Earl:
(..see Graaf's last post, he made the modest sarcastic comment but it was really to highlight the way he felt he was treated here, it really didn't have much to do with Markus's previous misunderstanding with Elias.)
..anyway,
I understand your viewpoint on directivity at low freq.s: that functionally it isn't practical. Nor would I argue with whats wrong with an open baffle midbass-bass operation (i.e. why it isn't practical), but my question is: what might be "right" with an open baffle woofer (beyond a very limited reduction in side-wall reflections at lower freq.s)?
I'd specifically like Elias to "chime-in" here on this topic - he has after all started it.
IMO, I've also discovered that YES - imaging can in fact get better with an open baffle operated midbass (vs. a more conventional monopole, modestly directive at higher freq.s or not). More specifically that some "images" become more apparent in depth, particularly lower freq. "images" like a drum kit. On most of my music (with drums), I'll find that they are "pushed" forward to some degree with a monopole midbass, whereas with a good open baffle they are not (i.e. they are set back behind performs in most instances).
Of course its one thing exclaim what is subjectivly correct, (at least what I and others I've "shown"), but I am very much interested in *why* this would be the case.
(..see Graaf's last post, he made the modest sarcastic comment but it was really to highlight the way he felt he was treated here, it really didn't have much to do with Markus's previous misunderstanding with Elias.)
..anyway,
I understand your viewpoint on directivity at low freq.s: that functionally it isn't practical. Nor would I argue with whats wrong with an open baffle midbass-bass operation (i.e. why it isn't practical), but my question is: what might be "right" with an open baffle woofer (beyond a very limited reduction in side-wall reflections at lower freq.s)?
I'd specifically like Elias to "chime-in" here on this topic - he has after all started it.
IMO, I've also discovered that YES - imaging can in fact get better with an open baffle operated midbass (vs. a more conventional monopole, modestly directive at higher freq.s or not). More specifically that some "images" become more apparent in depth, particularly lower freq. "images" like a drum kit. On most of my music (with drums), I'll find that they are "pushed" forward to some degree with a monopole midbass, whereas with a good open baffle they are not (i.e. they are set back behind performs in most instances).
Of course its one thing exclaim what is subjectivly correct, (at least what I and others I've "shown"), but I am very much interested in *why* this would be the case.
I wonder what are psychoacoustical differences between real sound source and phantom stereophonic sound source from spatial hearing perspective?
this seems to be crucial question from perspective of optimization of stereo sound reproduction
it seems that the two kinds of sound sources are defined by different sets of localization cues
in case of a real sound source there is complete set of coherent localization cues supporting each other
in case of phantom source there is incomplete set of localization cues which are partly different from natural localization cues and partly incoherent
e.g. in case of classic coincident/panpotted "X-Y stereo" there is:
1) Interaural Level Difference (ILD, also called Interaural Intensity Difference) which is unlike in real life - frequency independent and available across full spectrum
and there is
2) Interaural Phase Difference (also called ongoing Interaural Time Difference) which is not recorded but is a function of stereo loudspeakers spread and can possibly give cues incoherent with recorded ILD cues
in case of non coincident "A-B stereo" the situation is even more complicated and more likely to produce incoherent results because working localization cues are only of the ITD kind and "ongoing ITD" cues resulting from stereo loudspeakers spread can conflict with recorded "ongoing ITD" cues
so we are left with "onset ITD" cues
on the other hand loudspeakers as real physical sound sources are defined by complete set of coherent localization cues - ILD (at high frequencies), onset ITD (across the full spectrum, most important in reverberant space such as typical listening room), ongoing ITD (at low frequencies) and spectral cues (HRTF)
the situation is even more complicated because
1) typical listening room is a reverberant acoustical space
2) loudspeakers "send" recorded localization cues to both ears
ad 1)
under reverberant conditions spatial hearing gives special weighting to different localization cues
localization cue most potent under anechoic conditions - ongoing ITD - seems to be the least important in such situation because it can be ambiguous and misleading due to multiple reflections:
therefore localization cues reliable under reverberant conditions are most probably ILD, "onset ITD" and spectral cues (HRTF)
in case of "X-Y stereo" strong ILD cues available across the full spectrum seem to have to prevail over all others under reverberant conditions BUT ONLY in the highest frequencies (see point 2) below)
I doubt if in such a case the loudspeakers themselves can be said to be "perceived in stereo as early reflections of a sound whose direct version we missed" as Moulton put it because stereo ILDs are much higher than level differences between real reflections which occur in reverberant field
in case of "A-B stereo" under reverberant conditions the working mechanism of phantom sound source perception seems to be the recorded "onset ITD"
also in this case I doubt if the loudspeakers themselves can be said to be "perceived in stereo as early reflections of a sound whose direct version we missed" as Moulton put it because stereo "onset ITDs" are much smaller than time differences between real reflections which occur in reverberant field
ad 2)
in case of "X-Y stereo" loudspeaker crosstalk leads to invalidation of most part of recorded ILD localization cues
why?
limited angle of stereo triangle cause that even for high frequencies the acoustic shadow cast by the head is ineffective
therefore even at those high frequencies both ears hear exactly the same sound level which is sum of left and right channels levels
in result ILD localization/phantom image formation mechanism is effective only at the highest frequencies
the above is how I understand(/don't understand) it
please correct what is wrong in the above speculations
so it seems to me that the listener is presented simultaneously with two real and strongly spatially defined sound sources ie. the two loudspeakers which cannot be simply treated as "early reflections" (as Moulton suggests) and with one or more weakly spatially defined phantom sources
how are those real and phantom sound sources related to each other?
how are they able to produce any reliable auditory scene?
what is wrong? (if anything?)
what has to be done? (if anything?)
what can be done? (if anything?)
any* input welcome!
best regards!
graaf
* provided it is ad rem , please 🙂
yes, I already know that I am disrespectful troll - no point in reminding me this undisputable fact again 🙂
please accept as a kind of disability that I am apparently unable to show adequate respect to most respectable DIY audio members
please treat me as a kind of disabled person in this regard
please forgive and accept as an excuse that as I am genuinely interested in audio
this seems to be crucial question from perspective of optimization of stereo sound reproduction
it seems that the two kinds of sound sources are defined by different sets of localization cues
in case of a real sound source there is complete set of coherent localization cues supporting each other
in case of phantom source there is incomplete set of localization cues which are partly different from natural localization cues and partly incoherent
e.g. in case of classic coincident/panpotted "X-Y stereo" there is:
1) Interaural Level Difference (ILD, also called Interaural Intensity Difference) which is unlike in real life - frequency independent and available across full spectrum
and there is
2) Interaural Phase Difference (also called ongoing Interaural Time Difference) which is not recorded but is a function of stereo loudspeakers spread and can possibly give cues incoherent with recorded ILD cues
in case of non coincident "A-B stereo" the situation is even more complicated and more likely to produce incoherent results because working localization cues are only of the ITD kind and "ongoing ITD" cues resulting from stereo loudspeakers spread can conflict with recorded "ongoing ITD" cues
so we are left with "onset ITD" cues
on the other hand loudspeakers as real physical sound sources are defined by complete set of coherent localization cues - ILD (at high frequencies), onset ITD (across the full spectrum, most important in reverberant space such as typical listening room), ongoing ITD (at low frequencies) and spectral cues (HRTF)
the situation is even more complicated because
1) typical listening room is a reverberant acoustical space
2) loudspeakers "send" recorded localization cues to both ears
ad 1)
under reverberant conditions spatial hearing gives special weighting to different localization cues
localization cue most potent under anechoic conditions - ongoing ITD - seems to be the least important in such situation because it can be ambiguous and misleading due to multiple reflections:
therefore localization cues reliable under reverberant conditions are most probably ILD, "onset ITD" and spectral cues (HRTF)
in case of "X-Y stereo" strong ILD cues available across the full spectrum seem to have to prevail over all others under reverberant conditions BUT ONLY in the highest frequencies (see point 2) below)
I doubt if in such a case the loudspeakers themselves can be said to be "perceived in stereo as early reflections of a sound whose direct version we missed" as Moulton put it because stereo ILDs are much higher than level differences between real reflections which occur in reverberant field
in case of "A-B stereo" under reverberant conditions the working mechanism of phantom sound source perception seems to be the recorded "onset ITD"
also in this case I doubt if the loudspeakers themselves can be said to be "perceived in stereo as early reflections of a sound whose direct version we missed" as Moulton put it because stereo "onset ITDs" are much smaller than time differences between real reflections which occur in reverberant field
ad 2)
in case of "X-Y stereo" loudspeaker crosstalk leads to invalidation of most part of recorded ILD localization cues
why?
limited angle of stereo triangle cause that even for high frequencies the acoustic shadow cast by the head is ineffective
therefore even at those high frequencies both ears hear exactly the same sound level which is sum of left and right channels levels
in result ILD localization/phantom image formation mechanism is effective only at the highest frequencies
the above is how I understand(/don't understand) it
please correct what is wrong in the above speculations
so it seems to me that the listener is presented simultaneously with two real and strongly spatially defined sound sources ie. the two loudspeakers which cannot be simply treated as "early reflections" (as Moulton suggests) and with one or more weakly spatially defined phantom sources
how are those real and phantom sound sources related to each other?
how are they able to produce any reliable auditory scene?
what is wrong? (if anything?)
what has to be done? (if anything?)
what can be done? (if anything?)
any* input welcome!
best regards!
graaf
* provided it is ad rem , please 🙂
yes, I already know that I am disrespectful troll - no point in reminding me this undisputable fact again 🙂
please accept as a kind of disability that I am apparently unable to show adequate respect to most respectable DIY audio members
please treat me as a kind of disabled person in this regard
please forgive and accept as an excuse that as I am genuinely interested in audio
Ups, there's an error in posting #258: "Interaural time and phase differences form very stable phantom images..." should be "Interaural time and level differences - especially the latter - form very stable phantom images...".
Best, Markus
Best, Markus
ScottG said:
Its "not practical" or necessary. You left off this last word.
As far as your (and others) personal perceptions about the benefits of OB designs, I have not had the same experince with such designs. Your experinces may not be universal and as such may not have a well defined psychoacoustic rational. Much is said about OB and its benefits, but much of it has also been debunked. Such as the claim that dipoles excite fewer modes for a smoother LF response. This has been shown by several people to be completely false. OB has still not garnered a significant market share, which one might expect if their benefits were universally perceived. I'm not satying that your perceptions are right or wrong, only that there is not a substantial body of evidence to support them. The theoretical claims are weak at best and the subjective ones are never done under scientific conditions.
gedlee said:
This is my dipole configuration:
An externally hosted image should be here but it was not working when we last tested it.
And this happens, when I toe in the H frames:
An externally hosted image should be here but it was not working when we last tested it.
Black is with the H frames perpendicular to the front wall, green with the frames facing one another, red with 60° toe-in.
In the right diagram you can see how the +30 Hz resonance is "switched off" by rotating the dipole axis. In the left diagram the mic is in the null of the 30 Hz resonance, but you see better, how the +60 Hz resonance is switched off too.
I don´t claim that this is the solution to room mode problems. We want a better distribution of room modes more than non-existant room modes. But it clearly shows, that directivity of dipoles matters down to low frequencies. I´m very shure that such a dipole as a third sub in your proposed system would be of great value.
Rudolf said:
The fact is that this test really doen't indicate or prove anything. Only that your woofer at that location changes when you rotate it. Thats to be expected right? Monopoles are not sensitive to rotation, but they are to location. So moving a monopole to a different location would also change it. That too is to be expected. But none of this shows how dipoles and monopoles will work IN GENERAL. To do that you need to look at the problem in a little broader sense - diferent room sizes and shpes, etc.. When that is done one finds that monopoles and dipoles both have about the same "variation" in thier response, but the dipole falls at LF and the monopole doesn't.
I have also said on several occasions that it doesn't matter what the LF source configuration is, monopole, dipole, cardiod, all work in the multiple sub configuration just about the same, but they require substantially diferent amounts of power to achieve the same result.
gedlee said:
Well, it only proves that a dipole offers one more degree of freedom to interact with the room than the monopole. It is certainly up to discussion whether that is of any practical help or not.
But you already have stated that <Its "not practical" or necessary>.
No, you are mistating me. I said the LF directivity is not practical or necessary. The use of dipoles as subwofers as you have shown is a seperate issue.
As to the added degree of freedom, this could be useful in some circumstances for particular problems, but its hard to say that "in general" this would be an advantage. The extra degrees of freedom offered by multiple subs are more effective IMO, however the ability to rotate the sub to change the modal interactions is, as you say, an aditional dgeree of freedom. But at these LFs the poor radiation efficiency of the dipole is clearly a detriment. It is unclear if the added degree of freedom would outweigh the substantial added power and EQ that dipoles require.
The added degree of freedom is an interesting thought however.
As to the added degree of freedom, this could be useful in some circumstances for particular problems, but its hard to say that "in general" this would be an advantage. The extra degrees of freedom offered by multiple subs are more effective IMO, however the ability to rotate the sub to change the modal interactions is, as you say, an aditional dgeree of freedom. But at these LFs the poor radiation efficiency of the dipole is clearly a detriment. It is unclear if the added degree of freedom would outweigh the substantial added power and EQ that dipoles require.
The added degree of freedom is an interesting thought however.
Rudolf, sure it's a nice idea to have a dipole sub that can be freely rotated - why is yours fix mounted? It can help in cancelling big peaks at the listening position that otherwise would lead to the typical "booming" bass. But as your diagrams have shown the price is to have less bass instead of a linear bass. Only active (multisub) or passive absorption can help in having a linear bass response.
Best, Markus
Best, Markus
You are right. LF directivity discussed for the Nathan is not automatically the same issue as the directivity of a dipole. 🙁gedlee said:
Now it seems we are not that far apart. I certainly did not mean to substitude ALL monopole subs with dipoles - just one out of three perhaps. And certainly I don´t see that dipole as the solution to all remaining problems, but just as an alternative to try.
Did you realise by the way that a dipole at floor height should (by theory) have much in common with a monopole half up the wall?
Rudolf said:
In all honesty I don't see your point here, but I do realize that within my theory the actual type of source is almost irrelavent. I have said many times that the type of source is secondary to the number and location, so I have argued against dipoles, I have only argued FOR using more of them.
Markus,
that baffle above the H frame isn´t fixed at all. And yes - I already have tried different angles between OB and H frame. But the H is crossed at 250 Hz to the midrange, so I can´t separate them totally.
Regarding peak cancelling: You are perfectly right. I´m on the verge of installing a first monopole sub (spare SLS10 😀) in my system. I´m not that fixed to dipoles as my signature might suggest - just coming from the opposite direction. 😉
that baffle above the H frame isn´t fixed at all. And yes - I already have tried different angles between OB and H frame. But the H is crossed at 250 Hz to the midrange, so I can´t separate them totally.
Regarding peak cancelling: You are perfectly right. I´m on the verge of installing a first monopole sub (spare SLS10 😀) in my system. I´m not that fixed to dipoles as my signature might suggest - just coming from the opposite direction. 😉
Meet you in the middle - but don't forget to bring the beer with you as Earl was so nice to give away mine to others 🙂
Speaking of active (multisub) or passive absorption: What happens when the active component (the subs) stops playing? I have the suspicion that the rooms modal field still plays a big role. Here's the waterfall of a SBA (single bass array):
Two additional subs in the back (DBA):
The frequency response in the DBA setup is linear but you still can see the mode "ringing" at around 35 Hz.
Best, Markus
An externally hosted image should be here but it was not working when we last tested it.
Two additional subs in the back (DBA):
An externally hosted image should be here but it was not working when we last tested it.
The frequency response in the DBA setup is linear but you still can see the mode "ringing" at around 35 Hz.
Best, Markus
Markus
A good point, and one that I have addressed in the past, although you may not have recognized it as such.
In a large room the LFs ring on for a very long time yielding a very natural recognition of the LF signal. This is a major component of why lows sound so good in big rooms.
In a small room discrete modes ring on yielding a pronounced boomy effect. The solution is very high LF damping (as I have always suggested) and a LF reverb system (also suggested) to add back a modal free "ring-on" just like we find in large rooms. The Reverb is not resonant like the modal decay is and as such it won't sound boomy. The end effect is a very natural LF sound field that parallels the LFs in a large room - which always sound good.
A good point, and one that I have addressed in the past, although you may not have recognized it as such.
In a large room the LFs ring on for a very long time yielding a very natural recognition of the LF signal. This is a major component of why lows sound so good in big rooms.
In a small room discrete modes ring on yielding a pronounced boomy effect. The solution is very high LF damping (as I have always suggested) and a LF reverb system (also suggested) to add back a modal free "ring-on" just like we find in large rooms. The Reverb is not resonant like the modal decay is and as such it won't sound boomy. The end effect is a very natural LF sound field that parallels the LFs in a large room - which always sound good.
How to implement a LF reverb system? And can there ever be such a good LF absorption that you need to add back LF?
Best, Markus
Best, Markus
Markus
Beringer makes system that looks like it would work. Don't remember the model number.
Getting LF absorption is a difficult task, probably the most difficult thing to do. Short of doing it in the actual rooms construction I don't know how it could be done. Sound traps just don't work. They absorb a lot of highs, but little lows. I'm sorry that this is not a "convenient" solution, but it is one that I can attest to works.
It would be possible to create enough LF absorption that added gain was required at the LF. The point is that with my method of doing the LFs this comes naturally and no thought needs to actually be given to whether or not this is the case.
Active LF absorption has its attractions to me for those cases where it is impossible to get LF absoption mechanically. However, the systems that I have seen are too simplistic and I suspect that a lot more is required. In a very real sense multiple subs act like active absorbers. If a sub is turned on and the delay and or phase is such that the sound level actually decreases then this IS an active absorber. So in principle multiple subs, when set correctly act as both sources and sinks. This is a key aspect of the need to distribute them and to use several.
Several years ago I did a complex "in-house" study of what could be done at LFs in a small room. Its far too complex to get into in detail here, but the need for reverberation came out of this study. I looked at several sources in a small room and what kind of "filter" in each signal path would yield the smoothest response. Quite interestingly the answer came out to be a "decorrelation" filter, as its called in the literature. But a careful study of a decorrelation filters impulse response will show that it is a myriad of random impulses that decay in time. This is exactly reverberation. Thus a different set of reverberation filters in each sub channel can be shown to be the optimal set of filters for the smoothest LF response.
The fact that this is in essence creating the same sound field in a small room that is found naturally in a large room is not only elegant but downright profound.
Beringer makes system that looks like it would work. Don't remember the model number.
Getting LF absorption is a difficult task, probably the most difficult thing to do. Short of doing it in the actual rooms construction I don't know how it could be done. Sound traps just don't work. They absorb a lot of highs, but little lows. I'm sorry that this is not a "convenient" solution, but it is one that I can attest to works.
It would be possible to create enough LF absorption that added gain was required at the LF. The point is that with my method of doing the LFs this comes naturally and no thought needs to actually be given to whether or not this is the case.
Active LF absorption has its attractions to me for those cases where it is impossible to get LF absoption mechanically. However, the systems that I have seen are too simplistic and I suspect that a lot more is required. In a very real sense multiple subs act like active absorbers. If a sub is turned on and the delay and or phase is such that the sound level actually decreases then this IS an active absorber. So in principle multiple subs, when set correctly act as both sources and sinks. This is a key aspect of the need to distribute them and to use several.
Several years ago I did a complex "in-house" study of what could be done at LFs in a small room. Its far too complex to get into in detail here, but the need for reverberation came out of this study. I looked at several sources in a small room and what kind of "filter" in each signal path would yield the smoothest response. Quite interestingly the answer came out to be a "decorrelation" filter, as its called in the literature. But a careful study of a decorrelation filters impulse response will show that it is a myriad of random impulses that decay in time. This is exactly reverberation. Thus a different set of reverberation filters in each sub channel can be shown to be the optimal set of filters for the smoothest LF response.
The fact that this is in essence creating the same sound field in a small room that is found naturally in a large room is not only elegant but downright profound.
gedlee said:
No, I specifically did NOT leave out "necessary".😉
Of course I'm ALSO not claiming that an open baffle midbass is necessary either, the fact is that I don't know *why* it sounds better in the way it does - that's kind'a the point to this inquiry.
Yes, I'm well aware that a dipole's excitation of room modes at lower freq.s is a relatively baseless claim for any superiority. They of course excite modes to a similar degree (amplitude), just at a slightly differently (freq.) (..for a given location). (JohnK has done some excellent work here on this subject.) For some reason though this claim is continually used, despite being mostly erroneous. What I'm interested in are OTHER reason why an open baffle midbass might be better.😉
It's easy to dismiss it because it's subjective sentiment expressed by a few on a forum, and far easier to because you haven't noticed such a difference. That however does NOT mean that there isn't something "going on" that is more than perceptual fiction. That an open baffle midbass is or is not a commercially present design has almost no bearing to any audible appeal (or not) (..i.e.e a REALLY weak argument).
Why not try expending some mental effort to consider reasons why an open baffle config. at lower freq.s *might* have merit? (Of course - THEN you can always argue against those reasons.) 🙂
- Status
- Not open for further replies.