Actually, the direct sound + early reflections sounds best to my ears in this example!
Here's the link again:
Precomputed Wave Simulation for Real-time Sound Propagation of Dynamic Sources in Complex Scenes - YouTube
Tom, thanks for you contribution.
I have to say that the several times I set up my stereo outdoors I didn't notice the effect you describe. The first several times I did it, I was struck by the cleanness of the sound. The last time I noticed that the floorbounce had a great audible effect! This is probably because it is the only significant reflection.
BTW, your recording of the fireworks sounds amazing! The other ones are nice too, though.
I have to say that the several times I set up my stereo outdoors I didn't notice the effect you describe. The first several times I did it, I was struck by the cleanness of the sound. The last time I noticed that the floorbounce had a great audible effect! This is probably because it is the only significant reflection.
BTW, your recording of the fireworks sounds amazing! The other ones are nice too, though.
Certainly the 7 mSec arrival is more even in level except for some interference at 500Hz. To claim that the speaker is improved by a pillow blocking the direct sound isn't something I would agree with. The pillow only partially blocks the highest frequencies, only about 10dB below 1000Hz. Plus the 15 mSec arrival is rougher than the direct arrival.
I'd guess the system sounds like a speaker with a pillow in front of it.
Otherwise, I'm not usually afraid of a little bit of group delay but you are forcing an 8mSec delay between direct arriving low frequencies and wall bounced high frequencies.
And why are your wall bounces stronger than the direct response, even without the pillow?
David S.
There is a school of thought (not mine) that believes floor bounce is not such a problem because your brain expects it. It is part of evolution that tells us we are near the ground versus being perched on a tree branch.
Comb-filtering because of the floor bounce may be a real problem, but is the reflected sound really a problem for localization?
Comb-filtering because of the floor bounce may be a real problem, but is the reflected sound really a problem for localization?
Hi,
It may sound good sometimes even best, but it is against the modern research and thus you must be wrong with your opinion
I do research myself and I must say that many of modern researchers are totally lacking the big picture.
- Elias
It may sound good sometimes even best, but it is against the modern research and thus you must be wrong with your opinion
I do research myself and I must say that many of modern researchers are totally lacking the big picture.
- Elias
By "controlled directivity" I meant more narrow. Less likely to involve room acoustics any more than is necessary. I don't know of a situation where comb filtering is desirable in hi-fi reproduction. It's a fact of life (law of physics) unless you're floating in space more than 75 feet from anything, or in an anechoic chamber, that you would be wise to fully understand and deal with somehow. Maybe you could explain yourself in much more detail. I'm confident that you misunderstood what Toole was saying, or I just don't get what you are saying.
And I like the sound and feel of the Harley being started in my room, I could almost smell the exchaust fumes
@Tom
Besides the 'broadband CD point source' characteristic there are two other aspects of the SH50 which are not found in normal indoor hifi speakers:
- very narrow directivity (50 deg beamwidth)
- the lack of any baffle which probably reduces the amount of diffraction
Do you consider the second aspect as an important property for a loudspeaker?
@Tom
Besides the 'broadband CD point source' characteristic there are two other aspects of the SH50 which are not found in normal indoor hifi speakers:
- very narrow directivity (50 deg beamwidth)
- the lack of any baffle which probably reduces the amount of diffraction
Do you consider the second aspect as an important property for a loudspeaker?
Any reflections that are less than 1 mS can affect perceived image location on the horizontal axis dramatically. Proof of that pudding is demonstrated by the fact that the Carver Holographic generator uses 125 micro seconds of delay. Above about 1mS (milli second) the way room reflections affect image location is more about amplitude comparisons, part because your inter-aural crosstalk steps on imaging information in the lower midrange, and part because all the comb filter FR fluxuations will cause amplitude comparison analysis to be all over the place, depending on the exact location of your ears.
A good acoustician knows that you don't just separate the delay effects into 2 catagories. There are different psycho-acoustic effects for several different delay ranges (4 or 5 that I've read about and experimented with). According to David Griesinger, pioneer of digital reverb and long time senior consultant for Lexicon, voice intelligibility is most damaged by delays between about 50mS and 150mS. A 30mS delay might even make the voice easier to understand. Wider dispersion does the opposite of what you said; it increases room reflections, and the problems (or opportunities) that come with that. Delayed energies of less than about 6mS are never a good thing. Open baffle or dipole people would know this real well. That's why dipoles sound way better if they're at least 3 ft. out from any walls.
The good news is that on anything but the left to right axis, the only way we can sense image location is by reverb analysis, or in the case of changing sounds, consistent changes in upper-midrange frequency response due to how the pinea works (I'm trying to say a lot in very few words here). So I agree that floor bounce could be an important part of our ability to sense depth and height. .
The bad news is that the floor bounce information that we use to perceive depth and height, might already be embedded in the recording, and so now you've got two floor bounce "echos" and the ear-brain mechanism may get confused by that, just as it does with two sets of interaural crosstalk information when you don't have some variation of the Carver interaural crosstalk cancelling mechanism. In the real world, it may be pretty rare that any floor bounce info is in a recording (it seems like most recordings are done with individual close mics, or mic trees hung from the ceiling for most classical music stuff), so I generally agree that floor bounce is not likely your biggest problem. It could even be a good thing if it makes a band sound like they are on solid ground rather than floating in the cosmos.
This point was debated quite a bit in the "Flat is not correct for Stereo?" mega-thread a few months ago, and I'm not sure that there was any real conclusion, just conjecture on both sides.
Not wanting to rehash that discussion, but if we assume for a moment that comb filtering related to floor bounce could be used as a cue that helps us recognise the approximate height (maybe even depth) of a sound source then the question becomes, do you want to localize all reproduced sounds to the same location as the speakers, (along a line between the left/right drivers) or do you only want to hear height/depth cues that may be encoded in the recording ? (Assuming some sort of live recording rather than an artificial pan-potted multi-mono close-mic recording)
My position on this was and is that comb filtering from floor bounce is something that is not in the recording (at least not the same floor bounce) which is being overlaid on everything you play back, and is therefore a type of colouration, and in the upper bass/lower midrange it is most definitely a tonal colouration if nothing else.
Taking steps to eliminate or reduce the floor bounce (such as very low woofer + high somewhat directional midrange/tweeter) makes sense in that it removes one spurious cue that might otherwise tend to localize the height or depth of the sound to the same location as the speakers. A cue that may conflict with cues embedded in the recording, much like a very reverberant room will conflict with and often override reverberation embedded within a recording.
It's certainly very anecdotal, and only my opinion, but my experience comparing speakers that address the floor bounce notches between 100-300Hz with those that don't is that the ones with minimal floor bounce cancellation tend to have a greater sense of depth and "scale" to the sound field on some recordings, with a sound field that has depth that extends from the speakers to well behind them, even behind the wall behind them.
On the other hand a typical 2 way with a high midbass driver and a large discrete notch from the floor bounce tends to have a fairly flat depth perception - as if the sound comes from a flat plane at the same depth/distance as the speakers. You might have good width, some sense of height variations, but not much depth.
Even if we don't know for sure what's happening, I say better safe than sorry. The on topic part of this detour is that directivity (+ driver placement) is an important part of minimizing floor bounce.
Last edited:
From my experimentation with a pair of KEF Q15s in different orientations and positions in a rather small, live room I gathered I need reflections to create a believable soundstage (I can place them at least 1m from the walls).
Front wall reflections provide a believable depth, otherwise I can hear depth, but it's somehow stuck at (originates from) the position of the speakers. Same with lack of side wall reflections, soundstage width is there but it sounds too., "artificial".
It's the combination of the two (from my big speakers), but mostly the side wall reflections, which has given me an enveloping 3D soundstage.
Having the Q15s located on the floor and angled upward I mostly noticed no difference in height compared to the big speakers (accountable to quality differences) so I'm not quite convinced about the impact of floor bounce on height perception.
Sitting behind my laptop in the otherwise dark listening to Dead Can Dance I can't really tell the location of the speakers. I still have to build the "omni-directional" floor cubes but I think they'll be able to create a believable soundstage while being invisible.
Front wall reflections provide a believable depth, otherwise I can hear depth, but it's somehow stuck at (originates from) the position of the speakers. Same with lack of side wall reflections, soundstage width is there but it sounds too., "artificial".
It's the combination of the two (from my big speakers), but mostly the side wall reflections, which has given me an enveloping 3D soundstage.
Having the Q15s located on the floor and angled upward I mostly noticed no difference in height compared to the big speakers (accountable to quality differences) so I'm not quite convinced about the impact of floor bounce on height perception.
Sitting behind my laptop in the otherwise dark listening to Dead Can Dance I can't really tell the location of the speakers. I still have to build the "omni-directional" floor cubes but I think they'll be able to create a believable soundstage while being invisible.
I have a commercial 2-way bookshelf that does stellar imaging, especially depth. Also, on a solo cello recording done in a reverberant space, played through an Altec VOT, the cellist can be located firmly about 20 rows behind the front wall. The VOT system has the woofer way above the floor, and it doesn't have any floor bounce mitigation in place. These are just two instances where I've experienced stunning depth perception despite the presence of floor bounce.
So, while I understand the theory behind eliminating floor bounce, I have heard speakers behave the exact opposite in practice.
So, while I understand the theory behind eliminating floor bounce, I have heard speakers behave the exact opposite in practice.
The measure for vocal intelligibility compares energy from 0 to 30 msec vs. energy from 30 msec on. More early and less late leads to greater intelligibility. As a "good acoustician" I've worked out the reflective ceiling design for Harvard style lecture halls based on those numbers. Concert hall clarity, I believe, uses a 50 msec border.
Huh? Reread what I've said. Sitting off axis or having wider dispersion will "reduce the level of direct vs. all subsequent reflections" i.e. will increase the reverberant field re. the direct sound, in relative terms.
David S.
Some papers i found to be interesting concerning perceptional aspects of group delay
Blauert, J.; Laws, P.: Group Delay Distortions in Electroacoustical Systems. J. Acoust. Soc. Am., Vol. 63, No. 5, 1978, pp. 1478-1483.
http://community.klipsch.com/forums/storage/3/1027021/7805blauert.pdf
Goossens, S.: Wahrnehmbarkeit von Phasenverzerrungen. IRT München.
http://forum2.magnetofon.de/bildupload/goosphase.pdf
Paper is in german language, but Table 1 (upper right) gives an overview
concerning detection thresholds measured by different authors.
Lee, L.W.; Geddes E.R.: Audibility of Linear Distortion with Variations in Sound Pressure Level and Group Delay. AES Convention 2006.
http://www.gedlee.com/downloads/AES06Gedlee_ll.pdf
Schroeder; Mehgart: Auditory masking in the perception of speech.
(No online source found.)
Last edited:
Bold emphasised of what I actually said.
Your data: Earphone tests, anechoic conditions, 25us rectangular pulses etc...
Give me a break !
Your data: Earphone tests, anechoic conditions, 25us rectangular pulses etc...
Give me a break !
Surely under reverberant conditions you will get differning results
for many aspects according to the (especially early) reflections present.
What conclusion do we draw ?
It's an acoustical small room anyhow, lets mess transients up ?
Getting transients as good as possible and trying to have the reverberant
field sounding like a somewhat larger room when listening close
(by finding a suitable balance between a pronounced ITD gap in level
and duration and sufficient diffusivity of the reverb) seems more attractive
to me.
To achieve that without any room treatment seems hard i guess,
but my goal would be to achieve it with modest room treatment, that
can be integrated well in a living room.
for many aspects according to the (especially early) reflections present.
What conclusion do we draw ?
It's an acoustical small room anyhow, lets mess transients up ?
Getting transients as good as possible and trying to have the reverberant
field sounding like a somewhat larger room when listening close
(by finding a suitable balance between a pronounced ITD gap in level
and duration and sufficient diffusivity of the reverb) seems more attractive
to me.
To achieve that without any room treatment seems hard i guess,
but my goal would be to achieve it with modest room treatment, that
can be integrated well in a living room.
Last edited:
that our hobby is about realistic sound reproduction at home to our satisfaction and not to satisfaction of microphones and scientific theories based on piecemeal and mostly irrelevant research?
Your data: Earphone tests, anechoic conditions, 25us rectangular pulses etc...
Give me a break !
exactly!
Last edited:
I was never statisfied with systems that mess up transients too
much. Acceptable reproduction of transients is essential to me.
A (loudspeaker/room) system that gets the transients well, makes me
"lock in" to the music.
If a system is good at that and has low coloration i am very happy when
finding i am not restricted to a narrow sweetspot to get that.
And if overall reverberant characteristics offers envolvement i am
perfectly happy.
But to me there is a certain order in these desirable properties.
I personally cannot follow the reverse order like:
"Hey that's spacey and involving, but transients are smeared and
voices and strings sound somewhat colored."
The "luxury items" have to be based on a good foundation to really get
noticed.
much. Acceptable reproduction of transients is essential to me.
A (loudspeaker/room) system that gets the transients well, makes me
"lock in" to the music.
If a system is good at that and has low coloration i am very happy when
finding i am not restricted to a narrow sweetspot to get that.
And if overall reverberant characteristics offers envolvement i am
perfectly happy.
But to me there is a certain order in these desirable properties.
I personally cannot follow the reverse order like:
"Hey that's spacey and involving, but transients are smeared and
voices and strings sound somewhat colored."
The "luxury items" have to be based on a good foundation to really get
noticed.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.