I hear you but even if important, narrow vs. wide is only one aspect. The question is:
What is the ideal directivity pattern for stereo speakers?
So where does narrow end ? 120° (+-60°) or 180° ?
What about dipoles ? Are they narrow or are they "wider" perceptionally because of the rear radiation ?
What about cardio and what about vertical behavior ?
Should the directivity be constant or should it end in any kind of forward firing behavior ?
Questions, questions, questions.
-We have the direct sound comming from the instrument directly into the mic and from speaker directly to listener. This is the same for all situations.
-Then we have the recorded reflection. This is a indirect sound picked up by the mic and then reproduced by the speaker and directly to the listener. This is still the same for all situations (direct/omni/etc..) For good sound we need these reflections to be diffuse, frequency independant and long enough/short enough for the music style. Iaw: The recording acoustics need to be good. So far I think all can agree.
-Now we have the listening room reflections of the direct sound and recorded reflections.
So at time X we hear a reflection and at time X + Y we hear an other reflection. One is comming from the listening room the other is in the recording. This is a extremely simplified example in the real world we would have many many more reflections. But it makes this more eazy this way.
Questions to ask then are:
-In what situations can we differenciate between these two?
-In what situations can not differenciate between these two?
One solution to the last question: If we eliminate (up to a point) all listening room reflections so we can only hear the recording. This is the conventional way of dealing with it, and it works.
But is this realy the only one possible solution that works?
If we make sure the listening room reflections have the same properties/characteristics of the recorded reflections, how can we possibly differenciate between the two?
-Then we have the recorded reflection. This is a indirect sound picked up by the mic and then reproduced by the speaker and directly to the listener. This is still the same for all situations (direct/omni/etc..) For good sound we need these reflections to be diffuse, frequency independant and long enough/short enough for the music style. Iaw: The recording acoustics need to be good. So far I think all can agree.
-Now we have the listening room reflections of the direct sound and recorded reflections.
So at time X we hear a reflection and at time X + Y we hear an other reflection. One is comming from the listening room the other is in the recording. This is a extremely simplified example in the real world we would have many many more reflections. But it makes this more eazy this way.
Questions to ask then are:
-In what situations can we differenciate between these two?
-In what situations can not differenciate between these two?
One solution to the last question: If we eliminate (up to a point) all listening room reflections so we can only hear the recording. This is the conventional way of dealing with it, and it works.
But is this realy the only one possible solution that works?
If we make sure the listening room reflections have the same properties/characteristics of the recorded reflections, how can we possibly differenciate between the two?
yes, and a lot to talk about and to experiment with, lotta fun
first of all I think that the ideal directivity pattern for stereo speakers should be so that shape of the direct and all reflected onsets is basically the same - I think this is the reason why many conclude from their experiments/research that it is important to have the direct and reflected sound sufficiently spectrally similar - but it is in fact not about frequency spectrum as such but of what the soundwave composed of it looks alike - of the onset shape:
furthermore I think that the ideal directivity pattern for stereo speakers should be so that the floor reflection is eliminated/muted - and I am not talking about floor bounce dip and peaks - effects of the bounce on frequency response - I mean floor reflection as a spatial cue for the distance and height perception
as such a cue floor reflection unmasks loudspeaker as separate real sound source besides virtual sound sources - this cue cannot be filtered out becasue it is incongruous with distance and height cues of virtual sound sources that are in the recording
I think this is the reason why many conclude from their experiments/research that it is important to get rid of (or at least significantly mute) the floor reflection - by means of waveguides, line sources, line arrays etc - but again - it is in fact not about frequency spectrum effects but of incongruous spatial cue
furthermore I think that the ideal directivity pattern for stereo speakers should be so that it produces low IACC for reasons explained many times here on the forum (also in this thread IIRC)
furthermore I think that the ideal directivity pattern for stereo speakers should be so that it makes localization/percepction of the loudspeakers as real sound sources separate from virtual sound sources as difficult as possible by - besides of eliminating/muting of the speaker's distance/height cues from floor reflection - eliminating/muting cues for localization of the speakers in the lateral plane with the use of some kind of CTC (cross talk cancellation/muting like in Ambiohonics) or by integrating of the speakers into the room sides (Beveridge) or "stereolitically" (see Elias' stereolits) - in additional consequence the position of the speakers does not limit the spread of the soundstage anymore
That is not a correct conclusion. As pointed out, the numerical rankings were inconclusive (and Toole was being a little devious in his quote). They neither disproved degradation of imaging with wide dispersion, nor did they show that spaciousness was increased. Finally, the naive listeners that prefered omni dispersion in the forced choice test didn't rank it numerically higher.
Also pointed out previously, the professional group was fairly consistant in their views for a given musical type, preferring wide dispersion on some selections and focused imaging on others.
Occupational disease? Yes, a preference for more focused imaging when appropriate.
Putting your own spin on the facts doesn't prove your case, nomatter how frequently or "loudly" you repeat it.
David S.
no it doesn't because it is like throwing out the baby with the bath water
in that case the baby is spaciousness
if we assume (as You propose) that we have good recording then the listening room, which is an acostically small room, produces early reflections with delays < the delays we have in a real performance
so question is what happens when we have direct sound of the original direct sound plus it's copy reflected in our room and delayed by say 2 ms and then direct sound of the original reflection/reverberation delayed by say 20 ms plus it's copy reflected in our room and delayed by say additional 2 ms
answer is - we do not hear the room reflection as it is perceptually dominated by the original reflection as Dave Moulton put it:
(sorry for reposting it for a 1000 time here)
Moulton Laboratories :: Nick Batzdorf Interviews David Moulton
therefore Moulton asserts that the smaller the room the better - yes! (in that particular regard of course)
I think the above misuses the terms wide and narrow.
The notion of narrow directivity refers to beaming issues. When the researchers refer to wide-dispersion loudspeakers, they are not referring to omnis or dipoles, they are referring to forward-firing speakers that don't have beaming issues as the frequencies rise.
Omnis and dipoles are considered to be 'special cases', not 'wide dispersion'.
So if you think the research is saying "some sort of wide pattern creating the appropriate pattern of early reflections is needed", that simply means normal forward-firing cone-dome configurations.
180 is definitely wide, 120 is near the wide border. Narrow is more like 60.
Dipoles are a special case. Most dipoles beam significantly and don't emit to the side walls and by those two criteria they are narrow. However a well performing constant 120 degree dipole loudspeaker would be counted as wide in the treble.
The desirable vertical pattern is similar to the desirable horizontal pattern.
The general assumption is that the speakers are forward firing in the treble.
Arg
?? don't You know that in science until something is proven it is not proven? we cannot say "yes", can We?
exactly therefore the answer is "no"
perhaps the quote was devious but taken together with other opinions that Toole expresses in His book His position is clear - the preference for high directionality/low level of reflection among some professionals is an occupational bias (if You are offended with the world disease) and has nothing to do with alleged law of naturethat "reflections destroy imaging"
the mantra didn't stand the scientific test, it was not proven
please send complaints to Dr Toole - my understanding of the results of this study is simply His, it is His spin on the facts, and I somehow tend to believe that He is more impartial in this regard than You
Last edited:
?
the phrase "wide pattern" was not used in that particular post in any particularly defined meaning taken from any research papers, and I don't know that the meaning of it is restricted to "normal forward-firing cone-dome configurations", in that case I apologize for misunderstanding
OTOH I don't feel particulrarly guilty as it seems to me we discuss the question of directivity pattern in context of reflections pattern it produces, because this is it that matters in the end
and back wall reflection is the only thing that differs forward radiating proper wide directivity speakers differ in that regard from separate case omni speakers in that regard
Do You really think this justifies treating them as "separate case"?
and what difference does it make? Are accuracy requirements different in a mixing room and in a listening room at home?
I'm still a bit of a fence sitter when it comes to the question of whether the reflections should have a similar spectral balance to the direct signal or not, and whether it's really that important. I haven't made up my mind yet.
I'm not sure that it is important to perceived sound quality as such if there is some variation between on and off axis response, as long as the variation is not huge, and stays within a certain window. For example a 3dB power dip in the low treble at a crossover frequency is probably not important, whilst a DI shift of over 10dB from 2Khz to 10Khz (often seen in large FR drivers) is certainly an issue, even though its generally a monotonic increase. Such a speaker can still sound great, but will be very sensitive to room set-up/damping, speaker positioning and toe in etc, and may require some equalization for best perceived response when there is such a large shift in the DI.
If we are to assume that it is important though, and that we also want to reduce the amplitude of early reflections to a manageable level, that leads towards the constant directivity approach at high frequencies, with 90 degrees (horizontally) often being suggested as optimum. I don't think there is anything magic about 90 degrees specifically - probably anywhere between 60 and 120 degrees is fine, as long as its relatively constant and the off axis response is spectrally similar to the on axis response. (Albeit lower in level)
Now this I think we can agree on... the floor bounce effect, at all frequencies, is a lot more important than it's given credit for. I've always tried to eliminate it as much as possible, and I really do think doing so helps to disguise the true location of the speakers (in height and depth) and let the virtual cues in the recording come to the fore.
I like the approach of frequencies below the Schroeder frequency being produced very close to the floor, (no floor bounce reflection is possible if the sound source is at the floor) coupled with a directional midrange and tweeter which are both quite high off the floor at ear level, which minimize floor bounce by their directivity.
(A ribbon tweeter is particularly effective here with its extra vertical directivity, as their response at the steep angle of incidence of the floor reflection is greatly attenuated - same for the ceiling reflection)
I think both are important - spectral errors from comb filtering at low frequencies, and spatial cues at higher frequencies. Both need addressing.
Why ? Reflections from the ceiling tend to increase IACC.
There is some advantage in a "narrower vertically than horizontally" approach IMO.
Last edited:
yes, I agree completely, it was an omission on my part, I should have posted "it is in fact not just about frequency spectrum effects but of incongruous spatial cue"
What else would be the context for discussion ?
It is about "preferred" reflecions and their being beneficial/detrimental.
Ask Moulton, not me. Although he may not have expertise on home listening rooms. My guess is that they may well be quite different.
Reflections definitely change imaging in Toole's expressed view, by broadening it. And for stereo it may be an advantage that partially compensates for the format's limitations in terms of spaciousness and envelopment.
Not just pattern, also quantity. It is possible to have too much reflection even if the pattern is acceptable.
For sure.
Really ?
Horizontal reflections are known to give a sense of spaciousness, but what is the advantage to reflections from the floor and ceiling ? Nothing that I've read has ever suggested that floor or ceiling reflections are good, or anything other than harmful. Reducing floor and ceiling reflections with directivity has always sounded better to me.
You can debate about how much horizontal reflection you want from the side-walls (spaciousness vs imaging precision etc) and that may be a factor of taste combined with type of music, but I thought we were all agreed that early floor/ceiling reflections were bad?
Why is equal horizontal and vertical directivity desirable ? Our hearing system doesn't respond the same to vertically displaced reflections as it does to horizontally displaced ones, walls and floors are typically a different distance away from the drivers in a speaker, the list goes on.tnargs said:
Equal directivity in both axes can work ok, but I wouldn't say it's optimal, and there is a case to be made for more vertical directivity than horizontal, which is certainly something line arrays have in their favour...
I'm going to go out on a limb here in this whole debate and suggest that constant directivity 90 degrees horizontally and 40 or 60 degrees vertically both from about 1-2Khz up may be fairly close to "optimal" for most listening rooms, and still give a good degree of flexibility in results (spacious or precise) with toe in adjustment.
Last edited:
Somewhat shallower, sure, but the overall pattern should still be similar to the direct radiation.
My suggestion would be
Below Schroeder frequency
--------------------------
Everything able to achieve balanced modal
excitation with respect to the listening area.
(dipoles, multiple monopole subs, configurations making use
of decorrelation, ...)
From Schroeder frequency to "crossover region" (ITD dominance vs. ILD,
---------------------------------------------------------------------
HRTF, envelope ...)
-------------------
- narrowed vertical dispersion.
- no floor bounce
- horizontal dispersion compatible with room (the more live the room, the narrower the radiation pattern)
- phase coherent radiation
Above ITD/ILD (envelope, HRTF, ...whatever) crossover region
------------------------------------------------------------
- diffuse radiation
- average DI matching the midrange
- narrowing towards top end intended, otherwise some downshelving might be needed
- detection thresholds for group delay (direct sound) have to be undershoot, as well as in the audible band as a whole.
Below Schroeder frequency
--------------------------
Everything able to achieve balanced modal
excitation with respect to the listening area.
(dipoles, multiple monopole subs, configurations making use
of decorrelation, ...)
From Schroeder frequency to "crossover region" (ITD dominance vs. ILD,
---------------------------------------------------------------------
HRTF, envelope ...)
-------------------
- narrowed vertical dispersion.
- no floor bounce
- horizontal dispersion compatible with room (the more live the room, the narrower the radiation pattern)
- phase coherent radiation
Above ITD/ILD (envelope, HRTF, ...whatever) crossover region
------------------------------------------------------------
- diffuse radiation
- average DI matching the midrange
- narrowing towards top end intended, otherwise some downshelving might be needed
- detection thresholds for group delay (direct sound) have to be undershoot, as well as in the audible band as a whole.
Last edited:
you misquote me, I did not say directivity and equal, I said pattern and similar.
I agree with all your comments above.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.