I don't think it will happen unless he moves the crossover quite a bit higher in freq. than it's current <1.5 kHz crossover.
I certainly believe that people have different preferences for the direct to reflected ratio aspect. I had some KLH9 electrostats way back when. They were so directional that they were much like wearing headphones but with a front centered image. I certainly liked that at the time. Others prefer to have a very diffuse sound and hate being able to tell where the speaker is: "sound is coming from right there".
Maybe we can "wrench control from the room" with highly directional speakers, but we then have to be okay with the dryer sound that comes with it. Not everybody is, and it is hard to argue in absolutes.
Adjusting direct to reflected ratio when the reflected sound isn't smooth? Thats pretty sophisticated for a subconscious act.
David
We can get there with an 8" square waveguide and compression driver, currently available in both 90° and 120° x 50° pattens for $15.
Those working with dome tweeters and waveguides get there as well, as the waveguide increases the SPL at the low end with its narrowed beamwidth.
If a diffraction lens is not distasteful, Seas DXT'll do it, too....
Seems we're getting beyond the "flat is wrong" side dish finally and on to a meat--direct vs. reverberant. Here's some graphs of the monitors I've measured. All with reasonably flat axial responses:
Now the average of their forward radiation:
Any guesses as to how these might sound? Which would be most spacious? Most detailed? Best image? strongest bass? It's all right there.😱 One of those speakers goes against Dr. Toole's perceived detail rule(barely) unless he mentions something about detail perception and low diffraction, baffle rigidity or something else the Mackie is doing. Some might suggest it's the spike at 14 kHz. I have strong doubts as the other one doesn't have that spike and still sounds as detailed. Of course there are still some people who believe it could be any number of factors(cables, caps, nonlinear distortion, etc...) but these closely correlate with Dr. Toole's book without even looking into that stuff. Should we be surprised?
Dan
Now the average of their forward radiation:
Any guesses as to how these might sound? Which would be most spacious? Most detailed? Best image? strongest bass? It's all right there.😱 One of those speakers goes against Dr. Toole's perceived detail rule(barely) unless he mentions something about detail perception and low diffraction, baffle rigidity or something else the Mackie is doing. Some might suggest it's the spike at 14 kHz. I have strong doubts as the other one doesn't have that spike and still sounds as detailed. Of course there are still some people who believe it could be any number of factors(cables, caps, nonlinear distortion, etc...) but these closely correlate with Dr. Toole's book without even looking into that stuff. Should we be surprised?
Dan
If I had the opportunity to go to Leipzig I would certainly attend this workshop.
26th International Audio Convention - Abstract Service
"Some conclusions: most listening rooms benefit from directional loudspeakers"
26th International Audio Convention - Abstract Service
"Some conclusions: most listening rooms benefit from directional loudspeakers"
Hello Dave,
I have found head movement (even very tiny) to greatly effect the high freq localisation. I suppose the movement of the pinna notches are detected. In my case this leads to the localisation of the speaker tweeters! Very annoying 🙁
The vertical source increases direct to reverberant ratio. Is this the reason for improved imaging? I suppose yes, partly at least.
Does it mean that the first order floor reflection is the most harmful? In terms of source localization? In any other terms?
I would rather get rid of the first order floor reflcetion alltogether, line array does it nicely for me! 🙂
- Elias
I have found head movement (even very tiny) to greatly effect the high freq localisation. I suppose the movement of the pinna notches are detected. In my case this leads to the localisation of the speaker tweeters! Very annoying 🙁
The vertical source increases direct to reverberant ratio. Is this the reason for improved imaging? I suppose yes, partly at least.
Does it mean that the first order floor reflection is the most harmful? In terms of source localization? In any other terms?
I would rather get rid of the first order floor reflcetion alltogether, line array does it nicely for me! 🙂
- Elias
Hi David,
To me it sounds like the speaker should be a thin line array. A bibole line array perhaps
- Elias
To me it sounds like the speaker should be a thin line array. A bibole line array perhaps
- Elias
No need to add "of course" to observations of that sort.
And that is why these high-falutin' mathematics from Professors of Physics at Princeton University are of minimal practical value.... and secondarily, it raises doubts about those "doctor knows best" solutions from speaker designers which imply "my size fits all."
There are better and worser designs, but there's aren't designs which work in all rooms (or all heads, all chairs, all music....) unless the user has some friendly controls to adjust.
Last edited:
Hello Rob,
Nowadays I havent find this to be the case because I use very high directivity speakers. It hardly matters where they are placed in my room because the direct sound clearly dominates over the reflections (unless one goes to the extream locations which I dont).
What gomes to your questions, Very good questions! I wish I had the answers 😀
- Elias
Nowadays I havent find this to be the case because I use very high directivity speakers. It hardly matters where they are placed in my room because the direct sound clearly dominates over the reflections (unless one goes to the extream locations which I dont).
What gomes to your questions, Very good questions! I wish I had the answers 😀
- Elias
Hi Elias,
With very directional speakers, I also found image stability is extremely sensitive to head position. I think this is a very common finding for owners of big old stats for example. But this has as much to do with the narrow radiation pattern as the head itself. With music and a broad even dispersion speaker, I don't find too much sensitivity to head placement as long as the first side walls are 5' or so out, or this first reflection is well absorbed in the upper mids and above. IME, which is limited to my experience, its close side wall reflections which lead to greater sensitivity to head position in the higher frequencies.
I can understand the first floor reflection as being harmful for tonality, and its why you see so many tall speakers today with multiple spaced woofers, to spatially average out the floor notch. Makes sense given the vertical notches. Queen found line sources reduce vertical image spread theoretically, and I'd take that particular finding with a grain of salt. He based it on an intensity rule (direct/reverb), not a subjective finding. I can't say I ever noted that a design that managed the floor well also resulted in better imaging, but I've never tried the experiment.
Hold on a minute - you are treating the speaker sound as if it was a real sound. The vibrations coming from the speaker are just something cooked up so as to produce a satisfying image at the sweet listening spot (or spots). And "satisfying" often means a good accounting of space (or, at least, horizontal spread) and something like the experience of being in a real place and listening intently to the flute during a loud passage of orchestral music... as people do in real concert halls.
A real cello playing in the corner does not have this problem that Elias introduced. The brain can also keep track of a speaker qua sound source* - and that is a serious problem for "reproducing" a performance undertaken somewhere else.
Your brain's job is to keep the percept coherent even when you move your head around. The real cello seems to stay in the real corner. Brain can't do that with cooked sound.
*Let's say you have a bit of static coming from the amp driving the right speaker. Your brain sure knows it came directly out of that speaker and not in any phantom location elsewhere.
Footnote: it may surprise many to learn that one-ear-only listening has not-so-bad localization (just as single-eye vision has excellent depth perception, depending on the image, of course).
Last edited:
I agree with the observation but I always assumed it was a strong direct field that caused the effect. Shift your head a little and the timing shifts, which is of course strongly tied to perceived image direction. As your speakers move on the continuum towards the Bose 901 (negative d.i., no other slur intended) then you are in a created diffuse field and the stereophonic illusion of the recording becomes secondary. You are in the stereophonic field of your room rather than of the recording. Moving around doesn't matter.
I'd second this. Our sense of vertical directionality is so weak that floor and ceiling bounces resolve as response effects. It takes wide lateral bounces to give us the sense of space (low IACC).
Maybe we should all return to Ken Kantor's Magic speaker as a model.
David S.
Funny you mention that, I was looking again at that article last night. I think the Magic speaker was alot like the high DI designs of today, except it tried to add some sense of spacoisness by recognizing the importance of reflections > 10 ms delay. It was also the use of foam in waveguides 25 yrs before it became chic.
I never figured out a way to take Gresinger's work on IACC and apply it to home use, other than maybe adding sideways spaced velocity sources for bass reinforcement. But that's a whole different thread!
I'm not implying that image localization with stereo is as insensitive to head movement as the real thing. Nowhere did I state that and I don't know how that conclusion could have been reached.
It's a matter of degree. I don't find head position to be as sensitive to image location as some are stating here, but under the assumption that the speaker has broad and even dispersion and that the side walls aren't too close.
The only points I was raising is that the sensitivity of image location with head movement increases when speaker directivity increases, or when the side walls get closer than ~ 5'. I was careful to state that these were my experiences. Caveat emptor.
Dave
Let me try again since, in my humble opinion neither you nor Speaker Dave, otherwise estimable posters, have the concept right.
The brain can't process into a coherent percept ("redintegrate") the cooked sound stimulus coming from the pair of speakers unless you've met the location and head-forward constraints of the situation (and to add to the confusion, the incoherent reflections* arising from your music room too).
Some of your remarks about localization do apply to cellos and real sound sources. What comes from the speaker box isn't such a thing and can't be made into such a stimulus except maybe with headphones or with head constraints.
Your phrase, "image localization" doesn't have a meaning except when the user is sitting still using the sweet spot and then it applies only to the redintegration of the "original" source being reproduced (or to some fantasy pop sound target...) and not to the speaker box per se.
When you turn your head, all bets are off about what the brain is going to do with the sensory input from the cooked sound arising from the speakers because it makes no sense. I suppose the brain just treats it like any sound you hear coming out of a window as you walk down the street.
ATTENTION: "redintegration" is a really great word for this discussion. Sorry I didn't introduce it earlier. Check it out.
*Incoherent with respect to the original source but, of course, coherent (but challenging to process) with respect to the speakers.
I didn't realize we were at odds in our thinking. Lets try this as a statement of belief:
Reproduced stereo is a balance of two different aspects: an illusion of multiple well located sources and a creation of a sense of space. The sense of space, typically called envelopment comes from multiple delayed reflections with significant differences between what the left ear and right ear hear (low IACC is one measure of that).
For the multiple discrete sound sources we will have a combination of virtual sources and real sources. The real sources can be sounds that come from the location of the speakers and they are created by simply sending a signal to one channel only, assuming a direct enough speaker that it doesn't generate its own diffuseness. (the Bose 901 again).
Virtual sources require a particular feed to both speakers at once. If both speakers get signals that create at our ears signals where the level difference and the time difference between the ears are typical of what we are familiar with in the natural world, then we will perceive a virtual source somewhere between the speakers. (Careful subtractive of signals can also create sources outside the speakers.)
This is a fragile creation and requires the timing to be pretty specific to work. Left to right lateral head shift will kill the effect. On the other hand, the effects of spaciousness and of sound sources at the speakers are pretty robust and will remain even if the centered images break down. A lot of the time we listen to a system far enough off the center line that no central images exist, but we still have left and right sources and a sense of diffusion, so we're happy with the result.
We can also have a system where wide dispersion and/or high room reverberation lead to high diffusion. Even a dry feed to the left or right speaker alone will create a diffuse, non-localised image. This isn't really stereo anymore. We have allowed the room and speakers to dominate the effect.
Now, turning the head can still give inter-aural timing within the expected range. Also, since the speakers are in front of us (assuming a typical 50 degrees spacing) the combined sound parts will remain in front of us. Based on this, I don't know why head turning would kill the effect, although it might shift it some.
David S.
Very interesting discussion.
As you all agree, stereo playback is poor at localization and envelopment while you are not sitting at the sweet spot and head movement causes all kinds of unnatural perceptions.
I think it was dr. Griesinger who proposed that the recorging side can alleviate the problem somewhat by uncorrelating the signal between loudspeakers and using multichannel reproduction. That way head turning doesn't break the illusion of envelopment and localization doesn't depend so much on timing and intensity differences between widely spaced speakers.
Also toed in constant directivity speakers widen the sweetspot somewhat - as you move closer to one speaker, the other gets louder. Speaker Dave, you should be familiar with this?
As you all agree, stereo playback is poor at localization and envelopment while you are not sitting at the sweet spot and head movement causes all kinds of unnatural perceptions.
I think it was dr. Griesinger who proposed that the recorging side can alleviate the problem somewhat by uncorrelating the signal between loudspeakers and using multichannel reproduction. That way head turning doesn't break the illusion of envelopment and localization doesn't depend so much on timing and intensity differences between widely spaced speakers.
Also toed in constant directivity speakers widen the sweetspot somewhat - as you move closer to one speaker, the other gets louder. Speaker Dave, you should be familiar with this?
Has anyone besides Griesinger quantified the effect, is it even achievable in the modal region? I wonder how much envelopment is contributed by low IACC in the bass..
Last edited:
Speaker Dave -
I think the obstacle is that you are thinking in terms of the sound stimulus (and then only in terms of mathematically tractable aspects... not in terms of pinna reflection, etc. etc.) while I am encouraging you to think in terms of the brain creating a percept.
Three elements that contribute to this are the two stereo tracks, the speaker design, and a whole bunch of choices the listener has made.
The music producers are trying to engender a sense of something in the listener. They record or add a certain amount of echo (in situ echo or synthetic echo) because that makes it nicer. However, and here is a point you are confusing, the echo of the music room adds a certain sense of delightful fullness but it doesn't add to the coherence redintegrating the original sound. The room sound has nothing to do with the sound of Leontyne Price in Carnegie Hall - even if you are playing it back in Carnegie Hall*.
With some effort, we might think of some highly constrained original sound whose sound-stimulus we can reproduce wholly ("the sound of Leontyne Price as heard at the end of a 20 foot absorbent pipe"). But even with Etymotic in-ear headphones (the ultimate in controlled presentation), the brain can't re-create anything but an orchestra at indeterminate distance over the top of your head somewhere. Therefore, when you say "at our ears," you are thinking rather reductively only of very basic mathematically tractable dimensions such as time, freq compass, and so on... and overlooking a whole lot of other cues the brain is getting to "discredit" the illusion you think the Master Engineer ought to be delivering. The "discrediting" takes place whichever way your head is pointing and, as I said earlier, even with headphones.
There are hard-core headphone listeners who think speaker users are settling for pretty inferior auditory experience. Are they wrong - thinking only in terms of delivery of sound to the ears?
For sure, I am not saying "the bumblebee can't fly." Rather, I am saying we need to take a user-centered grasp of the cues presented. From there, music producers, speaker designers, and end-users can fashion satisfying sound.
Apropos the above comments, many times on this forum, I have posted disparagingly of the effort to "reproduce" Carnegie Hall in your living room. No question in my mind that music producers know more about listening than speaker designers seem to - esp, Professors or Physics at Princeton. They watch the score and when the flute plays something interesting, they crank it up - exactly as people do when sitting in Carnegie Hall. I bet a lot of readers don't understand what "crank it up... sitting in Carnegie Hall..." could possibly mean.
*Two loudspeakers do not emit sound like one large singer. Same is true even if you recorded her anechoically and played it back in Carnegie Hall, but closer.
I think the obstacle is that you are thinking in terms of the sound stimulus (and then only in terms of mathematically tractable aspects... not in terms of pinna reflection, etc. etc.) while I am encouraging you to think in terms of the brain creating a percept.
Three elements that contribute to this are the two stereo tracks, the speaker design, and a whole bunch of choices the listener has made.
The music producers are trying to engender a sense of something in the listener. They record or add a certain amount of echo (in situ echo or synthetic echo) because that makes it nicer. However, and here is a point you are confusing, the echo of the music room adds a certain sense of delightful fullness but it doesn't add to the coherence redintegrating the original sound. The room sound has nothing to do with the sound of Leontyne Price in Carnegie Hall - even if you are playing it back in Carnegie Hall*.
With some effort, we might think of some highly constrained original sound whose sound-stimulus we can reproduce wholly ("the sound of Leontyne Price as heard at the end of a 20 foot absorbent pipe"). But even with Etymotic in-ear headphones (the ultimate in controlled presentation), the brain can't re-create anything but an orchestra at indeterminate distance over the top of your head somewhere. Therefore, when you say "at our ears," you are thinking rather reductively only of very basic mathematically tractable dimensions such as time, freq compass, and so on... and overlooking a whole lot of other cues the brain is getting to "discredit" the illusion you think the Master Engineer ought to be delivering. The "discrediting" takes place whichever way your head is pointing and, as I said earlier, even with headphones.
There are hard-core headphone listeners who think speaker users are settling for pretty inferior auditory experience. Are they wrong - thinking only in terms of delivery of sound to the ears?
For sure, I am not saying "the bumblebee can't fly." Rather, I am saying we need to take a user-centered grasp of the cues presented. From there, music producers, speaker designers, and end-users can fashion satisfying sound.
Apropos the above comments, many times on this forum, I have posted disparagingly of the effort to "reproduce" Carnegie Hall in your living room. No question in my mind that music producers know more about listening than speaker designers seem to - esp, Professors or Physics at Princeton. They watch the score and when the flute plays something interesting, they crank it up - exactly as people do when sitting in Carnegie Hall. I bet a lot of readers don't understand what "crank it up... sitting in Carnegie Hall..." could possibly mean.
*Two loudspeakers do not emit sound like one large singer. Same is true even if you recorded her anechoically and played it back in Carnegie Hall, but closer.
Last edited:
It was Kates that first wrote about this in the 70s, I've been a fan of it forever. With abroad enough even dispersion, you can illuminate a wide sweet spot through overlap in front. It allows a wider 'intensity stereo" effect. It also lessens the strength of the side wall reflection above 2 kHz.
No ones quantified it in home to my knowledge. Below 500Hz, cues are more phase based than intensity. Trick is how to actually set up the sound field at such long wavelengths.
I certainly get it, its an illusion. Ben, i completely agree with you that the best training for speaker design, once you grasp the basic science, is learning how to mix and master. These people know how to craft the sound, the illusion. I more than get it. However, this doesn't mean the science of audition is not very helpful at the task.
For example, you wrote:
The room sound actually does help extract the sense of space from teh recording. Those wacky PHDs showed this in the 90s, I believe it was Olive with his detection of reflection study. What was found was that natural local reflections heighten the detection of the natural sense of space and ambient cues buried within the recording.
But you raise a good point. I for one am not skilled in the art of mixing and mastering, but I do know a thing or two about auditory illusions. I would love to see a thread plumb this topic exclusively, from the side of audition (my background) and from mixing and mastering.
Dave
- Status
- Not open for further replies.