Constant directivity. They allows reflections as delayed (and similar) replicas of direct sound and the ear (brain) ignores the room.
Constant directivity loudspeakers are:
- Omni-directional
- Dipole radiators
- Cardioids
According to Toole, real envelopment is actually a property of multi-channel reproduction. Stereo is always mainly frontal and lateral.
- Constant or controlled directivity designs as already mentioned
- Speakers that create a sufficient level of reflections => medium or wide dispersion (as implied by omni, dipole, cardioid)
- placement of the speakers >= 1m away from walls
With the precedence effect at play the indirect sound does not generate additional sound sources and the increased amount of reflections also masks the speakers at least a little. Ultimately your brain dials into the phantom image / stage.
I have a whole bunch of speakers that do the disappering act and none of them would fall into those categories...
A low diffraction signature is a boon to this task.
dave
I see two different ways to make loudspeakers disappear: a brute force method and a more subtle one.
Brute force is making reflections so overwhelming that the direct response of the speakers is for the most part disguised. This will even help with recordings which have each stereo channel "glued" solely to the corresponding speaker. But you pay with a loss in definition, precision, resolution of the stereo stage or however you call that.
The subtle method is making the speakers work as smooth as possible. Avoid loudspeakers which distort, resonate or have sudden changes in their directivity. Keep the geometry of the listening environment as symmetric as possible. This will do the disappearing act and keep the precision of the recorded stereo stage. But it might annoyingly uncover recordings, where nearfield mono takes have been panned to a single stereo channel.
Rudolf
Brute force is making reflections so overwhelming that the direct response of the speakers is for the most part disguised. This will even help with recordings which have each stereo channel "glued" solely to the corresponding speaker. But you pay with a loss in definition, precision, resolution of the stereo stage or however you call that.
The subtle method is making the speakers work as smooth as possible. Avoid loudspeakers which distort, resonate or have sudden changes in their directivity. Keep the geometry of the listening environment as symmetric as possible. This will do the disappearing act and keep the precision of the recorded stereo stage. But it might annoyingly uncover recordings, where nearfield mono takes have been panned to a single stereo channel.
Rudolf
not necessarily, check out tinitus' and Radugazon's reports on their flooder speaker tests
For me there are two issues that make a speaker dissapear. One is response related and the second is directivity related.
A speaker needs to have very low coloration, smooth response and freedom from resonances. These are all the same thing, primarily coming from smooth and flat axial frequency response. Speakers that suffer from not having smoothness end up having a "personality" that is always there reminding you that you are listening to a speaker.
I also think that a speaker with reasonably wide dispersion becomes less definite as a source in the room. This may be at odds with sharp imaging, but if you want a speaker that disappears then it tends to help.
I don't agree that flat reflections are mandated.
David S.
A speaker needs to have very low coloration, smooth response and freedom from resonances. These are all the same thing, primarily coming from smooth and flat axial frequency response. Speakers that suffer from not having smoothness end up having a "personality" that is always there reminding you that you are listening to a speaker.
I also think that a speaker with reasonably wide dispersion becomes less definite as a source in the room. This may be at odds with sharp imaging, but if you want a speaker that disappears then it tends to help.
I don't agree that flat reflections are mandated.
David S.
I pretty much agree. No significant resonances is the highest criterion on my list. I also agree that in general wide dispersion speakers do the disappearing act better than high directivity ones. In my experience it actually helps to have lots of strong very early reflections. In the living room the speakers are less than 2 feet from the side-walls and they 'disappear' very well. However, with the right recording, high directivity speakers disappear equally well. That means no sharp pan-potting and lots of reverb.
Thats right too. When the recording is dead and the speaker is directional, then the speaker becomes very obvious as a source. As the recording becomes more diffuse the opposite happens.
David
If you play one ESL 63 set at mid room height (say on a table) and play a dry voice, the sound “sounds like” it is coming from behind the speaker somewhere.
That speaker and other full range speakers (like a Manger) which radiate similarly (project a simple portion of a sphere) also have this property.
Another easy example is a small fostex driver on a large flat baffle. It may be limited at both ends but over much of the middle, it radiates as a simple point source.
AS I got the Unity and then Synergy horn drivers to knit together into one source and radiate that same portion of a sphere, that same effect happened.
If you played a voice through one and stood with your eyes closed, you could clearly hear what direction the speaker is but it got harder and harder to hear how far away it was when your eyes were closed.
Now As you walk forward and put your head into the horn mouth of say an SH-50 , the sound still sounds like it’s somewhere off in front of you, no clue there are three frequency ranges.
Conversely, a loudspeaker that radiates a complex interference pattern may measure identically but since you don’t measure from two points like you hear, this effect is not visible per say.
If when you close your eyes and listen to one, to the degree you can localize it’s physical depth, that is because of differences between what arrives at the right and left ears where a simple source radiates something much more similar or identical reaching both ears.
The differences allow one to identify the “place” the sound is originating. When I first noticed this effect back then, I called it source identity.
This is not what produces the best stereo image.
The “mono phantom image” is the basis of stereo reproduction, the more realistic or convincing the mono phantom is, the better a stereo image will be projected. The phantom is produced by the R and L speaker producing an identical signal and to the degree it is exactly identical, your brain determines it is coming from directly in front of you (where there is no speaker).
If the speaker radiates a normal complex pattern, then it will be easier to hear the speaker as a source in addition to the desired image.
When things are right, it will sound like a continuous stereo panorama, like a giant window in your room open to another space. This of course depends on a lot of things in addition being right, a minimum of close room reflections being critical among others.
Best,
Tom Danley
Danley Sound Labs
That speaker and other full range speakers (like a Manger) which radiate similarly (project a simple portion of a sphere) also have this property.
Another easy example is a small fostex driver on a large flat baffle. It may be limited at both ends but over much of the middle, it radiates as a simple point source.
AS I got the Unity and then Synergy horn drivers to knit together into one source and radiate that same portion of a sphere, that same effect happened.
If you played a voice through one and stood with your eyes closed, you could clearly hear what direction the speaker is but it got harder and harder to hear how far away it was when your eyes were closed.
Now As you walk forward and put your head into the horn mouth of say an SH-50 , the sound still sounds like it’s somewhere off in front of you, no clue there are three frequency ranges.
Conversely, a loudspeaker that radiates a complex interference pattern may measure identically but since you don’t measure from two points like you hear, this effect is not visible per say.
If when you close your eyes and listen to one, to the degree you can localize it’s physical depth, that is because of differences between what arrives at the right and left ears where a simple source radiates something much more similar or identical reaching both ears.
The differences allow one to identify the “place” the sound is originating. When I first noticed this effect back then, I called it source identity.
This is not what produces the best stereo image.
The “mono phantom image” is the basis of stereo reproduction, the more realistic or convincing the mono phantom is, the better a stereo image will be projected. The phantom is produced by the R and L speaker producing an identical signal and to the degree it is exactly identical, your brain determines it is coming from directly in front of you (where there is no speaker).
If the speaker radiates a normal complex pattern, then it will be easier to hear the speaker as a source in addition to the desired image.
When things are right, it will sound like a continuous stereo panorama, like a giant window in your room open to another space. This of course depends on a lot of things in addition being right, a minimum of close room reflections being critical among others.
Best,
Tom Danley
Danley Sound Labs
What's your opinion on how best to avoid source identity with a conventional, vertically-stacked multi-way speaker?
454casull said:
Well if one were building one, some of the things one could do would be;
Use a small tweeter that works well and is flush to minimize that part of the diffraction. I have tested several Acuton tweeters which while expensive perform very well as a simple source.
Mount these on a flat baffle covered with absorption and put a nice big radius on the cabinet edges.
The acoustic output from drivers can add together coherently like signals adding through resistors but this only happens when the sources are less than about ¼ wavelength apart. At this acoustically small spacing, the driver add coherently and if you reversed one, it would cancel out adding to the other.
When the spacing is larger say ½ wavelength or more, then the sources radiate independently and produce an interference pattern made of a series of lobes and nulls. The lobes are produced in regions where the two sources add constructively and the nulls where they cancel each other out.
To avoid an interference pattern (instead of a simple spherical source), put two mids as close as you can just above and just below the Hf. IF POSSIBLE choose the drivers such that the spacing between the hf and mid is no more than ¼ wl at the predicted crossover Frequency.
This places the mids ½ wl apart at the higest frequency they operate at with a source in the center, that narrows the vertical directivity but in this case, not too much.
The same thing would apply for the woofers, they would go above and below the mids and the same quarter wl rule applies.
Here, all the physical spacing’s (with proper crossover) would avoid lobes and nulls at the crossovers, the driver to driver spacing in the mid and lows are close enough not to produce an interference pattern.. The only aspect not dealt with totally is the shape of the radiation pattern (confining it to a smaller angles) but that is one reason why they have horns, they extend the baffle and limit the angle to some less than 180 degree value. Where a box speaker has a “baffle step” a horn has a “pattern loss” frequency where they begin to loose pattern control and the angle widens.
Both are the point where the angle and dimension determine the Frequency.
An alternate solution might be a large flat baffle, a little Fostex full range with an optimum sealed rear and a ring of little woofers around it?
Hope that helps,
Best,
Tom Danley
Yes / no, if you were outdoors then directivity is not an issue, but in a room, then the reflections are all things which are not in the recording.Professor Smith said:
Your ears and brain produce an image but it is a complicated and learned process. A measurement microphone samples pressure from one point in space and has no directionality BUT your ears are very different. For example, the shape of your ear changes the frequency response according to height. Over time, we hear all the angle dependent alterations not as comb filtering but as the source location. The changes that your ears and head cause are how we hear behind, above side to side and because headphones do not produce any of those clues, our brain interprets the image as floating in your or above your head.
A real problem is capturing a stereo image too, most music is a contrived stereo image but that is something I have always been interested in. I am working on new way to capture a stereo image, it is a work in progress but it has glimmers of promise I think.
If interested, try some of the recordings at the bottom of this page from work, use headphones first if you have them..
Danley | Technical Downloads
Best,
Tom Danley
Danley Sound Labs
Happy Thanksgiving
Last edited by a moderator:
But from the other threads we've learnt this should lead to pin point imaging. No ? When the reflective disturbances have been cleared out the phantom image is said to be rock solid. Are you telling now that you too have problems in perceiving phantom images but are perceiving the speakers ? - Elias
In my opinion high directivity speakers still have the edge, because they can do both sharp imaging and 'the disappearing-act'. It all depends on the recording. You can even have both virtues with the same recording, e.g. a well-centered singer with a spacious/diffuse sounding big-band supporting him.
In my experience tricks with phase and decorrelation (like the Q-sound effects in Roger Waters' Amused to Death album) work much better with high-directivity speakers an no early reflections. Q-sound is an extreme example, but in other recordings a similar effect is used to increase spaciousness and envelopment. With my main system these effects work really well, while in the living room (wide-dispersion speakers relatively close to the side walls) the room-sound is much stronger, so I hear less of the recording itself and the effect is not as strong.
In my experience tricks with phase and decorrelation (like the Q-sound effects in Roger Waters' Amused to Death album) work much better with high-directivity speakers an no early reflections. Q-sound is an extreme example, but in other recordings a similar effect is used to increase spaciousness and envelopment. With my main system these effects work really well, while in the living room (wide-dispersion speakers relatively close to the side walls) the room-sound is much stronger, so I hear less of the recording itself and the effect is not as strong.
I have no problem with higher directivity and the sharper imaging that it provides. I also don't think that sounds imaging at the speaker is a problem. If the recording is dry, or some elements are fed to left only or right only, then it is totally appropriate for the sound image to be right there at the speaker.
However, the OP asked how to make a speaker disappear as the source....
David S.
However, the OP asked how to make a speaker disappear as the source....
David S.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.