From the child`s experience: Walking down the church hall up to the organ gallery and discovering that the notes came from different parts of the organ. Playing at the other end of the corridor, while my mother was playing the piano in her room. Listening to self-made omnis from varying distances.
To me it is an iron law: If you leave the vicinity (nearfield) of the sound source and let the room do (too much of) its work, you loose the detail and the resolution, but you gain envelopment and the sources become bigger.
Of course it was only recently that I became aware in some detail how these things work in conjunction with acoustics and the human brain.
Rudolf
The Toole/Olive study with 4 speakers in 4 different rooms was informative in the subject. The detrimental earlier reflection(if they can be called that) didn't influence which speaker was preferred in every room negatively or positively. I think we are essentially overanalyzing this of course. They have effects..good and bad, but overall is the speaker measures smooth/flat and set up with some care, it's going to make an enjoyable noise with enjoyable material.
Oh, no offense intended Dave! I thought it sounded funny so I had to write it. Probably should have thrown a smiley on there, but the deadpan can be more effective sometime.
Someone mentioned accuracy a few pages ago. We will never be close to accurate for the original event, but we can get close to the recording. If there were one "ideal" pattern, we know what we are stuck with still.
Dan
I mentioned accuracy in post 1469.
Sharp imaging, I think, is the key issue is this whole directivity discussion. Sharp imaging is required in the production end, as most recordings are done with a microphone placed close to the instrument and positioned in the acoustic space afterwards. It is not the 'original event', but a production.
As we know from this discussion: magnitude, time-delay, direction and number of reflections have influence on imaging and envelopment, exact numbers are unknown. I would say that at the reproduction end the loudspeakers must be able to deliver the same sharpness in imaging as the production end, just like a 1080P HD movie doesn't come to it's full potential on a 20 inch SD television. Omni's, dipoles and conventional boxes are all perfectly capable of delivering this sharp image, but in general they need more distance to the walls to deliver the same sharpness as the narrow directivity waveguide speakers can deliver. So, room size dictates the directivity pattern.
Remains spaciousness and ASW. Both should, in my opinion, be controlled at the prodcution side. It can be done, a nice example is this one:
Not sure what you describe is the same scenario. Equidistant stereo triangle delivers sharp center phantom images. Now add ipsilateral reflections and the center becomes ambiguous. Why is this happening? Instead of two false pinna cues we now have 4 but they are still symmetrical. Seems like the two pairs of reflections are perceptually summed and increase the opening angle of the stereo triangle which creates the very same sensation of a "missing" phantom center.
Unless the loudspeakers are close to the side walls, I've not heard center become "ambiguous", or even altered much in any fashion (..again, assuming equidistant speakers that are equidistant to their ipsilateral wall.)
So, for me and those I've tested - it isn't happening.
Now alter the direction of the loudspeaker (i.e. "toe-out"), and yes - center *can* start to become more ambiguous.
Is this a function of the ipsilateral wall, or that of the driver + baffle? Considering I've used heavy insulation to test this on the wall I can say without a doubt that it's the loudspeaker, NOT the wall.
Assuming a member has access to loudspeakers, it's not difficult to test any of this.
Of course none of this is to say that the wall doesn't contribute as well in these circumstances, rather that the manner of it's contribution is an alteration that doesn't really impact it's clarity with respect to position.
Now the *contralateral* wall, that can often "stretch" and image somewhat - in a similar manner to "apparent source width" for acoustics. Of course it also has other negative qualities with respect to cross-talk, tending to "shrink" sound-stage width to the boundaries of the loudspeakers depending on the horizontal polar pattern of the loudspeaker.
Yes, this is function of cross-talk generally.
I'm not sure about "envelopment" however - I suppose it would depend on your definition.
I see my dipoles as being effectively controlled directivity below about 500HZ due to the cancellations at their sides, which makes embedded timing cue info less damaged by lateral room reflections in the 100HZ - 500HZ region, which I like when I'm sitting in the sweet spot. I noticed that in my room anyway, when I'm significantly outside of the sweet spot, the rear emission reflections off the front wall become a bigger part of what I hear, which makes the speakers sound more like ambience speakers, which may be preferred for non-critical listeners anyway. They're not perfect, but I still like the set of tradeoffs. My openbaffle dipoles give a very enjoyable listening experience. Coarse, I'm also using an inter-aural cancellation circuit, so lower mid timing cue info can really come to life.
Inter-aural crosstalk screws up our ability to use lower midrange timing cue information embedded in a recording almost completely (if there is any). Sidewall reflections that might generate a sense of spaciousness in the upper frequencies (above 1kHZ), may compound the timing cue image information problem in the lower mid freqs., to the point of weakening the overall sense of center image. That's my present belief. When using my inter-aural cancellation circuit, I've found in many different rooms over 25 years, that sidewall reflections are very detrimental to it working well.
Member
Joined 2009
I observe the same phenomena. Furthermore, I find that strong contralateral reflections (with extreme toe-in) tend to exagerate the center phantom image while shrinking the width of the soundstage.
I don't think that we can fully understand how this is happening without answering the fundamental question "why do we hear a center phantom image in stereo".
Well, that one is fairly simple. This was nicely described in some papers by Ben Bauer that were published in JASA in the early 1960's. The topic was called the "theory of stereophony".
I did observe this "hole in the middle" with recordings using interchannel level differences only. It's the very same sensation as using an opening angle >60°.
Very true.
Boris' (and my) question goes beyond that level of explanation.
Didn't one of the references posted a few pages back say that when it comes to Azimuth estimation at high frequencies, amplitude and arrival time differences and HRTF contouring are all used together, but that some dominate for some of the angular range of azimuth and others dominate for the rest of the range ?
I don't remember which way around it was, but it makes some sense if the amplitude and arrival time comparison at high frequencies dominates from the median line out to, lets say, 30-45 degrees or so, and the HRTF response contour of the near ear dominates from say 45 degrees to about 120 degrees.
Why ? Because if we start at directly ahead and start to go off axis to the right, for the first 30 degrees or so there will be a fairly rapid loss of high frequencies in the left ear as it becomes occluded, as well as a change in arrival times of impulsive sounds, and whilst there is a change in the HRTF of the near ear, it's not a huge.
By the time you get to 45 degrees or so the amplitude of high frequencies in the opposite ear has dropped about as far as it is probably going to, and is of no further use in determining azimuth, the arrival time difference with angle will become smaller and smaller rapidly as we approach 90 degrees, so it becomes of little use either.
The only strong cue left in this sector is the HRTF of the near ear which, conveniently, has a very exaggerated change in response (and therefore a fine degree of resolution) in the vicinity of 90 degree azimuth where the other two sources of localization must be weak and ambiguous.
From this I postulate that a phantom image can form when the speakers are at or closer than +/-30 degrees, because:
(a) both ears receive exactly equal amplitude, (although comb filtering will mess this up slightly if you're not dead centre)
(b) both ears perceive equal arrival time of the initial impulse, and the delayed cross-talk from the opposite speaker also arrives slightly later but at the same time in both ears
(c) the conflicting cue of HRTF response is ignored because it corresponds to a narrow angle where the brain gives priority to amplitude comparison and arrival time. It may also be ignored because the HRTF cues of the left and right ears don't agree on a discrete source location - each ear says the sound is coming from its side, because they are both being fooled by hearing the near side speaker arrival first, whereas with a single sound source to one side the HRTF's of the two ears would be for complimentary angles.
Now put your sound sources out much wider, lets say +/- 60 degrees. Most people who have tried really wide angular separations like this will have found that the centre phantom image doesn't form well (if at all) with it being stretched and indistinct, perhaps even starting to sound like its coming from somewhere to each side.
Why ? Perhaps its as simple as the pinnae HRTF starting to exert dominance over localization at the wide angle. For a normal sound source at 60 degrees or more most (if I'm right in my theory above) of the localization is coming from the HRTF of the near ear and very little from amplitude and time comparisons.
The pinnae HRTF contribution of azimuth detection says 60 degrees, so the brain gives high priority to this, despite the amplitude and time delay still being equal. The contradiction between the two cues is probably why the location becomes indistinct and is neither directly ahead, nor completely in the direction of the speakers.
Strong specular side-wall reflections that lead to early reflections from wide enough angles where pinnae HRTF cues are dominant would therefore be detrimental to the phantom image formation.
Yes, the phantom image is an auditory illusion, and if the above explanation is right it only works in the first place because the pinnae cue can be suppressed and overridden by amplitude and arrival time cues, but only out to a certain maximum angle.
This also means if we want precise phantom channel imaging and excellent envelopment at the same time, wall reflections are only ever going to be a compromise between the two, and our two best hopes of achieving both at once are either to use very wide speaker separation together with an actual centre channel, (either discrete 3 channel or derived from stereo) or to use a conventional +/- 30 degree 2 speaker stereo triangle with the addition of +/- 60 degree "envelopment" speakers, preferably from discrete channels available on the recording, or if not, derived from the left-right ambience information and with the centre phantom image cancelled as much as possible in these wide speakers. (We don't want anything from the centre appearing at this wide angle)
Last edited:
Let's not forget that strong toe in and toe out have an effect on the direct to reflected ratio.
We use directivity index as the measure of general directivity of a speaker. It compares the power response of a given system to the power response of a perfectly omni directional system sharing the same on axis frequency response. If a speaker radiates into half space it has a d.i. of 3dB (radiates half the power of a perfect omni. A d.i. of 6dB is typical in the midrange and 10dB in the upper range of a typical system (or perhaps across the board for a CD horn).
Now the concept of d.i. always assumes you are on axis, listening at the loudest point of the polar curve. If, however, you aim the hot spot away from you, you get a lower level of energy at the listening position but the total power radiated into the room remains the same. Likely you will turn the level up a little to get the direct level higher and further raise the reverberent field.
So toeing in or out strongly will automatically reduce the direct to reflect ratio, involving the room more in what you hear.
David S.
We use directivity index as the measure of general directivity of a speaker. It compares the power response of a given system to the power response of a perfectly omni directional system sharing the same on axis frequency response. If a speaker radiates into half space it has a d.i. of 3dB (radiates half the power of a perfect omni. A d.i. of 6dB is typical in the midrange and 10dB in the upper range of a typical system (or perhaps across the board for a CD horn).
Now the concept of d.i. always assumes you are on axis, listening at the loudest point of the polar curve. If, however, you aim the hot spot away from you, you get a lower level of energy at the listening position but the total power radiated into the room remains the same. Likely you will turn the level up a little to get the direct level higher and further raise the reverberent field.
So toeing in or out strongly will automatically reduce the direct to reflect ratio, involving the room more in what you hear.
David S.
Is that true about "most people"? I always found that the very wide angular spacing (nearly sitting between the speakers) gave a strong center phantom. As we talked about before, it also eliminated the comb filtering of being a little off the center line.
I think with narrow angular spacing most people don't realize they are a little off the center line (typically we are) and that there is no phantom center. The hole in the middle isn't so bad if the speakers aren't far apart.
Wide spacing works well if you are well entered.
David S.
Yes, headphones give a very strong phantom center
(Sorry for trying to be funny)While listening to such a setup I get a lot of images at the sides, less in between and a center that jumps with little head movement.
I still like the idea of a center speaker utilizing room boundaries. Has anyone experience with arrays like the ones proposed by Hooley? See Cambridge Mechatronics (CML) conceives, develops and pioneers innovative actuation solutions for consumer electronics. This involves the development and provision of semiconductor devices, software and other Intellectual Property. CML has a portfolio
Yes, headphones do give a very strong phantom center. That is because the time delay difference between channels is always zero and the signal is presented nearly anechoically. Of course, the image tends to be inside the head (that is, neither in front of or behind the listener).
I find if I aim my speakers towards each other and sit in the middle, well centered, the experience is identical to headphones.
David S.
Re: phantom center and Stereophony.
Actually, I am not convinced that it does go beyond this. Consider this issue: what arrives at the ears is the resultant of the various waveforms (in a vector sense of the word). That resultant is the waveform that the two ears deal with. Again, this is stereophony and how a phantom center is created.
I had a personal demo of the 1 Limited unit given by Tony Hooley. It was later marketed by Pioneer for $50,000 (at least you only had to buy one!)
http://www.jstage.jst.go.jp/article/ast/27/6/354/_pdf
This was a fascinating device. It is essentially a line array (actually a plane array) with steerable "fingers" of sound. The theory is much like what the military uses to create programable, steerable antenna arrays.
The intention is to input 5 channels of surround material and the array processes each input and applies it to the array to give individual sound rays. The sound rays have to be narrow enough that a listener in front hears nothing but the center channel input, from that location. All the other channels go off to sides and corners to bounce back at the listener from the spots where left, right, left surround, and right surround should be. Of course, if you don't have a wall at the right spots you are in trouble.
Tony put on some single channel test signals while I moved around in front of the system. The beams were very tight, a foot or so left or right and the sound fell off considerably.
The Yamaha Sound Bar series follows the same theory with a wide but short array.
It was one of the cooler technical achievements that I have witnessed but I'm not convinced that it is a strong commercial prospect.
David S.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.