How do you get good imaging?

[a_tewinkel]

Originally posted by tinitus



Yes, and NO

A small amount of reflection is good...its a matter of finding the right amount fore that particular speaker in that very room

There is a difference between reflections and reflections. It's all about perception. Early reflections ( < 10-20 ms) should be minimal, but a reverberant field is necessary. Listening in a completely dead room is not very enjoyable. In a dead room, there is only the reverberant field information recorded in the source material and not from the room, but it will not be perceived as a reverberant field.

However, when early reflections cannot be avoided it is probably the best if they are spectral copies of the direct sound from at least a few hundred Hertz up, so that the human hearing perceptually masks them and does not identify them as a wavefront from a different source. Nevertheless, even these reflections will affect virtual source localization (imaging) as they are cues of the position of the speaker itself.

Also, as Dr. Geddes points out in one of his papers, early reflections may cause comb filtering which will affect the perceived frequency response.

A flat power response is also preferable for a neutral sounding reverberant field.

[gedlee]

Originally posted by a_tewinkel

Also, as Dr. Geddes points out in one of his papers, early reflections may cause comb filtering which will affect the perceived frequency response.

I'm a little less inclined to make that statement today because of some discussion with Dr. Toole about his book and further investigations.

Very early reflections < 10 ms can fuse with the direct sound to create an image ambiguity (poor image) - the shorter the delay the worse, which makes cabinet edge diffraction and nearby object diffraction a major issue. Good imaging requires a perfectly clean direct signal and impulse response for at least the first 5-10 ms. After this the reflections become masked by the precidence effects and it then appears that comb filtering effects are not significant. Floyd Toole still suggests that all reflections are perceived as positive, although this might be bacause they increase the stage width and the feeling of spaciuosness. There is likely a real tradeoff between spaciousness and image for very early reflections - one gets better only at the sake of the other, although there may be an ideal. I believe that the ideal exists when the first few ms range is completely free of reflections and diffraction, but after that ALL reflections only add spaciousness without a loss of image. Controlling the first reflections without reducing later ones is the key.

[tinitus]

Which is why I fill the tweeter faceplate screw holes with black wax...makes quite a difference
It is also known that the "edge" around the dome on a Vifa XT25 should be rounded
I believe new models are all nicely rounded at this point
Its all only logical to be that way...I mean, noone would ever think of mounting drivers without countersinking
Funny that professional driver designers should learn that from DIYers

Wonder when they will take the next step getting rid of ALL chassis screw holes...they should know by now thats its not just a cosmetic thing

[Patrick Bateman]

Originally posted by a_tewinkel


There is a difference between reflections and reflections. It's all about perception. Early reflections ( < 10-20 ms) should be minimal, but a reverberant field is necessary. Listening in a completely dead room is not very enjoyable. In a dead room, there is only the reverberant field information recorded in the source material and not from the room, but it will not be perceived as a reverberant field.

However, when early reflections cannot be avoided it is probably the best if they are spectral copies of the direct sound from at least a few hundred Hertz up, so that the human hearing perceptually masks them and does not identify them as a wavefront from a different source. Nevertheless, even these reflections will affect virtual source localization (imaging) as they are cues of the position of the speaker itself.

Also, as Dr. Geddes points out in one of his papers, early reflections may cause comb filtering which will affect the perceived frequency response.

A flat power response is also preferable for a neutral sounding reverberant field.

Nicely summarized! I'm glad we're drawing a disctinction between early and late reflections.

One thing I've noticed with waveguides is that you can bend the rules a bit. For example, you wouldn't want a conventional speaker too close to the walls. But with a waveguide it's possible if the wall forms a boundary with the waveguide.

For instance, I used to have my speakers set up in a very small room. With the speakers perpendicular to the rear wall, the image lacked spaciousness. But the sense of spaciousness increased dramatically when I oriented the speakers so that the mouth of the waveguide formed a boundary with the side wall.

It was quite amazing to turn the lights off and hear a soundstage which extended beyond the confines of the room!

Of course this isn't ideal, I really needed a bigger room. But if you're cursed with a small room, waveguides can offer unique advantages if you use room boundaries to full effect. (The same thing applies to cars also, and it's the reason waveguides sound excellent in a vehicle.)

[HK26147]

Excuse Me, Dr. Geddes:
(1) Is it safe to assume that the Hass Effect window is predominate in the auditory region utilized for speech ( approx 170 - 4Khz ).
(2) Is there a perception vs frequency analysis?
(3) The 5 - 10ms window would imply that there should be no reflecting surfaces within 40cm of a speaker?
Thanks again
( I have my eraser with the safety off just in case )

[tinitus]

Originally posted by Patrick Bateman


For example, you wouldn't want a conventional speaker too close to the walls.

I dont know about other conventional speakers...but my "conventional" speakers doesnt seem to have any problem with sidewalls or any other things around them, none at all...I would say its mostly a matter of crossover, but it may be that its easier with "born" directive transducers like waveguides, ribbons and such

Hiquephon seems to take the diffration issue more seriously than the competition...unfortunately there still is a very small sharp edge left

[gedlee]

Originally posted by HK26147
Excuse Me, Dr. Geddes:
(1) Is it safe to assume that the Hass Effect window is predominate in the auditory region utilized for speech ( approx 170 - 4Khz ).
(2) Is there a perception vs frequency analysis?
(3) The 5 - 10ms window would imply that there should be no reflecting surfaces within 40cm of a speaker?
Thanks again
( I have my eraser with the safety off just in case )

I think of the Haas effect as existing in time not frequency and I'm not sure what data there is on the effect with frequency mixed in. From what I know all auditory effects are at their maximum in the speech range, but I would consider this somewhat higher than 170 Hz. Maybe 300 Hz. 300-3 kHz is what the phone company considered was necessary for speech. I think that they were right.

I would think that most of the time issue are pretty constant with frequency within the speech range, but I am not positive about that.

The 40 cm is a delta - a path length difference. I'd have to calculate what this meant in terms of absolute distance to a reflection. I think that its much larger.

[Patrick Bateman]

Originally posted by gedlee



I think of the Haas effect as existing in time not frequency and I'm not sure what data there is on the effect with frequency mixed in. From what I know all auditory effects are at their maximum in the speech range, but I would consider this somewhat higher than 170 Hz. Maybe 300 Hz. 300-3 kHz is what the phone company considered was necessary for speech. I think that they were right.

Recently I upgraded from three subwoofers to seven, and I noticed an improvement in imaging in my mains. I have a theory about this; bear with me.

Previously there were three subs were mounted in the three corners of the room. They had a cutoff of 80hz. My theory is that reflections from the *subwoofers*, not the mains, could have muddied the image.

Even with an 80hz cutoff there's third harmonic distortion present at 240hz. Even worse, with the subs in the corners that distortion at 240hz gets a nice reflection off the ceiling, two walls, AND the floor!

(I believe this is one of the reasons that subs should be physically separate from the mains, right? Even with a 100db filter at 80hz they will muddy the image due to 2nd and 3rd harmonic distortion.)

Long story short, the imaging on the mains has improved noticeably, and I didn't touch them. All I did was add subs.

[Pano]

Originally posted by gedlee
Controlling the first reflections without reducing later ones is the key.

Makes sense to me!

[gedlee]

Originally posted by Patrick Bateman

Long story short, the imaging on the mains has improved noticeably, and I didn't touch them. All I did was add subs.

John - not what I would expect, thats for sure. Maybe a smoother bass helps? I don't know. If what you say is true then the image should be level dependent. Is that true?