Terry if it were only so simple.
Floor bounce is a tough one. The amount of absorption provided by a carpet & pad is not very effective overall. Where it is effective is restricted to the top octave or two at best. Very little floor bounce will be reduced with a carpet.
If anyone is interested, have a look at tables of sound absorption coefficients for various materials and implementations.
Last edited:
Studies show that floor bounce is worth dealing with. Binaural hearing lets us discriminate between lateral bounces and the direct sound, our lateral direction sense is strong. But for the floor bounce, poor vertical discrimination means we have a hard time separating it from the direct sound. This results in an audible frequency coloration. The best papers on this were by Soren Bech. He simulated all the reflections of a typical room in an anechoic chamber and found floor bounce highly likely to be audible and detrimental.
Since treatment is difficult the best approach is to design a speaker to deal with it, either a line array or an Allison type approach with the woofer on the floor boundary.
Typical audio effect (say, with a bookshelf speaker on a stand 2-3 meters away) is a big dip at 200 Hz or so that thins out vocal fundamentals.
David S.
What aspect of floor bounce are you referring to ?
There is the first frequency of cancellation, (sometimes called the "suck out" region) which typically occurs in the upper bass / lower midrange somewhere between 150-500Hz as a result of the reflected path being half a wavelength longer than the direct path. This will obviously vary with driver heights, listener height, and listener distance, and will generate a steady state notch in the response, sometimes in excess of 10-15dB, roughly 1/3 octave wide.
(A handy on-line calculator to calculate the frequency can be found here: Floor/Ceiling Reflection Calculator )
The other effect (which is really a different way of looking at the same thing, but is treated a bit differently by the hearing mechanism) is essentially a time delayed duplicate of the speakers direct signal. (Albeit somewhat different in frequency response balance due to the vertical off axis response of the speaker, and absorption of the carpet, if present)
Although the steady state result of this is comb filtering, if the time delay is enough (I forget the figure, but its somewhere around 10ms) then the brain separates this from the direct signal and "ignores" the effect of the delayed signal. This is the same thing it does for early reflections from side-walls, or front wall, as long as they are not too early. (Or much too late for that matter, when they become echoes)
This ability to separate the first arrival from early reflections only happens at higher (~midrange and up) frequencies, at bass frequencies you perceive the steady state result of the time delayed reflection, eg comb filtering - and mainly the first (lowest frequency) notch.
Although higher notches are usually present, they will be reduced by driver directivity, and also start to land in the higher frequency range where the brain can separate the direct and delayed signal - thus go unnoticed even though a microphone measuring steady state response would measure them.
So there are two different things to focus on to try to minimize the perceived effect of the floor bounce:
1) Try to avoid or minimize the steady state notch in the upper bass at the listening position through choice of overall speaker topology, driver placement and crossover frequencies. This generally involves having the low frequency driver fairly low to the floor.
As the low frequency driver is moved closer to the floor the time delay between direct signal and reflection is reduced, pushing the first floor bounce notch higher in frequency. Once it's well beyond the crossover frequency, the problem - at least for that driver, is solved.
Going too low will introduce other issues (listener being too far off axis from the low frequency driver, colouration of the lower midrange by resonances of the floor, too much floor gain which will need compensating for in the network etc) so there is an optimal range of heights which is dependant on the size of the driver, the baffle size, the crossover frequencies etc. Generally you wouldn't want to go lower than is necessary to solve the floor-bounce notch.
2) Try to ensure at midrange and treble frequencies there is sufficient time delay of the reflection to put the reflected signal in the time window where the brain will treat it as a separate arrival, and thus ignore it. This usually involves having the midrange and treble drivers quite high off the floor, and having them both as close as possible to sitting ear level is usually high enough.
Another benefit of having the midrange/tweeter high off the floor is that the angle of floor reflection is much steeper so the natural off axis roll-off of the drivers will reduce the amplitude of the reflection at higher frequencies. Having midrange and treble drivers that are a bit more directional than usual is also beneficial in minimizing the effects of floor bounce in this frequency range.
I'll come right out and say it - a 2 way design (with single mid bass driver) will never solve problems 1 and 2 at once - they are mutually exclusive when a single driver is producing the entire bass and midrange from one point in space.
Put the driver low enough to push the notch up out of the upper bass region and now the midrange will both be coloured by any floor resonances, (such as resonating floor boards) and the time delay of the reflection will be insufficient to allow the brain to separate the delayed image - so they will smear together.
Put the driver high enough to optimize the midrange (and treble, if you want to keep the two drivers as close together as possible) and now you push the cancellation notch down into the upper bass region, causing a "suckout".
A 2.5 way or 3 way system allows for both conflicting problems to be optimized somewhat at the same time.
In a typical narrow 2.5 way system you'll get some reduction in the depth of the notch since you have two drivers working together at different heights, so they both form notches at different frequencies. Where one driver would be causing an exact notch, the other driver "fills in" the response somewhat and visa versa. The result isn't perfect though, because both drivers are still operating through their floor bounce cancellation frequency.
A 3 way design improves matters a bit further, especially if the woofer is physically a lot larger and it's centre significantly displaced from the midrange driver. (By sheer necessity of size if nothing else)
If you're willing to cross over from the woofer fairly high (~300Hz) you can actually conspire to have the woofer low enough to the floor to push it's floor bounce cancellation frequency above the crossover frequency, whilst a high enough midrange driver will put it's floor bounce cancellation frequency just below the crossover frequency.
This means you'll get additional reduction of the apparent dips due to the crossover. Even though the midrange driver is still experiencing a steady state dip near or below the crossover the time delay of it's reflection means the brain largely ignores this.
In short, a large 3 way system with the midrange and tweeter near ear level, (~90-100cm) and the woofer "somewhat" lower (say about 40-50cm from floor to centre) all but eliminates the worst effects of floor bounce, and certainly does away with upper bass "suck-out".
Last edited:
Does floor bounce cause any trouble hearing a live saxophone?
I think it is inevitable to conclude the deader-the-better for rooms with speakers and the more reflection the worse the sound. That's not to say that a Toole-endorsed speaker made for today's typical hard-surfaced rooms would sound good in a dead room.
It also means than speakers that are anechoically flat are going to be pretty deficient real rooms without tweaking.
Makes it sound simpler to create sound systems for large halls and cinemas where all the echos are vastly less prominent than the direct sound.
I think it is inevitable to conclude the deader-the-better for rooms with speakers and the more reflection the worse the sound. That's not to say that a Toole-endorsed speaker made for today's typical hard-surfaced rooms would sound good in a dead room.
It also means than speakers that are anechoically flat are going to be pretty deficient real rooms without tweaking.
Makes it sound simpler to create sound systems for large halls and cinemas where all the echos are vastly less prominent than the direct sound.
Last edited:
Terragon,
I don't believe I said that it was that simple. I believe that I stated that it would help to diminish floor bounce, although the lower the frequencies, the less effect it will have. But it is a start that is easily achieved and may prompt the listener into going further if he achieves a noticeable improvement. If your talking about Bass, then there is very little that will help with floor bounce other than having the woofer next to the floor itself. For a three-way speaker this is much less challenging than a two-way speaker would be, given the frequencies of the woofer/tweeter crossover point and the necessity of close proximity between them. Eigenmodes, within a normal listening room are much harder to deal with and I believe that outside of a correctly installed distributed bass set-up employing 3 or 4 subwoofers, only rarified atmosphere pressure will help.
YMMV.
TerryO
Terry, again have a look at the table of coefficients. A carpet will absorb very little even at higher frequencies. It really does not help much with floor bounce.
The nice thing about absorbents is that they sum everywhere they are - an absorbent tucked away behind a couch or under a coffee table works fine since sound (or noise from the highway outside) bounces around lots before reaching your ears. Likewise, I think wall-to-wall carpeting (with felt underlayer) helps eat low frequencies.
Personally, the higher frequencies would bother me the most. However, I have as a matter of fact, seen plenty of tables with sound absorption coefficients.
My next listening room will use a suspended cargo net for a floor, with a 523.707 foot pit under it filled with foo-foo dust.
Best Regards,
TerryO
No,
Because the brain uses the floor bounce effect as one of the cues to help estimate the height of the instrument, even if we are not consciously aware of it.
However you don't want your entire band to sound like it's coming from the exact same height as your speakers regardless of original instrument heights due to the speakers introducing their own pronounced floor bounce effect in your living room.
If you record a saxophone with a microphone that is positioned roughly at the same height as a real listener would be present, the microphone will record the floor bounce signature in the recording, which includes both the perceived steady state cancellation at upper bass frequencies as well as the overall time delayed broad spectrum reflected signal.
If you play this signal back on a speaker system designed to minimize floor bounce effect (as described in my previous post) the cues present in the recording can take control and provide a sensation of height of the original instruments as each instrument (assuming a live non-nearfield mic setup) will have it's own unique floor-bounce signature encoded in the recording which the brain can interpret.
If the speaker however introduces an obvious floor bounce signature of it's own (particularly a large suck-out in the upper bass) the brain gets a conflicting cue - now it's getting a cue in the lower frequencies that indicate the height of the speakers, so everything in the recording starts to sound like it's coming from the same height as the speakers.
I think this is why some speakers, particularly larger ones can reproduce a sense of vertical "space" while others can't, and only produce width and depth.
Of course it's highly dependant on the recording - a close mic'ed heavily processed pan-potted recording has no inherent floor-bounce information encoded in the recording, so it would only apply to live recordings made at a normal "listening distance" where such effects could develop to be recorded in the first place.
DBMandrake - Could you please annotate your previous post by indicating parts that are evidence-based and parts that are your personal expectations of how things ought to work in a perfect world.
When I am being careful, I remember to use terms like "I guess" or "maybe" to suggest when some assertion is... ah... just that.
When I am being careful, I remember to use terms like "I guess" or "maybe" to suggest when some assertion is... ah... just that.
Good old Snell design. Tonally, I found it nice to listen to. Is it accurate? Well, if you believe that the recording engineer got it "right" and if his environment had the floor bounce, then so should yours.
Paul Barton's alluded to use of the line of woofers technique in his PSB designs, to reduce the detrimental effect of the floor bounce. It also benefits making the cabinet narrower and less visually intrusive.
If you listen to smaller scale music, the floor bounce provides cues that the sound is generated in your room, and in this case that may be a good thing for you. If however you want to give the ilussion that you are in the recording space, then getting rid of the floor bounce and early reflections is a worthy goal.
The discussion over the last few pages has centered on perceived frequency response impacts due to early reflections, but Olive showed its helpful to have some local physical reflections to help detect the size of the space the recording itself was made in. Now, if you could provide these (diffuse) reflections, time delayed enough to not impact perceived frequency response but still early enough to help detect he space encoded in the recording, then it may be possible to start to have the room "disappear". I don't know if Olive looked at how long the reflections could be held off while still supporting this trick (I'd have to reread the paper).
Finally, there is spaciousness to think about. How far off can the delay of the first reflections be held without overly reducing the sense of spaciousness? That info is buried in Gresinger's papers.
Dave Dal Farra
Good and once again, subtle.
Are you saying Olive (or Toole??) showed these reflections provided a verifiably accurate perception of the recording space or just a good-feeling about the ambiance of the recording space?
Are you saying Olive (or Toole??) showed these reflections provided a verifiably accurate perception of the recording space or just a good-feeling about the ambiance of the recording space?
I don't think that can be done with two speakers in a small domestic room but, with more speakers - sure, why not?
Frankly I’m a little bemused by all the “early reflection” and “floor bounce” arguments . . . it’s a *lot* more complicated than drawing a couple rays from the speaker to the listener. For a start you have to draw *lots* of rays . . .
When I seat our orchestra they are on a floor. They (and the audience) much prefer if I can put a shell as close as possible behind them. The instruments are all at pretty much the same level when played. And the audience hears *far* more reflected than direct sound (except, of course, at our outdoor concerts . . . whole different ball-o-wax there). Reflections, including early reflections, are not necessarily detrimental to the sound . . . at least some are clearly preferred and desired. I’m nowhere near as convinced as some here that early reflections in the listening room, blended as of necessity they will be with the early and late reflections captured in the recording, are horribly detrimental to the perceived sound. In fact in the case of “close mic’ed” recordings (which all are relative to normal audience seating positions) reflections in the listening environment may result in a more natural sounding presentation.
Add to that the even more confounding ability of the ear to “tune out” familiar “room effects”, and a lot of the “authoritative” statements made here about measurements and response just don’t translate very well to what we actually hear. And the notion that we can somehow generate a realistic “vertical image” from the sound captured by microphones 20 feet in the air is pure fantasy.
When I seat our orchestra they are on a floor. They (and the audience) much prefer if I can put a shell as close as possible behind them. The instruments are all at pretty much the same level when played. And the audience hears *far* more reflected than direct sound (except, of course, at our outdoor concerts . . . whole different ball-o-wax there). Reflections, including early reflections, are not necessarily detrimental to the sound . . . at least some are clearly preferred and desired. I’m nowhere near as convinced as some here that early reflections in the listening room, blended as of necessity they will be with the early and late reflections captured in the recording, are horribly detrimental to the perceived sound. In fact in the case of “close mic’ed” recordings (which all are relative to normal audience seating positions) reflections in the listening environment may result in a more natural sounding presentation.
Add to that the even more confounding ability of the ear to “tune out” familiar “room effects”, and a lot of the “authoritative” statements made here about measurements and response just don’t translate very well to what we actually hear. And the notion that we can somehow generate a realistic “vertical image” from the sound captured by microphones 20 feet in the air is pure fantasy.
FURTHER INVESTIGATION INTO THE LOUDNESS OF RUNNING REVERBERATION
D. Griesinger
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.40.7215&rep=rep1&type=pdf
D. Griesinger
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.40.7215&rep=rep1&type=pdf
Ben, I know Dr. Toole says early reflections expose more detail about the recording--and that's great! It's the whole you are there vs. they are here argument of yesteryear. The problem is that they also come with timbre aberrations and source broadening(which could be read as image smearing). They also expose off axis speaker problems to a great degree. Not a lot of speakers have great off axis behavior. I actually think that the things I 'hear' from an RT20,30,or 60 are more early reflection or local issues.
Pick your poison, but you can have accurate imaging and envelopment. It will cost you.
Dan
Pick your poison, but you can have accurate imaging and envelopment. It will cost you.
Dan
Last edited:
Think theres 3 differnt subjects here..
1) The reasons and understanding why speakers can't be flat
2) how to tune the listening environment
3) How is human perception in correlation with measurements
There may be many ideas and tools that can adjust and change some elements.. but the fundamental understanding of the first subject is essential... I fact so much that if you get the first one really right the latter seems to loose their importance...or at least be reduced to have a few plants and place a rug on the floor...
1) The reasons and understanding why speakers can't be flat
2) how to tune the listening environment
3) How is human perception in correlation with measurements
There may be many ideas and tools that can adjust and change some elements.. but the fundamental understanding of the first subject is essential... I fact so much that if you get the first one really right the latter seems to loose their importance...or at least be reduced to have a few plants and place a rug on the floor...
Well I'm not going to play that game, as I'm just offering my comments "as is", so take them or leave them, but I will say this.
I first noticed over 10 years ago that certain speaker designs which included having their woofer very close to the floor (and a few other characteristics) could on some recordings provide an almost eerily realistic sense of depth, height, and "space". Perhaps not strictly accurate, and maybe somewhat "larger than life", but the effect was there - the actual position and height of the speakers seemed to melt away leaving a sense of space that had a height component as well as depth and width. As I say, on some recordings the effect was not apparent.
Systems that had their woofer much higher off the floor or were 2 way seemed largely unable to do this. I suspect it might be related to floor bounce cancellation in the upper bass but wasn't sure.
A few years later when experimenting with a design I had a configuration that had 12" woofer, 8" full range (each in separate stacked cabinets) and ribbon tweeter, and I was able to switch it in many different configurations - 3 way, 2 way with just the full-range and tweeter, 2 way with the woofer and full range, or full range only. One of the things this allowed me to explore was the effect of floor bounce in the upper bass, as without moving any speakers I could switch from 2 way (with a response good down to 35Hz in room) to 3 way, with a lower woofer.
At the listening position floor bounce cancellation was very obvious in the 2 way mode and non-existent in 3 way mode. No matter how I EQ'ed the 2 way system - to give the same near field bass response as the 3 way system or otherwise - the sense of space and depth and "being there" was no-where near as convincing as the 3 way.
The only difference between the two configurations in the controlled tests was the height and directivity of the woofer, so I again I put it down to the floor bounce cancellation notch, although at the time I didn't understand how floor-bounce might affect a perception of height.
Fast forward to a couple of months ago and I came across an article that discussed how floor bounce effect was one cue used by the human hearing system to estimate the height of objects, and that these cues could be captured during the recording process, and suddenly my observations from years ago clicked into place.
I'm struggling to remember the name of the article or who it was by, but if I can track it down I'll post it.
- Status
- Not open for further replies.