or make it worse?
No difference, if the port is tuned under 100Hz (as most are). When speakers are elevated to ear level, the floor bounce notch is most severe between 100Hz and 150Hz. To smooth it, you can blend a couple sound sources in this range, one a little lower in height than the other. One way to do that is to put a woofer closer to the floor and blend it with the mains up to ~200Hz. Naturally you'll want a speaker that's clean at this frequency, but most are, even subs.
Absolutely. Response irregularities in this region are very noticeable. The 100Hz to 5kHz is the range we're most sensitive to. So while I think bass smoothness is also very important, I would suggest that the floor bounce notch be at least as high on the list of priority as lower frequency room modes.
Subjectively, bass anomalies and lower midrange anomalies sound a little different to me. Lumpy bass is most noticeable in bass note progressions. They go through a sort of loudness roller coaster, up and down in volume as the note progression goes through the scale. Lower midrange is sometimes the same, for example, instruments like piano will betray a floor bounce notch. But vocals and many instruments have rich harmonics that often mask the floor bounce notch.
When listening to vocals and instruments with rich harmonics, the average loudness doesn't seem to change when the fundamental of a note progression passes through the floor bounce notch frequency. Instead, what I notice is a throatiness in male voices when there is a peak in the lower midrange, or a thinness when there's a big notch. Sometimes this thinning of the lower range from a floor bounce notch can make female vocals sound a little strident, even shrill.
One of the things that bothered me in the early multisub discussions (debates) on AudioRoundTable.com was the fact that this lower midrange transition region wasn't addressed by everyone. I think all agreed that the bass range under 100Hz could be improved with multiple sound sources but not everyone agreed that the 100Hz to 150Hz range was equally important.
- Midbass/Midrange Blending in the "Speaker Placement - The Subwoofer Thing" thread
I think most agree that absorbent materials are not very effective at 100Hz to 200Hz. It's a hard range to deal with. You can damp the bass with large panels, even drywall damps it somewhat. Acoustic insulation is fine at the higher frequencies. But the lower midrange seems harder to tame. I think source blending here is very important. Flanking subs or woofer/midrange blending in the 100Hz to 200Hz range is the best approach, in my opinion.
Dr. Geddes, I'm a bit disappointed that you accept this as evidence. Now it looks like you'll accept 'proof' if it suits your standpoint. I have listened to Dave's MP3's and although the effect does indeed sound rather awful, "It doesn't prove anything!". What it does do, is give us an idea of what might happen to the sound, but you will have to do real life experimenting to get proof that it does indeed deteriorate the sound quality to a significant extend. Daves could probably just as easily 'prove' that for example late reflections mess up the sound - even though it is widely acknowledged that they improve spaciousness.
While I'm reading more of the discussion, I see other people already put forward the same issue.
You completely missed the point. I neither accepted it nor rejected it, in fact I never even listened to it. My point was to put this demo on an equal footing with the kinds of evidence offered up from the other side, which is largely circumstantial and subjective and also didn't "prove" anything. The demo was a singular piece of evidence thats all - and it is "evidence", even if not conclusive. But, to me, it was far more substantial than the usual "Well it sounds good to me" argument put forward on the other side.
Are you sure about these frequencies? Say you have a dip right at 125 hz. That corresponds to a pathlength difference of 1.37 meters. I don't see that happening with a floor reflection. I'd expect a floor bounce notch close to 300 hz in most situations.
Scroll back a few pages to posts #1438, #1451, and #1454 where this was being discussed, that should answer your question 🙂
I'm sure about those frequencies. The strongest notch you'll measure from floor bounce is the one that is caused from the reflection straight down, not the one from path length difference. This is also true for the notch from the wall behind the speakers.
What do you measure if the height of the woofer off the floor (lets say 0.8m for a stand mounted 2 way) is the same as the distance from the woofer to the front wall and/or side wall - do the notches all add up and give a horrendously deep hole, even though the notches are forming from 3 different physical axes relative to the listener ?
Last edited:
It makes just one notch, as you might expect. The notch is pretty deep in either case, whether just a floor bounce (like measured outdoors) or when the wall behind the speakers is the same distance as the floor. I've measured this effect outdoors - source off the ground with wall behind - because indoors there is too much other stuff going on to isolate individual boundaries.
These are reasons why I like putting the speakers very near boundaries, when possible. The constant directivity cornerhorn approach does this nicely. Otherwise, like when using two-ways on stands, I usually use flanking subs with low-pass near the notch frequency. Blending with the mains mitigates the notch, and the relatively close proximity prevents localization. This approach can be incorporated with more distant subs in a multisub approach, although the distant subs should obviously be low-passed at a lower frequency.
These are reasons why I like putting the speakers very near boundaries, when possible. The constant directivity cornerhorn approach does this nicely. Otherwise, like when using two-ways on stands, I usually use flanking subs with low-pass near the notch frequency. Blending with the mains mitigates the notch, and the relatively close proximity prevents localization. This approach can be incorporated with more distant subs in a multisub approach, although the distant subs should obviously be low-passed at a lower frequency.
Last edited:
But what can cause the lower frequency 1/4 wavelength notch outdoors though ?
I've been playing around with the FRD consortium room response calculator, which is based on modelling large (configurable) numbers of room reflection rays, and if I set the reflection coefficient of all walls and ceiling to 0, leaving only the floor set to 0.7 the calculated response shows the expected "floor bounce" comb filtering which is due to distance differential between direct and reflected signal, but it doesn't show any notch at the lower frequency 1/4 wavelength height.
This makes sense from a purely direct/reflected interference stand point, because if you think about it on an infinite ground plane there are only two rays that can reach the listener - direct and reflected at one specific angle toward the listener - classic floor bounce.
However if a notch is still occurring at 1/4 wavelength above the ground plane in actual measurements something else must be going on. Simplistic reflection modelling assumes that the boundary reflections don't affect the actual output of the speaker, but this seems unlikely.
Perhaps when the speaker is 1/4 wavelength off the ground (and thus the returned reflection is 1/2 wavelength out of phase) the effect of the destructive interference is to increase the loading on the speaker at that frequency and thus reduce it's excursion and therefore output in all directions ?
The problem with this theory is such an effect should be obviously measurable in near-field bass measurements and I'm not sure that I've noticed it, although I have sometimes noticed anomalies in the 100-200Hz range in near-field measurements which I assumed were cabinet panel modes, but didn't try altering the height of the speaker to see if they moved in frequency...
Assuming this is actually happening it means different speakers will respond differently to this effect, depending on how sensitive their alignment is to changes in acoustic loading. For example a highly sensitive woofer with light cone, large Vas from very soft suspension in a large box could see more reduction in output than a woofer with a very heavy cone and inefficient alignment in a small box, while a servo feedback system would see no change. (assuming the feedback could go that high in frequency)
Has anyone looked into this possibility ?
Last edited:
Thanks!
I don't really understand. Sound that travels straight down from the woofer will move straight up after reflecting against the floor. I wonder why you would assume it to diffract and cause a null at the listening position. However, I don't question your and DBMandrake's measurement. Especially since adding a woofer near the floor made things look better, I think that the dip might be caused by the second order standing wave between floor and ceiling. What do you think?
EDIT: a couple new messages were posted while I was reading older posts and typing this one. Now I see the dip is also present when measuring outside. That makes my suggestion moot.
Last edited:
I measure several ways outdoors. My usual course is to use the ground plane "pit" method, which I then extend a false baffle to meet the loudspeaker baffle. Sometimes I just get lazy and lie the speaker on its back, outdoors on the ground, which is obviously a little different. It's useful though, because the pit method removes any baffle directivity shift, i.e. baffle step.
Another thing I've done is to raise the speaker off the ground with a stand. This shows the floor bounce notch. And I've also pushed the speaker back against a brick wall at various distances. And lastly, into a trihedral corner. All this, outdoors.
When indoors, room modes come into play. It's useful to know how they act too, obviously, but each room is different in frequency and severity of the modes. Outdoors measurements tell me what is happening without room modes, and let me isolate each reflection independently.
The thing is, the notch from floor bounce and the nearest wall get very close to the modal region, many times even falling below the Schroeder frequency. So the techniques effective at smoothing room modes sort of apply here too. It's just a matter of acoustic scale. The transition region between the modal and reverberent range have some interesting problems to solve. Baffle directivity often shifts in this region, and floor and nearest wall reflections also happen here.
Another thing I've done is to raise the speaker off the ground with a stand. This shows the floor bounce notch. And I've also pushed the speaker back against a brick wall at various distances. And lastly, into a trihedral corner. All this, outdoors.
When indoors, room modes come into play. It's useful to know how they act too, obviously, but each room is different in frequency and severity of the modes. Outdoors measurements tell me what is happening without room modes, and let me isolate each reflection independently.
The thing is, the notch from floor bounce and the nearest wall get very close to the modal region, many times even falling below the Schroeder frequency. So the techniques effective at smoothing room modes sort of apply here too. It's just a matter of acoustic scale. The transition region between the modal and reverberent range have some interesting problems to solve. Baffle directivity often shifts in this region, and floor and nearest wall reflections also happen here.
I think you have a good point about the second order standing wave from floor to ceiling - it's something that is very likely having an effect in room, and something I've given thought to over the years.
Many ceilings are 2.4 metres (8 feet) high, and the 2nd order mode for that occurs at 143Hz.
Points of maximum cancellation would be heights of 0.6m and 1.8m, with the first being relatively near typical woofer heights.
I'd actually always assumed this was the main reason for the notch between 100-200Hz, and as I've never tried to measure this outdoors, I didn't really think twice about it.
An easy way to separate the two effects though, is to vary the height from floor level up to 1m or so and watch the frequency where the notch occurs.
If it's only boundary cancellation the notch depth will stay similar but reduce in frequency as the height rises, whereas if it's due to vertical room mode excitation the frequency of the notch will stay the same but the depth will increase to a maximum at the 0.6 and 1.8m heights.
Possibly both effects will be present at the same time, with the notches coinciding at certain heights.
Just yet another good reason to put the woofer close to the floor 🙂
Absolutely. This is why I like to use a couple sound sources in the 100-200Hz range, similar to the multisub approach but a little bit closer together to be in the right acoustic scale.
When I measure outside, I usually do two things: 1) put the speaker as high above the floor as possible and do a gated measurement, and 2) put the speaker on the floor and do a ground-plane measurement. Next time I take my stuff outside I have a third measurement to perform!
thats because there isn't one, at least not from the floor. The claims of 1/4 wave cancellation are incorrect. Its 1/2 wave path length differences that cause the notch and nothing else.
That asumption is correct. There is no effect on the speaker from a reflection. At least not the floor one.
And to a point that I made earlier about using a rug. The floor bounce is by its nature at a grazing incidence and this allow for a much greater absorption effect from a given surface. There is even an angle of "coincidence" which can actually have 100% absorption, although this effect, in general varies with frequency. Its the high absorption at grazing incidence that makes the groud plan measurements so suspect. Because even apparently hard surfaces can have substantial sbsorption at grazing over long distances.
Last edited:
Dr. Geddes, what are your thoughts on the dip between 100 and 200 hz Wayne Parham measures even outside. Have you ever measured this floor bounce dip? The explanation of the vertical reflection that somehow messes up things at a couple meters distance seems a bit improbable to me.
- Status
- Not open for further replies.