Member
Joined 2009
The 10ms time gap between the direct signal and the first reflection was already mentioned earlier in this thread. Linkwitz suggests >6ms. In reality that translates to a big listening room with a pair of speakers in the middle of it. I think many will agree that this is a hard to accomplish luxury.
I have a wide dispersion system and I get satisfactory results with a time gap of just 3ms. I don't claim my setup is optimal and there's certainly limits of how loud I can play until room reverberation gets in the way. All in all I get a very good sound and imaging.
My point is that there might be something else going on and the ETC (energy time curve) might not be the single defining factor. Furthermore, ETC is given in 2 dimensions and I think we dealing with 3 variables - time, intensity and frequency. Also I speculate that the ETC will be relative and not imperative as Dave suggests in the previous post. For example if the room is bigger it can be more reflective, if it's more absorbent it can be smaller.
Attachments
Last edited:
I would never claim that you can't get satisfactory results, any more than I would tell someone that dipoles can't work, or omnis are never any good. The question is whether you might get some improvement with a more optimal dispersion speaker or whether pushing your reflection at 3ms down in level or further back in time wouldn't make a noticable improvement.
The first hurdle to answering that, is to answer "is the reflection audible?" Bech studied that by simulating (in an anechoic chamber with lots of surrounding speaker) a typical speaker in a typical room and turning the level of each reflection up and down relative to its typical level.
Some of what he found: "The results have confirmed the finding (of his previous paper) that the first floor reflection will contribute on an individual basis to the timbre of a noise signal." "...an increase in the level of individual reflections for a noise signal is most likely to be audible for the first order floor reflection and for reflections from the wall to the left of the listener (the near wall). The experiment has further shown that the first order reflection from the wall behind the listener also belongs to this group. For a speech signal only the first order floor reflection is most likely to produce an audible effect."
So the floor bounce is always an issue. The side and back wall can also be troublesome. This suggests to me that an Allison approach to deal with the floor bounce, along with some upper range directivity will help with the floor bounce. Aiming speakers to miss the floor and sidewalls at mid frequencies (such as lots of toe in and maybe an upwards tilt) should help plus diffusion or absorption at those particular points.
Note that another interesting part of Bech's study was that he had improved the accuracy of his simulations by factoring in a particular speakers frequency response at the angle of every reflecting ray. At some reflection angles the speaker had a dipped out response due to crossover interference and driver directivity. This lead to lower detection (a good thing). So the notion that power response dips will hurt, or that every reflection needs the same spectral response as the direct response may not be true.
David S.
A short answer to the original question would be, it depends on how the speakers are going to interact with the room acoustics. In a really bad room I might argue that controlled directivity is likely to be desirable. In a pretty balanced and good sounding room, wide dispersion is probably a better choice.
The long answer involves understanding the psycho-acoustic effects of various colorations and delayed (reflected) signals. How does our perception change over frequency? Why and where does decorrelation have its place? What does each reflection do to the final result and why?
I've read and am convinced that we detect image location in the lower midrange by timing comparisons even more than amplitude comparisons. So a vertical array of five inch drivers for example might smear latteral timing info minimally (that was my choice and no regrets). A vertical array also has a rolled off high frequency response at the ceiling and floor, so you've minimized floor and ceiling bounce to the degree that is practical. I'm also convinced that we detect image location in the upper midrange (roughly 1k - 6kHZ ) primarily by amplitude comparisons, so we want identical frequency responses to be delivered by each speaker. Since there will be many room reflections in the real world, each creating comb filter damage, but likely different on one side than the other, perhaps the additional reflections created by a vertical line array would be one of the best optimizations. Each additional reflected signal has a certain probablity of filling in the cancellations of any other combination of direct signal mixing with a delayed (reflected) signal. I've read that above about 5kHZ, we don't have nearly the sensitivity to exact image location. That range seems to be more of a coloration of what is below 5kHZ, and because we perceive frequencies above about 2kHZ as sounding like they are soming from higher up physically even when they aren't, you could say those frequencies put "air" or height into the soundfield . If one tweeter on each side gives you significant nodal lines (and it should, especially with mono or center images), then it seems to me that using an array of tweeters might effectively decorrelate the signals by causing so many different nodal patterns right on top of each other that the end result may well be that they are no longer perceptable. Virtually all the cancellations will be largely filled in.
Reflections will be there if you're in a room. The question is how can you work with that "mechanism", such that its damage is minimal, or in some cases an enhancement of the listening experience? (I knew I shouldn't have drank that coffee...)
The long answer involves understanding the psycho-acoustic effects of various colorations and delayed (reflected) signals. How does our perception change over frequency? Why and where does decorrelation have its place? What does each reflection do to the final result and why?
I've read and am convinced that we detect image location in the lower midrange by timing comparisons even more than amplitude comparisons. So a vertical array of five inch drivers for example might smear latteral timing info minimally (that was my choice and no regrets). A vertical array also has a rolled off high frequency response at the ceiling and floor, so you've minimized floor and ceiling bounce to the degree that is practical. I'm also convinced that we detect image location in the upper midrange (roughly 1k - 6kHZ ) primarily by amplitude comparisons, so we want identical frequency responses to be delivered by each speaker. Since there will be many room reflections in the real world, each creating comb filter damage, but likely different on one side than the other, perhaps the additional reflections created by a vertical line array would be one of the best optimizations. Each additional reflected signal has a certain probablity of filling in the cancellations of any other combination of direct signal mixing with a delayed (reflected) signal. I've read that above about 5kHZ, we don't have nearly the sensitivity to exact image location. That range seems to be more of a coloration of what is below 5kHZ, and because we perceive frequencies above about 2kHZ as sounding like they are soming from higher up physically even when they aren't, you could say those frequencies put "air" or height into the soundfield . If one tweeter on each side gives you significant nodal lines (and it should, especially with mono or center images), then it seems to me that using an array of tweeters might effectively decorrelate the signals by causing so many different nodal patterns right on top of each other that the end result may well be that they are no longer perceptable. Virtually all the cancellations will be largely filled in.
Reflections will be there if you're in a room. The question is how can you work with that "mechanism", such that its damage is minimal, or in some cases an enhancement of the listening experience? (I knew I shouldn't have drank that coffee...)
One point that I'm missing in this thread, is how the listener would like to experience the sound.
That can be like a real live band in your own (living)room, a (virtual) concert hall (sort of expanding the acoustics) or just hear what "original" is recorded.
I can imagine that each option has a different approach and is maybe even before the "first hurdle" of David S.
It would be my first counter-question to keyser to go further into the subject.
What are you looking for?
In the discussion with graaf different estimations of usefulness/harmfulness of early reflections showed up.
Graaf posted that Toole citation in #153, so i just re-cited it to bring back into mind, that according to
Toole without reflections there is
"...acoustical crosstalk* that plagues stereo phantom images is present in its naked ugliness..."
* Own supplement: acoustical crosstalk and interspeaker interference
___
To me the question arose, whether smoothing interference patterns (by also early reflections) might be
a major reason for some listeners to prefer wide dispersion speakers, as the enjoyable listening zone might
be considerably widened, even if the zone of "proper imageing" will not be widened to same extent.
Preference for wide dispersion speakers would then as a consequence (at least partially) originate in a
shortcoming of the stereophonic system itself and not (solely) in (postulated) advantages in speaker/listening
room interaction.
If smoothing of interspeaker interference patterns can be achieved in a different way (not just using early
reflections introduced through wide dispersion), then also estimation of usefullness and enjoyability of more narrow
radiating speakers might change considerably even for that share of listeners preferring (conventional phase coherent)
wide dispersion speakers.
I suggested to partially overcome the interference by using radiators, which decorrelate phase between left and
right speaker above say 2KHz, if the listener is not in the median plane between the speakers.
That decorrelation can be achieved without exceeeding limits of group delay, which are accepted for high quality
reproduction.
Last edited:
You need multi channel for that, plus appropriate studio production. For two channel just try to get the recorded sound into the living room without too much mangling.
Direct to reflected ratio does seem to be a critical factor, especially for side-wall reflections, and as you say, the optimal directivity pattern may depend on how dead the room is, along with the listening orientation in the room. (Speakers firing across or along a rectangular room)
I have speakers that are moderately directional above 1-2Khz and I've recently realised that for the room I currently have (fairly small 4.8m x 3.45m and quite reflective/reverberant, with not a lot I can do to change it) that the optimal toe in of the speakers is different depending on whether the listening axis is along or across the room.
For a long time the speakers were on the long wall firing across the short axis of the room. This arrangement worked quite well and put the speakers some 1.5 metres from the side-walls. Imaging was excellent, and I found I got the best results with the speakers toed to converge about 2 feet behind the listening position.
If I toed them to converge in front of the listening position by an equal amount the result was a strong centre phantom image, but very little apparent source width, and a very dry overall sound. I definitely preferred the more toed out configuration.
Some time later the room was reconfigured to put the speakers on one of the short walls firing along the long axis of the room. (For other room use reasons, not for sound...)
For quite a while I was unhappy with the results of this. Although this orientation can work in a large room I've found it generally doesn't work well in a reverberant room this small. It put the speakers less than a metre from the side walls and increased the listening distance considerably, while reducing the speaker to speaker spacing, making the angular separation quite small.
With the speakers in their normal toe in to converge behind the listener configuration sweet spot was narrow and vague, imaging in general was vague and disappointing, and the reverberent nature of the room was all too apparent, with a very strong room characteristic.
Some time later on a "what do I have to lose" hunch I decided to try some strong toe in to cross them over well in front of the listener. (~2 feet) I'm not a fan of crossing speakers in front of the listener but to my chagrin and delight the difference was quite staggering.
It went from sounding like I was sitting well back in the reverberant field with very diffuse imaging to sounding like I was once again in the direct field with pin point imaging, wide sweet spot, and a lot less contribution from the room. In fact overall the result with toeing in front of the listener was very similar to toeing behind the listener in the other room configuration, and overall is very satisfactory, considering that it's the "non-optimal" room orientation.
Thinking about it since then it seems obvious that adjusting the toe in of somewhat directional speakers is directly manipulating the direct to reflected ratio of the side-wall reflections to get a "pleasing" balance.
If there isn't enough side wall reflection, as in the wide room orientation, toe the speakers less, illuminating the side wall more. If there is too much side wall reflection, as in the narrow room configuration, toe them in more.
The implications are interesting, if perhaps a little obvious - if you like precise imaging rather than feeling buried in the reverberant field, more directional speakers are probably better suited to use on the short wall of a rectangular room than speakers with very wide dispersion, and with the directional speakers you can gain more control of the direct/reflected ratio of the first reflection with toe in adjustment. This makes toe in far more critical, but at the same time gives you more control to get the balance you want without needing to change the speaker design or room damping.
On the other hand for speakers mounted along the long wall of a room with side-walls fairly distant, wider dispersion may be more appropriate, and in the case of directional speakers, a fairly minimal toe in. (Crossed behind the listener)
If the toe in is largely about controlling the side wall reflection that also suggests that the optimal toe in angle will change depending on how dead the room is, especially the side-walls.
Can't see anything to disagree with here. Having some vertical directivity in the high midrange and treble is quite doable and does seem to improve imaging at a distance in my experience, and more toe in goes along with what I was just talking about.
In a room with the speakers on the short wall I've found diffusion/absorption on the side-walls near the early reflection point can be beneficial. In one house where I had a 4m wide 8m long room with the speakers on the short wall I had floor to ceiling curtains on both side-walls for the first 3 metres from the speaker side-wall - just normal curtain material, nothing thick, but with large S shaped folds of a depth of about 6 inches. Worked really well at both eliminating the slap echo in the room and controlling the side-wall reflection but without making the room too dead. The improvement in imaging with the curtains in place was like night and day.
For front wall reflections (did you mean back or front wall?) in the midrange I think one of the best solutions is to simply use a wide baffle design to put the baffle step frequency below most of the midrange. Unfashionable today perhaps, and then you have to contend with increased baffle diffraction, but I think a combination of wide baffle and relatively directional drivers (to minimize high frequency baffle diffraction effects on the large baffle) can be quite successful.
Interesting.
Last edited:
10ms to first reflection is a complete fantasy for anyone living in Europe, even 6ms is impractical.
As you say, 3ms can give a satisfactory result, if overall reflectivity and reverberation are kept down to a manageable level. With such short time delays it would seem that direct to reflected ratio becomes a lot more critical factor, thus room damping, speaker positioning and toe in all become more critical and influential.
I agree with that but one minor correction in terminology: Controlled directivity is not the opposite of wide
dispersion. The opposite is narrow dispersion and both patterns can be controlled or "close to constant".
This is pretty much an audio myth. Even Toole says it somewhere (' don't have the book handy). This is only a good proof
that mics "listen" differently than ears/brain.
Reflections will always cause "combing" but certain types and amounts of reflections are desired. So the unwanted ones
should be given real names otherwise people will never stop believing that "combing" is bad in general.
10ms to first reflection is a complete fantasy for anyone living in Europe, even 6ms is impractical.
Which is why you want speaker directivity or room treatment to help to deal with those early reflections.
Isn't that exactly your experience with the two placements in your room? When your speakers got closer to the side wall you weren't happy with the sound until extreme toe in reduced the level of the side wall reflections. You used the speaker's dirctivity to change the ETC.
I completely agree. Speaker directivity and room treatment are both tools to control the direct to reflected ratio, particularly of the early reflections, but also of the reverberant field in general.
If you have more directional speakers you can make do with less room treatment or vica versa. In many listening rooms treatment that doesn't consist entirely of normal "decorative" adornments is entirely out of question for WAF reasons, so more directional speakers - provided they're not large or ugly looking could be a more realistic option than extensive room treatment for many people.
Yes, although I think there is more than one factor involved - shorter time delay to first side-wall reflection, increased amplitude of the side-wall reflection, (less path length differential) and a lower direct to reverberant ratio due to the increased listening distance.
The references I've seen suggest that with a reduced time delay for the first reflection the amplitude threshold for detection of the reflection shifts - the reflection needs to be at a lower amplitude than what is tolerable with a greater time time delay, and yet the closer proximity to the wall will increase the amplitude.
Turning the speakers inwards is reducing the side-wall reflection amplitude considerably, particularly when the tweeters are approximately constant directivity in the horizontal plane, which would cause a fairly rapid fall off in reflection amplitude when passing the critical angle where speaker directivity meets the incident angle of the wall reflection to the listener.
Large time delays to first reflection look good on paper, but in so many living rooms they just aren't practical. The only other option is to reduce the amplitude of the reflections, and getting the right balance of direct to reflected ratio for a given reflection time delay may be the magic ingredient to getting the most subjectively pleasing result, with a greater direct to reflected ratio required for smaller time delays.
Last edited:
So in your particular experience a wider dispersion speaker would have made the room problem less solveable? Is that part of the answer to the original question?
I think you can deal with the early reflections without much effect to the direct to reflected ratio. Killing the first 3 bounces with aborption wouldn't pull that much out of the power response of the system (the aspect that feeds the reverberent field) and certainly putting diffusors at the bounce positions would have no impact at all. (Diffusors reduce the specular reflection by about 10dB by scattering the reflected energy to more angles.)
I think it allows the main problem of this room configuration to be largely solved by just turning the speakers in more, yes, rather than having to add more diffusion/absorption to the side-walls, which I think would be necessary in the same circumstances with a speaker with much wider dispersion.
I didn't turn them a huge amount either, the difference between toed behind me to toed in front of me is probably about 20 degrees, with each speaker going from about 10 degrees off axis (from the listener) one way to 10 degrees off axis the other way.
So it's the direct to reflected ratio of the early reflections that matters then ? And the overall reverberation should be considered and treated separately based on its RT60 ?
Direct to reflected ratio as in the classic calculations of direct level vs. distance and how it compares to the level of the steady state reverberent field (Hopkins Stryker calculations).
Acousticians do look at early to late ratios, for example speech inteligibility is related to energy level before vs. after 30 msec. I wasn't really going there, more thinking of the total reverberent level vs. the direct. As you sit off axis or use a wider dispersion speaker you will reduce the level of direct vs. all subsequent reflections. This was the macro view of acoustics.
The micro view is, what do the directional properties of the speaker infer about the first 5 or 10 room reflections? Taking out a few early reflections doesn't drastically alter the direct to reflected ratio but may make big improvements in sound.
David S.
David S.
Hello,
Good question, but why the system specification of the stereophonic reproduction system does not specify this reflection pattern? The inventor of the stereo was clearly lacking in this regard. Since there is no specification where to aim, can there be an answer to the question either? No. That's why everyone is finding his own answers.
We can disscuss since it's fun and perhaps educating sometimes, but that discussion will have no end since the lack of specs of what the reflection pattern should be. No one knows, simply.
For the ETC I'm not looking any of them since they are limited view on what is happening in the room. Better to do binaural wavelet analysis I think.
- Elias
Good question, but why the system specification of the stereophonic reproduction system does not specify this reflection pattern? The inventor of the stereo was clearly lacking in this regard. Since there is no specification where to aim, can there be an answer to the question either? No. That's why everyone is finding his own answers.
We can disscuss since it's fun and perhaps educating sometimes, but that discussion will have no end since the lack of specs of what the reflection pattern should be. No one knows, simply.
For the ETC I'm not looking any of them since they are limited view on what is happening in the room. Better to do binaural wavelet analysis I think.
- Elias
I think there is definitely merit in that, and it does make a strong argument for speaker directivity that minimizes the earliest reflections.
Another thing that popped into my head that I wonder if everyone is overlooking, is what about early reflections whose source are close to the listener rather than the speaker ?
When we speak about early reflections I get the impression everyone is talking about the first reflective surface encountered by a ray leaving the speaker, (front/side walls and floor) but a reflective wall immediately behind the listener can add a reflection that can be comparable in time delay to the side-wall / floor reflection, and more importantly it's a reflection whose amplitude will be very minimally affected by any speaker directivity.
Furthermore, this reflection "contains" additional delayed copies of all the other reflections whose source are physically near the speaker, such as side-wall reflections.
What is the effect of this on localization, tonal colouration etc ? The obvious cure is "don't sit too close to the rear wall" with too close probably being <1 metre, but like the 10ms argument it simply isn't practical in many smaller rooms where you might be forced to have your sofa up against the wall, putting your head only about 0.5m from the rear wall.
At that distance the delay of a reflection from the wall behind you is around 2.9ms. At the very least it should cause some comb filtering in the lower midrange, and I'm pretty sure I've observed that before.
As a mitigating factor a lot of the high frequencies in the reflection will be attenuated by our HRTF, but this still leaves the lower midrange and bass regions. Speaker directivity can't help, and I don't think diffusion along the wall behind you would help either, just spreading the apparent source of the reflection further to the sides, possibly making things worse.
Absorption on the wall behind you seems to be the only course of action if moving further from the wall isn't possible, but such absorption is unlikely to be effective at lower midrange frequencies. Is being forced to listen so close to the rear wall just a fundamentally compromised situation that we can't do much about ?
Last edited:
Subjective opinion is pretty much the worst part of any hobby. The only subjective opinion that should matter is yours. You honestly should never really read other subjective opinion because they will seldom have the same testing variables, expectation variables that your own conclusions will have.
When you consider any product anywhere, you can choose to read about the science behind the product class then you can priortize what matters to you.
Its easy to explain "reviews", they are tainted, exaggerated wasted bandwidth that shoul be ignored IF the purpose of reading them is to find truth about a product.
Now if you just want to read someone experience knowing that they most likely didnt control the test properly and they have elevated imaginations about differences then enjoy the review but you should make sure you understand their state of mind and all the variables involved (or lack of) in making their conclusions.
Hi Folks
While working on the Unity and then Synergy horns over the last 12 years, I noticed something puzzling / interesting which is related to the stereo image.
I figured since a few of you experiment and that is how new things are often discovered, I thought I would mention my observation and see what you find when you try it.
The object in these horns was to make the drivers combine into one source and that required a quarter wavelength spacing maximum and gradually it was possible to also eliminate the phase shift normally associated with crossovers.
As these became more and more like a simple single broadband source radiating a single lobe with no nulls, I noticed a really weird thing.
Obviously it took a large number of TEF measurements getting there but naturally I did a lot of listening too.
With a single speaker playing a voice, It became harder and harder to identify the loudspeakers location in physical depth. In other words, with your eyes closed, it was easy to tell the direction it was coming from but got harder to “hear” how far away it was when your eyes were closed.
My conclusion was that there was something about the way speakers radiate that gives aural clues as to it’s location in depth. I figured this had to be because the old way of doing it resulted in the speaker radiating a complex field which included these clues while as the horns became more like one driver, they radiated a simple field which was more homogeneous, presented the same signal to the R and L ears and so there were fewer clues as to the source location.
Anyway, I was reminded of this by the discussion here and a comment the other day from a contractor who installed a large Synergy horn in a stadium and commented that even across the street, the announcers voice seemed to just float in front of him (the speaker being about 700 feet away at that point). Also he observed that in a strong crosswind the sound was hardly affected at all whereas the large line arrays he replaced (which depend on a complex interference pattern) were severely affected by even a modest crosswind.
While few large installations are stereo, what I noticed was that when used in the home, the large reduction of source identity made for a much stronger stereo image. A couple years ago, a co-worker and acoustics professor at Columbia College whose interest is the stereo image commented he never heard anything like the strength of image he heard from the sh-50’s in my living room.
While we don’t market to the home market yet, I believe this is a real effect or enough to notice as in the last two years KEF has introduced the “blade” which while is not constant directivity, cannot reproduce a square wave over a broad band, they claim it does radiate as if it were a single point source producing an expanding spherical segment like the Synergy horns.
Anyway, the point was part of stereo imaging is the loudspeaker itself and that most / all of the testing done is in rooms using speakers that have limited directivity and radiate an interference pattern over some / much of their range which makes a single speaker easily localizable in depth and that effects how well they can image aside from all the room effects.
An experiment one can do is set up any stereo speakers outdoors in a normal geometry and listen to them under conditions where there are no room effects. Normally, they have anemic bass BUT the stereo image is usually much better than indoors.
Part B of the experiment is to obtain a pair of full range drivers like fostex makes and mount them on a largish flat baffle with a sealed back and do this indoors or better outdoors. Over much of their range, they radiate a portion of an expanding hemisphere and little or no an interference pattern. If you can ignore the limited LF and hf droop, you will hear a stereo image that is usually much better than typical multi-way speakers.
My conclusion after many speaker listening and BBQ events was that reflections only harm the preservation of the recorded image although not having a “without condition” to compare to, they may seem natural when added to the original signal.
Also, that because we hear with a pair of ears, that gives one the ability to localize each source in depth when what reaches the R and L ears is different enough even though from the same source.
Lastly, that when the R/L ear differences from each source are removed, then a pair of those speakers can produce a much stronger mono phantom image and also any stereo image.
Now a warning, nearly everything I have ever read about this area was written by someone who had a financial interest in what they were saying including me. While we don’t sell hifi speakers it is something I hope to do as this is my lifelong hobby and what has driven the route I have taken at work.
For that reason I would urge you not to take what I am saying blindly but to perform the outdoor listening experiment AND the small full range speaker experiment to see if you would agree with what I have described.
Lastly, for those who have headphones hooked up to your computer, you can try out a couple recordings I made using a microphone array I am working on which is my approach to capturing a more lifelike stereo image. Since the speakers can reproduce a solid image anywhere between the them, what I am working for is a way to capture live events with the same imaging.
Most recordings are produced at the mixing stage and capturing the real thing is actually rather difficult.
What you will hear in these is the front two channels which is the image roughly corresponding to ones field of view in a 360 degree panorama.
Try the Harley recording and the trains here at the bottom of the page;
Danley | Technical Downloads
Best,
Tom Danley
Danley Sound Labs, Inc. | Facebook
While working on the Unity and then Synergy horns over the last 12 years, I noticed something puzzling / interesting which is related to the stereo image.
I figured since a few of you experiment and that is how new things are often discovered, I thought I would mention my observation and see what you find when you try it.
The object in these horns was to make the drivers combine into one source and that required a quarter wavelength spacing maximum and gradually it was possible to also eliminate the phase shift normally associated with crossovers.
As these became more and more like a simple single broadband source radiating a single lobe with no nulls, I noticed a really weird thing.
Obviously it took a large number of TEF measurements getting there but naturally I did a lot of listening too.
With a single speaker playing a voice, It became harder and harder to identify the loudspeakers location in physical depth. In other words, with your eyes closed, it was easy to tell the direction it was coming from but got harder to “hear” how far away it was when your eyes were closed.
My conclusion was that there was something about the way speakers radiate that gives aural clues as to it’s location in depth. I figured this had to be because the old way of doing it resulted in the speaker radiating a complex field which included these clues while as the horns became more like one driver, they radiated a simple field which was more homogeneous, presented the same signal to the R and L ears and so there were fewer clues as to the source location.
Anyway, I was reminded of this by the discussion here and a comment the other day from a contractor who installed a large Synergy horn in a stadium and commented that even across the street, the announcers voice seemed to just float in front of him (the speaker being about 700 feet away at that point). Also he observed that in a strong crosswind the sound was hardly affected at all whereas the large line arrays he replaced (which depend on a complex interference pattern) were severely affected by even a modest crosswind.
While few large installations are stereo, what I noticed was that when used in the home, the large reduction of source identity made for a much stronger stereo image. A couple years ago, a co-worker and acoustics professor at Columbia College whose interest is the stereo image commented he never heard anything like the strength of image he heard from the sh-50’s in my living room.
While we don’t market to the home market yet, I believe this is a real effect or enough to notice as in the last two years KEF has introduced the “blade” which while is not constant directivity, cannot reproduce a square wave over a broad band, they claim it does radiate as if it were a single point source producing an expanding spherical segment like the Synergy horns.
Anyway, the point was part of stereo imaging is the loudspeaker itself and that most / all of the testing done is in rooms using speakers that have limited directivity and radiate an interference pattern over some / much of their range which makes a single speaker easily localizable in depth and that effects how well they can image aside from all the room effects.
An experiment one can do is set up any stereo speakers outdoors in a normal geometry and listen to them under conditions where there are no room effects. Normally, they have anemic bass BUT the stereo image is usually much better than indoors.
Part B of the experiment is to obtain a pair of full range drivers like fostex makes and mount them on a largish flat baffle with a sealed back and do this indoors or better outdoors. Over much of their range, they radiate a portion of an expanding hemisphere and little or no an interference pattern. If you can ignore the limited LF and hf droop, you will hear a stereo image that is usually much better than typical multi-way speakers.
My conclusion after many speaker listening and BBQ events was that reflections only harm the preservation of the recorded image although not having a “without condition” to compare to, they may seem natural when added to the original signal.
Also, that because we hear with a pair of ears, that gives one the ability to localize each source in depth when what reaches the R and L ears is different enough even though from the same source.
Lastly, that when the R/L ear differences from each source are removed, then a pair of those speakers can produce a much stronger mono phantom image and also any stereo image.
Now a warning, nearly everything I have ever read about this area was written by someone who had a financial interest in what they were saying including me. While we don’t sell hifi speakers it is something I hope to do as this is my lifelong hobby and what has driven the route I have taken at work.
For that reason I would urge you not to take what I am saying blindly but to perform the outdoor listening experiment AND the small full range speaker experiment to see if you would agree with what I have described.
Lastly, for those who have headphones hooked up to your computer, you can try out a couple recordings I made using a microphone array I am working on which is my approach to capturing a more lifelike stereo image. Since the speakers can reproduce a solid image anywhere between the them, what I am working for is a way to capture live events with the same imaging.
Most recordings are produced at the mixing stage and capturing the real thing is actually rather difficult.
What you will hear in these is the front two channels which is the image roughly corresponding to ones field of view in a 360 degree panorama.
Try the Harley recording and the trains here at the bottom of the page;
Danley | Technical Downloads
Best,
Tom Danley
Danley Sound Labs, Inc. | Facebook
I'm sure that close rear reflections create the same issues tht frontal reflections do. If you are sitting closer to the boundary than the speaker is to its nearest boundary then, yes, rear wall bounces will arrive prior to some other early reflections.
The problem in all this is that there are too many variables, leading to too many potential listening rooms, to do a definitive study of every case.
It isn't insurmountable though. Some simplifyiing assumptions can be applied. Most studies look at audibility of reflections vs. strength and delay time. Due to our poor sense of image height, I think directional issues can be thought of with height ignored. They then become a function of angle of horizontal grazing incidence. For most tests, sounds are either frontal, full lateral or something in between. If we consider rearward reflections as straight from the rear, or lateral, or something in between, they should duplicate the frontal results (for the corresponding frontal angle) as long as we can factor in the front vs. rear HRTF. That is, rear reflections would be similar to a duller version of the frontal reflections.
I certainly think that sound directly from the rear (not the sides) would be equally hard to separate from the direct sound as its frontal counterpart.
David S.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.