Controlled vs wide dispersion in a normal living room environment..

[graaf]

Yes, but they affect it.

You seem to presume that it can only be to the detriment which in fact is highly subjective

research suggests that some reflections can increase realism of the otherwise artificially and flat sounding stereo image
it is about things like IACC, ASW and LEV, discussed in Dr Toole's book too

even such a prominent advocate of CD speakers as Dr Geddes (just for example) contends that reflections delayed more than 20 ms are beneficial for stereo experiencem, without them the experience is worse, less realistic

well, ok., I know, You probably won't agree that hifi sound is about "realistic sound reproduction"

apparently our goals in all this are just different

Perfect in the sense that the recorded image will be perfectly reconstructed - with all drawbacks which stereo has.

so it is perfect in the sense of imperfect?

I am not talking of 1st order reflections only. You have to take higher order reflections into account, too.

yes, of course, but the main controversy is over 1st order reflections

it is not that difficult to raytrace higher order reflections either

No. Make a test with a an omni and a slightly directive speaker. Make sure that the latter is (nearly) CD (Yes, this is a constraint; non-CD speakers may be inferior to omnis, I was not precise in my statement). Under exactly the same conditions, the directive speaker will have the more precise stereo image and fine detail. The overall sound may not necessarily be better for you, but in these two disciplines, the directive speaker will win.
Baseballbat

believe me I did all kind of experiments and my feelings are quite the opposite, well controlled wide directivity gives a solid and stable, even palpable phantom images (perhaps because of increased ASW) while moderately controlled directivity speaker produces vague and flat phantoms that just fall apart as soon as I move my head

but perhaps really good high and true constant directivity speaker is needed for such a comparison to be fair, but such speakers are in fact very rare (Geddes, Danley)

OTOH I can agree that perhaps some tiny details can suffer, and while I know that it is typical for audiophiles to have this fetish for tiny details this is not the same as quality of spatial reproduction, two different things

for my part while I would like to have both and I enjoy earphones reproduction of tiny details I greatly prefer realism of spatial presentation by loudspeakers over those details

in fact when one attends real live music event or even plays an instrument oneself one doesn't hear most of those little clicks and scratches that might be lost

 

[graaf]

It was filtered to the averaged material recommendations of slight emphasis on midrange output from the front speaker, and large emphasis on rear output in the high frequency range.

"sounds" very much like CLS's setup with flooded tweeters hidden behind midrange baffles

 

[graaf]

Why not make a simple decent set-up of the closest thing we can come to an omni speaker for a reasonable cost, ie. a bipolar.

With a bipolar speaker you can vary the amount of output from the rear speakers very easily
...

yeah! such a bipolar is indeed best for tests but it is not "the closest thing we can come to an omni speaker for a reasonable cost" - a flooder is

please tell more about those tests done that You have "hinted" at

 

[Baseballbat]

research suggests that some reflections can increase realism of the otherwise artificially and flat sounding stereo image
it is about things like IACC, ASW and LEV, discussed in Dr Toole's book too

Yes, of course. But it affects the image that is on your record.

If you listen to a headphone, it will have a pinpoint stereo image (which is also a wrong one unless the headphone is not perfectly matched to your ears). If you add artificial reflections to it, the stereo image will blur. This might sound more "realistic", but is not "correct" in means of perfect reconstruction of the signal on your record. That is the main drawback of stereo (and surround) recordings, that you have to recreate the complete sound field with a limited (i. e. too low) number of sources.

well, ok., I know, You probably won't agree that hifi sound is about "realistic sound reproduction"

No, I like "realistic sound reproduction" more than perfect reproduction. Because "perfect reproduction" uncovers all flaws of stereo, while "realistic reproduction" sounds more natural.

But maybe the word "realistic" has different meanings for us?

so it is perfect in the sense of imperfect?

Yes.

while moderately controlled directivity speaker produces vague and flat phantoms that just fall apart as soon as I move my head

Yes, but that's stereo. Except of the "vague" phantoms; I always perceive them more precise (pinpoint like with headphones), but of course they fall apart and they are not realistic. That's one reason why I do not have a very high directivity speaker at home, but with a moderately controlled vertical directivity, while the lateral directivity is very low. I deliberately traded "precision" for "naturalness", and I'm fine with it. I tried very wide dispersive speakers before, and that was too much "naturalness" and not enough "precision". And these findings are what many other people I know feel.

Of course, an omni is good if you are close enough and/or your room is very dry. In fact, if I had to built some nearfield speakers, I would go to omnis.

but perhaps really good high and true constant directivity speaker is needed for such a comparison to be fair, but such speakers are in fact very rare (Geddes, Danley)

It is quite easy to build a 2-pi-speaker (big baffle, 5" woofer, 2" midrange (e. g. Aura NSW2, do not use a dome), 0.75" tweeter). It will win against an omni (4-pi).

Baseballbat

 

[markus76]

I strongly disagree with that, and I sincerely believe you would too given the chance to test it.

Maybe it's time to make this more scientific instead of just quoting works of other, and me also vaguely referring to research I cannot disclose.

Why not make a simple decent set-up of the closest thing we can come to an omni speaker for a reasonable cost, ie. a bipolar.

With a bipolar speaker you can vary the amount of output from the rear speakers very easily by just dividing the cabinet in two parts, so the woofers are in separate cabinets as would be required. Whether it's split down the middle horizontally or vertically wouldn't matter.

When the speaker is built each pair just needs to be driven by a separate amp and now you can perform all sorts of test (which I as hinted have already done on numerous test subjects).

The beauty of this set-up is that it's basically impossible for the listener to have any knowledge on when the set-up is changed as all since all that is required is to switch to a different balance and filter setting. Prior calibration must be made to ensure that total output at the listening position is the same in all settings.

Not sure how the speaker you describe should look like. Would a 2'-3' baffle, very shallow box with a small fullrange at the front and one at the back suffice? Guess a sub would need to complement the lower frequencies? Or is directional bass also important per your "research" (by the way, one of the requirements for real scientific research is "full disclosure"!).

Then I still don't know how such a speaker would satisfy most people that listen to music, TV and movies over the very same speakers. You've said that different types of signals had different D/R preferences (http://www.diyaudio.com/forums/mult...al-living-room-environment-2.html#post3165745).

By the way, is it just D/R or is level, direction, delay and spectrum of single reflections important too? After all the sound field in acoustically small rooms is dominated by discrete reflections and not by a diffuse (homogeneous and isotropic) sound field.

 

[graaf]

If you add artificial reflections to it, the stereo image will blur. This might sound more "realistic", but is not "correct" in means of perfect reconstruction of the signal on your record.

but it is a better reconctruction of the real thing and isn't the point of hifi reconstruction of the real thing? is a recording a thing in itself or just one of the means used in the process?

But maybe the word "realistic" has different meanings for us?

why? rather I believe that our necessarily limited experience with loudspeakers and rooms is different so far

Yes, but that's stereo. Except of the "vague" phantoms; I always perceive them more precise (pinpoint like with headphones), but of course they fall apart

I mean "vague" in the sense of "unstable", they seem constantly to be on the verge of falling apart because human head moves unconsciously when human sense of hearing is "tracking" a sound source

I do not have a very high directivity speaker at home, but with a moderately controlled vertical directivity, while the lateral directivity is very low.

so what's the difference -with regard to laterally reflected sound - between Your speakers - one with a 180 degrees uniform coverage laterally (as far as I understand) and a speaker that is 360 degrees? is it just the presence or absence of front (behind the stereo base) wall reflection ?

It is quite easy to build a 2-pi-speaker (big baffle, 5" woofer, 2" midrange (e. g. Aura NSW2, do not use a dome), 0.75" tweeter). It will win against an omni (4-pi).
Baseballbat

I understand that it will win for You, subjectively?
Objectively, technically, do You think that front wall 1st order reflection is more detrimental to sound reproduction than diffraction "noise" of such a 2-pi speaker on a wide baffle?
Most of the research done seem to suggest quite otherwise.

ps.

But there is a simple solution to this: avoid reflections.

this is certainly not what You proposed above, You avoid just a front wall reflection, that's all

 

[Baseballbat]

I mean "vague" in the sense of "unstable"

Then I agree.

so what's the difference -with regard to laterally reflected sound - between Your speakers - one with a 180 degrees uniform coverage laterally (as far as I understand) and a speaker that is 360 degrees? is it just the presence or absence of front (behind the stereo base) wall reflection ?

No. My speakers do not only have no (less) first order reflection from the front wall, but also a reduced vertical reflection and higher order reflection pattern, because of their overall higher directivity. I know that first order front wall reflection does improve stereo image localization, but we do not have only first order reflection in rooms.

Objectively, technically, do You think that front wall 1st order reflection is more detrimental to sound reproduction than diffraction "noise" of such a 2-pi speaker on a wide baffle?

If you carefully design the 2-pi speaker, you have only diffraction "noise" from the nearby drivers. A quite good design is this: http://www.grimmaudio.com/whitepapers/speakers.pdf although I would have chosen at least an odd order crossover to avoid the power response dip.

this is certainly not what You proposed above, You avoid just a front wall reflection, that's all

Yes. And half of the higher order reflections.

Baseballbat

 

[markus76]

I know that first order front wall reflection does improve stereo image localization, but we do not have only first order reflection in rooms.

What specific direction of the front wall reflection do you have in mind? How do you define "improved image localization"? Increased image focus (less blurry/vague/ambiguous) or added spaciousness?

 

[graaf]

No. My speakers do not only have no (less) first order reflection from the front wall, but also a reduced vertical reflection and higher order reflection pattern, because of their overall higher directivity. I know that first order front wall reflection does improve stereo image localization, but we do not have only first order reflection in rooms.

still I can't see how You can agree with Remlab who claims that sidewall reflections are worst enemies of good imaging

for my part I appreciate restricted vertical dispersion because I believe that floor and ceiling reflections (from conventional loudspeaker, not a ceiling flooder which probably functions as a kind of virtual ceiling speaker) are the worst of them all
results of research done in the Archimedes project seem to confirm this

 

[Baseballbat]

What specific direction of the front wall reflection do you have in mind? How do you define "improved image localization"? Increased image focus (less blurry/vague/ambiguous) or added spaciousness?

If I talk of first order front wall reflection, there is no big selection, hm?

In this case, "improved image localization" means more focus.

graaf,
I didn't see the post of Remlab about sidewall reflections, and was only referring to general reflections. However, I agree to this, and stated it already. In my room, the wide lateral dispersion works quite good. In more reverberant rooms, or with a larger listening distance, it is probably too much dispersion.

Baseballbat

 

[markus76]

If I talk of first order front wall reflection, there is no big selection, hm?

In this case, "improved image localization" means more focus.

Well then I have to strongly disagree. Those reflections are early and strong. They will shift (phantom) sources towards their direction and color the direct sound. This trading effect will also negatively affect image focus.
Toole specifically recommends to treat the front wall because of this. And he's not the first researcher in psychoacoustics to do so.

This can easily be tested with a second speaker playing the same signal as the first. Then adjust level and delay of the second speaker.

 

[Baseballbat]

I may have misunderstood it. Can you give me the page where Toole mentions it in his book?

 

[markus76]

I may have misunderstood it. Can you give me the page where Toole mentions it in his book?

Chapter 8, page 116.

I don't think any reflections can help with improving localization focus for phantom sources as those reflections will always deliver spatial cues that point to the speaker itself and not in the direction of the phantom source.

 

[Baseballbat]

Indeed, you are right. He recommends adding absorbing material to the front wall.

Hm. I'm sure I've read somewhere about the benefits of front wall reflections. Might be a wrong remembrance.

 

[qusp]

I have to thank all those who have contributed here so far, great timing as i'm just finalizing the design of my baffles and thinking about setting up my small quite difficult room. even though i'll be using digital XO and convolving, I want to get it as 'right' as I can acoustically first.

 

[markus76]

Indeed, you are right. He recommends adding absorbing material to the front wall.

Hm. I'm sure I've read somewhere about the benefits of front wall reflections. Might be a wrong remembrance.

One can read all sorts of things in these threads

As I've said before, there might be beneficial constellations for REAL sound sources as reflections might provide a "second view" on the original event (see speech intelligibility tests). They might also deliver cues to the auditory system about the location of the sound source but this is not true for phantom sources created by two (or more) speakers. In that case the reflections would only reveal information about the speaker's locations.

 

[graaf]

Well then I have to strongly disagree. Those reflections are early and strong. They will shift (phantom) sources towards their direction and color the direct sound.

Can you give me the page where Toole mentions it in his book?

Chapter 8, page 116.

Indeed, you are right. He recommends adding absorbing material to the front wall.

but not because of any "image shifting"

indeed Markus is right that:

One can read all sorts of things in these threads

unfortunately it applies to Markus' statements regarding "image shifting" too

fortunately we can also just post some little scans from Dr Toole's book, to clarify things a bit, for educational purposes
see below, so much for Markus' theories of image shifting caused by normal reflections in a normal room

Hm. I'm sure I've read somewhere about the benefits of front wall reflections. Might be a wrong remembrance.

there are benefits from higher order reflections that originate in first order front wall reflections

 

[graaf]

ps.
three more ("image shift/broadening or spreading" is not at all the same as defocusing/blurring, and this is not summing localisation Markus):

 

[6.283]

Why not make a simple decent set-up of the closest thing we can come to an omni speaker for a reasonable cost

That is pretty much what I did and invite everybody to read about the results (if not known already); for what it's worth.
In my view these speakers are much better than you could think from comments, papers, wild guessing, you name it. Please also follow the links.

Demokrit

Now working my way towards a version with a pinch more directivity and the changes are very interesting indeed. More to follow here soon.

Demokrit-T

 

[FrankWW]

Problems in typical living rooms -

the absolute SPL of reflections is much higher than in normal performance spaces.

much less audible decorrelated sound than performance space

paucity of modes compared to performance space