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

[graaf]

Just compared to wavelength.
There are omnis that are rather big like German Physiks. I am not convinced that they provide the full omni experience without other side effects.

perhaps, but why not?

Not exactly sure what you mean but the direct sound diminishes anyway. Flatter though than 6db per double distance like shown in Toole's.

yeah, You're right, I understand, I misunderstood it at first because the comparison between direct SPL and reverberant SPL is abstract, it is a comparison between SPL under anechoic and reverberant conditions

in fact there is no such perceiveable thing as "direct sound SPL" in a listening room at all, there is only the first wavefront and then the SPL is perceived in ~30 ms "samples" - this is why the reverberant field is indeed stronger perceptually as far as frequency related qualities - pitch and timbre - are concerned

In my room an setup, 1.8m is the max. listening distance with Demokrit. And based on several estimations, the critical distance for an omni is indeed below 1 m in my room. So the example you posted is not unrealistic at all and shows exactly the point that I was mentioning about distance.

I hoped it's not unrealistic

still I don't know what is going on after a certain distance in a listening room is exceeded? in other words - what happens when reflected SPL exceeds direct SPL (theoretical - anaechoic) by more than 5 dB? What sort of degradation of quality of reproduction can be experienced and exactly why?

 

[markus76]

The precedence effect according to puplished literature does not explain this.

It doesn't explain this because it's simply not a case that was ever studied in detail. Nonetheless studies know the "breakdown of precedence" and this is exactly what can happen when we put a speaker in a room. Reflections are likely perceptually lumped together which obviously can cause a breakdown of precendence. Additionally there's a multitude of other factors. Lot's of open questions

 

[graaf]

This can change if we hear recorded reverberation being reproduced over loudspeakers. It is more difficult to "de-reverberate" the recorded sound when reproduced than in the live setting

yes indeed - and exactly therefore it is easy to hear all recorded spatial cues through a listening room acoustics which is unable to dominate or mask them

see also:http://www.diyaudio.com/forums/multi-way/121385-loudspeakers-room-system-35.html#post2077643

Simple Dereverberation Experiment
Last Summer, I visited Michigan State professor William Hartmann who is a well known psychoacoustics researcher.
(...)
He suggested we do a little experiment right there.
(...)
Experiment’s conclusion: recorded reverberation is far more audible than live reverberation, at least for a simple recording technique.

 

[graaf]

Toole is next to God, right ?
- Elias

wrong! You got it all wrong! it's Blauert!

Toole's book is just a body of data

 

[Melo theory]

I know the feeling I propably should stop writing at this thread and go back to my own thread lol

haha!
Well, the problem with psychoacoustics is that there is no hard evidence confirming what people perceive.
Toole, linkwitz, and plenty of others are peddling their theories just like we are. The only difference is that they wrote books or white papers or have convincing arguments on there websites, so we refer to them as if they are fact, but we have to because we can't refer to our own perception.
That's why these types of threads are so long and inconclusive.
For what it's worth, I agree with you that the precedence effect is not as prevalent in small rooms, "quick" reflections will be perceived as a widening of the image. This is my argument against line sources. (not that I don't like them)
Also, as per your own matrix invention, steering sound into a diffused field will result in shifting the image away from the source.

 

[graaf]

i would liken it to playing electric guitar wearing headphones. Then i would add a small pinch of reverb, just enough so that it isnt 'anechoic' in the headphones. Then remove the headphones and play through the speaker instead. The sound doesnt sound the same.

it's a monophonic forward radiating highly directional big guitar speaker, isn't it?

 

[graaf]

For a mind teaser, take a look at the attached wavelet plot from a small room. There is a direct sound from a single loudspeaker at 0 ms, and first strong reflection comes at 5 ms and after that comes multitude of reflections. According to precedence effect we would expect the perceived direction to be at the location of the speaker. But, that is not the perception. In this case there is no perceived direction at all, the sound is coming from somewhere at frontal hemisphere, the location cannot be expressed more accurately than that. The precedence effect according to puplished literature does not explain this.

but the direct sound and the reflection are in this case significantly different, therefore precedence effect probably doesn't apply
perhaps different and partly conflicting and at the same time equally strong cues from both - the original and the secondary - sound sources can create such an effect of "somewhere there"?

 

[Melo theory]

perhaps different and partly conflicting and at the same time equally strong cues from both - the original and the secondary - sound sources can create such an effect of "somewhere there"?

This is also a problem of stereophonics itself.
It cannot produce a convincing center image at higher frequencies.
Even though the signal is produced at the same time, it widens the image due to comb filtering.
This is why I don't understand why stereophonics is so easily accepted.
If I built a 6 ft tall loudspeaker and placed 2 tweeters, 1 on each end, 6 ft apart, it would be blaspheme!

 

[Kindhornman]

One thing that we should do is take devices and not necessarily categorize it into a specific category and not see that it is a design choice. I could easily produce a horn that is omni instead of directional, I have seen designs in threads here on the DIY forum where that was the design concept. So we should not make blanket statement. but qualify the individual applications. Yes most horn systems are designed as directional devices, but many times the off axis polar responses are anything but. I would say that same would go for a dipole or bipolar response characteristic. Depending on the baffle size and the directionality of the devices used that may not always be the case either. In general they will follow certain rules but not when we look at specific implementations. An 18" cone driver on the same baffle as a smaller driver could have extremely different polar responses and then you are back to the same situation as a small monitor in a bass reflex enclosure and the spreading dispersion patterns of each device. I am not arguing with anyone here, just saying we need to be careful with generalities.

 

[Pano]

Does this one need to be split? Just say the word.

 

[Kindhornman]

Pano,
How would you split this discussion. Polar responses of differing speaker types and implementation and the original discussion of Controlled vs wide dispersion in a normal living room environment?

Steven

 

[6.283]

I think omnis belong to this handicapped category as well, since they suffer from distance problems.

Hey, where's your "reinventing the wheel" sloagan gone ?
If you say this then you might as well blame the room.

perhaps, but why not?

The perfect omni is what ? A pulsating sphere of infinitesimal small size. If a transducer emits sound of let's say 1" wavelength to the front, the back and the sides but the acoustical sources on the surface of the transducer are separated by several inches you will have timing issues at least and not only in the reverberant field where it is less severe but also in the direct sound. It becomes a smeared representation.
If you happen to play the same freq. range over two transducers (cross-over region) and the two sources are too far apart then you see phase issues that cannot even be fixed with real time delay.

still I don't know what is going on after a certain distance in a listening room is exceeded? in other words - what happens when reflected SPL exceeds direct SPL (theoretical - anaechoic) by more than 5 dB? What sort of degradation of quality of reproduction can be experienced and exactly why?

Speech intelligibility suffers, bass becomes boomy, muddy and single note bass and the rooms sonic signature starts to dominate the perceived sound. You hear your room with a speaker placed in it. The sound field is too diffuse. There is less clarity and focus, the details seem to be gone or diluted and the phantom images fade out.
In short, with increasing distance it becomes more and more sound field soup and it fieels like swimming in it.

 

[graaf]

Polar responses of differing speaker types and implementation and the original discussion of Controlled vs wide dispersion in a normal living room environment?

Elias meant "conventional vs unconventional" split, it happened before:

http://www.diyaudio.com/forums/multi-way/195124-what-ideal-directivity-pattern-stereo-speakers.html
http://www.diyaudio.com/forums/mult...ques-achieving-oustanding-stereo-imaging.html (quite sarcastic, isn't it?)

some people feel painfully uncomfortable about anything unconventional

 

[graaf]

I think omnis belong to this handicapped category as well, since they suffer from distance problems.

Hey, where's your "reinventing the wheel" sloagan gone ?
If you say this then you might as well blame the room.

exactly but it is not that omnis "suffer" from "a weakness" and this is not a question of a particular pattern of 1st order reflections from adjacent side wall, front wall etc.

IF as a rule (following from the results of Klippel's experimental study) the sound quality deteriorates when the distance from a speaker exceeds twice the critical distance then it's a matter of that particular room average absorption coefficient because it's a matter of the critical distance itself - so if You want to listen to omnis from a greater distance then You need to introduce some more absorption into the room - but not at all in the points of first order reflections

The perfect omni is what ? A pulsating sphere of infinitesimal small size. If a transducer emits sound of let's say 1" wavelength to the front, the back and the sides but the acoustical sources on the surface of the transducer are separated by several inches you will have timing issues at least and not only in the reverberant field where it is less severe but also in the direct sound. It becomes a smeared representation.

yes, agreed - time coherence is very important - I believe that for an omni to work all 1st order reflections must suffciently preserve the shape of the first wavefront and therefore polydirectional speakers like Duevel Planets for example - mistakenly called omnidirectional - tend to suffer from lack of focus

so several inches can be a problem, those big German Physics are particularly big

Speech intelligibility suffers, bass becomes boomy, muddy and single note bass and the rooms sonic signature starts to dominate the perceived sound. You hear your room with a speaker placed in it. The sound field is too diffuse. There is less clarity and focus, the details seem to be gone or diluted and the phantom images fade out.

sounds like it is partly a bass related problem really because muddy bass can really spoil everything else including imaging

In short, with increasing distance it becomes more and more sound field soup and it fieels like swimming in it.

I still wonder what - besides influence of deteriorated bass - causes all this?

 

[a_tewinkel]

Precedence effect. There are numerous problems in current psychoacoustic studies and their applicability in living room environment as is the title of this thread. Almost all of the perception studies with numerical data employ only one single reflection in addition of direct sound. This is not the case in a small room. What the precedence effect does not tell you is what happends when there are multiple of reflections in a short time window. That is the small room case. I could even go to great lengths and declare the applicability of precedence effect in small room is questionable at best.

For a mind teaser, take a look at the attached wavelet plot from a small room. There is a direct sound from a single loudspeaker at 0 ms, and first strong reflection comes at 5 ms and after that comes multitude of reflections. According to precedence effect we would expect the perceived direction to be at the location of the speaker. But, that is not the perception. In this case there is no perceived direction at all, the sound is coming from somewhere at frontal hemisphere, the location cannot be expressed more accurately than that. The precedence effect according to puplished literature does not explain this.

Well, you have a point here about the majority of experiments being done with a single reflection.

But I think that according to most research, such strong localization blur as described here only occurs when there are multiple, relatively strong reflections < 5 ms or so. Above that, even in a small room, the amplitude of early reflections will be sufficiently reduced to allow "reasonable" localization.

There are, however, some experiments which have involved multiple early reflections. You're probably familiar with the Bech experiments? They do not really give us a theoretical explanation, but they do give us practical results to work with. For me, that's good enough for now. More research would certainly be interesting however

 

[6.283]

exactly but it is not that omnis "suffer" from "a weakness" and this is not a question of a particular pattern of 1st order reflections from adjacent side wall, front wall etc.

But omnis make the sound field more diffuse to begin with.

IF as a rule (following from the results of Klippel's experimental study) the sound quality deteriorates when the distance from a speaker exceeds twice the critical distance...

Watch out, this is not what Klippel says ! He says the optimum pleasantness and naturalness is at 5dB reflected/direct for music. It takes more distance until it finally deteriorates as described. Before it is less pleasing.

then it's a matter of that particular room average absorption coefficient because it's a matter of the critical distance itself - so if You want to listen to omnis from a greater distance then You need to introduce some more absorption into the room - but not at all in the points of first order reflections

That is what I say in my "what is the ideal radiation pattern...."
If you select a radiation pattern based on preference you might need to deal with consequences. If you choose to drive an Audi Q7 then you should think about the garage.

I believe that for an omni to work all 1st order reflections must suffciently preserve the shape of the first wavefront...

and the reflections should be at a lower level and sufficiently delayed.

I still wonder what - besides influence of deteriorated bass - causes all this?

The missing level of direct sound as reference, which is responsible for phantom sources and ensures that the precedence effect is at play ?

 

[DBMandrake]

Interesting observation I made just today: I can hear the difference between correlated and inverted (one channel) HF pink noise (HP 48dB 4kHz) with speakers whereas I can't hear it with headphones.

Here're the test signals I've used:
https://dl.dropbox.com/u/21936387/Pink-Noise_stereo_correlated_HP48dB4kHz.wav
https://dl.dropbox.com/u/21936387/Pink-Noise_stereo_inverted_HP48dB4kHz.wav

Can you hear a difference on the headphones if you low pass filter the noise to below 1Khz ?

It's well known that we can't detect inter aural phase differences of non-transient sounds above about 2Khz. (try a 3 KHz tone in one ear, then another 3Khz tone in the other ear with variable phase from the first ear and you won't hear any difference even if the phase of one is inverted)

Perhaps with broadband noise the lack of phase discrimination of the high frequencies dominates the perception and overrides the small cue of low frequency phase correlation. If so low pass filtering should make it audible.

The reason you hear a difference on speakers is comb filtering affecting the frequency response...

 

[graaf]

But omnis make the sound field more diffuse to begin with.

therefore more absorption needs to be introduced into the listening room to balance it but it doesn't mean that absorption should be put in the areas of 1st order reflections

Watch out, this is not what Klippes says ! He says the optimum pleasantness and naturalness is at 5dB reflected/direct for music. It takes more distance until it finally deteriorates as described. Before it is less pleasing.

I know - it is what You say and Your experience happens to correspond somewhat to what with Klippel found out during his tests

If you choose to drive an Audi Q7 then you should think about the garage.

OTOH buiyng thicker curtains or carpet or a couch with softer upholstery is much less demanding

and the reflections should be at a lower level and suddiciently delayed.

You mean specific 1st order reflections? But does it follow from any studies done? Because studies done and quoted in Toole for example and also in this thread suggest that typical room reflections are not strong enough to introduce anything else than increased sense of spaciousness regardless of delay

The missing level of direct sound, which is responsible for phantom sources and ensures that the precedence effect is at play ?

but there is no "level of direct sound" in a listening room AT ALL, You cannot perceive anything like "direct SPL" in a room, even with Your head stuck into a speaker You always perceive just a combined direct/reflected SPL sampled in ~30 ms intervals

comparison is theoretical - between anaechoic level of direct sound and the level of sound in a listening room

 

[graaf]

There are, however, some experiments which have involved multiple early reflections. You're probably familiar with the Bech experiments? They do not really give us a theoretical explanation, but they do give us practical results to work with.

and what are those practical results?

actually they are that Bech is another prominent advocate of wide directivity - he was the leader of Archimedes research project and now is the Head of Research at Bang & Olufsen

praised here on diyaudio for His scientific publications Bech is in fact the man behind B&O-Moulton coooperation and Beolab line of 180 degress uniform horizontal directivity speakers

 

[markus76]

Can you hear a difference on the headphones if you low pass filter the noise to below 1Khz ?

Yes sure.

The reason you hear a difference on speakers is comb filtering affecting the frequency response...

It's not that there is a difference (which is obvious why) but that it's the same "out of phase" sensation I get from LF noise <2kHz.