Without comment. Impulse response at my listning position.
Craw monopole- front driver only:
Monopole both sources connected:
CRAW, dipole mode:
CRAW, cardioid mode:
Craw monopole- front driver only:
An externally hosted image should be here but it was not working when we last tested it.
Monopole both sources connected:
An externally hosted image should be here but it was not working when we last tested it.
CRAW, dipole mode:
An externally hosted image should be here but it was not working when we last tested it.
CRAW, cardioid mode:
An externally hosted image should be here but it was not working when we last tested it.
It's all explained at Comparison of different near field subwoofer configurations
The spectrogram is supposed to show in-room low frequency modal ringing. As always the window time is a tradeoff between frequency and time resolution. I doubt 300ms is too long for what I wanted to show.
I can save data fro ATRT as I do not have a registered copy. My W95 PC is not working so I can't use the IMP. So pictures are the best I can do at the moment. The thing the struck me was that with the mono, connecting the rear source, which is 18" behind the front source, seems the make the impulse ring on much longer, not to mention the apparent dead spot right after the initial spike which is then followed by a broad up spike.
My software can use up to six room locations and I suggest at least three, but more is better. The sources cannot be moved about once the data has been taken, but the processing can change. In the end it creates a situation where if you turn off one source, you will in general see both an increase and a decrease in different parts of the spectrum. This means that this sub is both sourcing and sinking. In general I find that when a good setup is obtained that all sources act as sources and sinks in different frequency ranges. To me this has a certain appeal to it.
Early next year I am on hook to do a presentation on using multiple subs to the lcal audio group. I will be doing and documenting the whole procedure in my room and I can probably post all this once I have it. I hope to do it over the holidays because between now and then I am swamped.
Depends on source spacing. I was thinking of two woofers well spaced giving double the power (+3dB) but double the pressure for a listener equal distance to both. No different than dual mids in an array giving +6 on axis but +3 for random phase summing (or a 3dB increase in d.i).
I realize woofers tend to be close relative to the wavelengths but if they are well spaced you may get a benefit towards the top of the range.
As to structural borne bass, that is hard to calculate. My comments were for air borne using the standard room acoustics approach. Structural borne is easy to deal with in terms of compliant feet. Do we know that it is ever a bigger factor than acoustical coupling from room to room?
I'd have to think about dipole vs. monopole with regard to vibration. I would guess the other way, that a dipole, that has to accelerate the woofer more for a given far field SPL, would tend to cause more adjacent vibration, but its not good to generalize too far.
Certainly, in a free field you are correct about nil difference between ears at bass frequencies (both for amplitude or phase). I believe Griesinger is strictly talking about in-room cases with differential between ears coming from proximity to standing wave nulls.
Regards,
David S.
For home situation I don't have much experience, but for cars I have a lot and in cars structural borne is a big deal - on par with airborne in some situations. In buildings I know that it is also a factor. That leads me to believe that it should not be overlooked. But yes, I completely agree that putting the sub on resiliant feet is a big step in the right direction and I do this for all my subs, but I don't think that it is the norm.
But isn't this what I said as well? That the dipole would have the greater mechanical force to its foundation.
Example of structure borne problem: I have an air-to-air heat exchanger on my HVAC system. When I rigidly mounted it to the house structure it made lots of noise (just two fans which are supposed to be balanced so where is the vibration coming from?). Then I suspended the entire unit from nylon cord - my first cut solution to all problems like this! - and the noise simply disappeared. A large sub cannot be much different than this system in terms of its size, weight and vibration force.
I still have a lot of problems with "LF localization". I keep going back to Blauert where all of his data stops at 250-500 Hz and is dropping below that. Obviously he never even consider LFs to have any localization and yet its talked about here as if its some kind of "known fact". I just don't see that at all. Even if there is "more LF localization than has been assumed in the past", it was assumed to be zero, so even "more" is still pretty insignificant. That it cannot be even comparable to frequencies above 500 Hz is clearly the case, but I have not seen any data that shows what it actually is in a small room where all sounds are basically steady state, all sources are basically close and highly correlated. I'm just not buying "LF localization" as a factor. It is proably artifacts > 250 Hz that are being detected.
Last edited:
Indeed, I read it backwards.
I don't think that Griesinger is talking about localization as we would think about it. In fact he is looking for the opposite. He wants rapid variation in the presentation to the ears to give a sense of envelopment.
This is a bit as we got into some pages back. Everyone else is pursuing some version of accuracy while he is actually going for less accuracy but a more pleasant since of involvement. His concert hall papers are following a similar vein arguing that engagement should be the primary design objective:
http://www.davidgriesinger.com/direct_sound.ppt
http://www.davidgriesinger.com/
Unconventional but worth thinking about.
David S.
I don't think that Griesinger is talking about localization as we would think about it. In fact he is looking for the opposite. He wants rapid variation in the presentation to the ears to give a sense of envelopment.
This is a bit as we got into some pages back. Everyone else is pursuing some version of accuracy while he is actually going for less accuracy but a more pleasant since of involvement. His concert hall papers are following a similar vein arguing that engagement should be the primary design objective:
http://www.davidgriesinger.com/direct_sound.ppt
http://www.davidgriesinger.com/
Unconventional but worth thinking about.
David S.
Indeed.
You can read that paper several times, and while it "reads" pretty well - it's difficult to truly understand, often presenting a slightly different perspective on each reading.
He needs a better editor (..or you know, one at all).
I would propose that in a small room, much as this completely disagrees with my approach above Fs, I am not sure that direct accuracy is warranted either. That is because a small room deviates so much from any room that one would ever listen to music in. So a direct one-to-one accuracy would not necessarily be what I would be looking for at these frequencies. It's at LFs that we need to "hide the room" and "trick our senses" because otherwise it just does not sound right. Again, thats because it is not right, its a small room not a large one. I use heavy EQ < Fs, and none above, I suggest heavy damping < Fs and very low amounts above. Its simply a completely different situation and has to be dealt with completely different. This is why I am not convinced that the shortest impulse response is desirable at LFs. What's the justification? More reverb at LFs would yield a greater IACC than a shorter impulse because reverberation tends towards uncorrelated sound.
I glanced through the Greisinger paper and found it quite compelling. He is saying much of the same things that I have been saying, that localization is the key to envelopment, that this is dominated at frequencies > 700 Hz, etc. etc. I never did read anywhere where these ideas had any implications for LF sound, but I didn't read it all. It is hard to read because it is not a paper, but notes from a presentation. He also seemed to be refering more to concert halls than small room reproduction systems. But what I did read I did not disagree with at all. It was completely parallel to what I have been saying and doing and I found his term "envelopment" to encompass precisely what I am looking for in a sound system. I might start using this term!! I like it! "Waveguides create Envelopment"! Cool!
Last edited:
Don't be so quick to dismiss gradients between both ears at bass frequencies. At 20Hz ? No. At 100Hz ? Yes definitely, I've seen the effect as low as 60-70Hz in some rooms.
Although wavelengths are long, if there are strong lateral standing waves present notches at certain frequencies due to destructive interference can be very deep and very localised spatially.
In one of my previous rooms there was a spot almost exactly at the listening position where 65Hz was almost perfectly notched out, at least for one ear at a time. Moving my head sideways would put one ear or the other directly in the notch, allowing the opposite ear to hear something, a very disconcerting effect much like listening to bass in one ear with headphones, with a phase reversal between the ears as the centre of the head crosses the cancellation point.
When phase between the direct path and reflection are almost exactly 180 degrees and the levels are very close to equal the change in amplitude with small phase shifts (physical displacement) is very rapid, many orders of magnitude more rapid than if they were summing in phase. It's easy to do the calculations based on summing sine waves at different phase angles to see this.
With normal in phase mono bass this effect will generally only happen if there are serious standing wave problems in the room, so bass quality overall is going to be poor anyway... but there is some music that puts the bass either in only one channel or in deliberate reverse phase between left and right (often with some levels mismatch on purpose to prevent total cancellation) and in these circumstances a bass gradient can happen between the ears especially if you're near equidistant from the speakers.
Again, the effect is quite disconcerting and I don't like it, but it does have an interesting effect on localisation.
Now is this effect true "localisation" or is it just an artefact of standing waves and destructive interference fooling the ears by causing a localised imbalance of bass ? The latter I think, as the exact perceived effect is highly dependant on the room standing wave structure, the precise listener position and so on, so while an interesting gimmick its not a reliable or reproducible way of introducing a sense of localisation at low frequencies. Certainly it is not reproducing any localisation captured during the recording process, and is therefore just a reproduction artefact.
A good multi-sub setup is going to almost completely eliminate any type of low frequency localisation by minimising standing wave issues and therefore steep gradients, and if its a mono bass setup it will eliminate reverse phase bass effects as well. (I'm not convinced however that a multi-sub setup should necessarily be configured for mono bass reproduction)
Last edited:
Simon,
that's basically what I've been saying in #518.
Especially when using close-field woofers/subs where 1/r level differences are still effective**) it is possible to get almost headphone-like L/R-seperation down low, with the appropriate signals coded in the source material.
**) and dipoles are best for this because there are two 1/r mechanisms
that's basically what I've been saying in #518.
Especially when using close-field woofers/subs where 1/r level differences are still effective**) it is possible to get almost headphone-like L/R-seperation down low, with the appropriate signals coded in the source material.
**) and dipoles are best for this because there are two 1/r mechanisms
Last edited:
Sorry, I've only been following the thread with one eye, so I missed post #518.
There's no question that imbalance in bass between left and right ears will cause a sense of left-right localisation (one need only try it with headphones to prove that) the question has only ever been whether a sufficient gradient between the ears can occur without headphones so the effect can manifest in non-headphone listening.
In free space with a single source - no. In rooms with complex standing wave patterns and boundary cancellation effects, yes, at certain frequencies and in certain circumstances it can occur, but usually not in any predictable way that could be exploited by a recording engineer.
I had the same debate with a certain well known member of rec.audio.tech about 10 years ago, he also didn't believe that a sufficient gradient was possible between the ears at bass frequencies even when listening near points of cancellation. (Simple calculations and empirical testing prove otherwise however)
There's no question that imbalance in bass between left and right ears will cause a sense of left-right localisation (one need only try it with headphones to prove that) the question has only ever been whether a sufficient gradient between the ears can occur without headphones so the effect can manifest in non-headphone listening.
In free space with a single source - no. In rooms with complex standing wave patterns and boundary cancellation effects, yes, at certain frequencies and in certain circumstances it can occur, but usually not in any predictable way that could be exploited by a recording engineer.
I had the same debate with a certain well known member of rec.audio.tech about 10 years ago, he also didn't believe that a sufficient gradient was possible between the ears at bass frequencies even when listening near points of cancellation. (Simple calculations and empirical testing prove otherwise however)
Last edited:
Thanks to Markus and John for their good work and posting the graphs
These graphs provides ample objective evidence to the subjective perceptions of many favoring dipoles and the comments I had posted in #458 :
"Even with the same diameter drivers, equalised to same LF extension, listened at moderate SPL, to even nearfield (2-3 Meters) there is that slight overhang in sealed units that makes it sound like a "speaker" which is not apparent in a dipole(H or Uframe) making it sound that last bit different ... or "natural'."
It is great to know that 3 independent experimenters (including Elias) using 3 different measuring methods have shown readings which show that a dipole as opposed to a monopole, has a better transient response and smoother frequency response both in nearfield and farfield listening.
It might be that multiple equalised monopoles as proposed by Dr. Geddes is the best "cost no object" way to get good bass but for an average room for stereo speakers, dipoles are as good if not better and I would agree with Markus in post #539
"I believe it's THE solution for everybody that wants really good bass and doesn't have a dedicated listening room with proper acoustic treatments. It's also limited to a few seats within one row but is stereo/multichannel anything more than a one seat solution?"
These graphs provides ample objective evidence to the subjective perceptions of many favoring dipoles and the comments I had posted in #458 :
"Even with the same diameter drivers, equalised to same LF extension, listened at moderate SPL, to even nearfield (2-3 Meters) there is that slight overhang in sealed units that makes it sound like a "speaker" which is not apparent in a dipole(H or Uframe) making it sound that last bit different ... or "natural'."
It is great to know that 3 independent experimenters (including Elias) using 3 different measuring methods have shown readings which show that a dipole as opposed to a monopole, has a better transient response and smoother frequency response both in nearfield and farfield listening.
It might be that multiple equalised monopoles as proposed by Dr. Geddes is the best "cost no object" way to get good bass but for an average room for stereo speakers, dipoles are as good if not better and I would agree with Markus in post #539
"I believe it's THE solution for everybody that wants really good bass and doesn't have a dedicated listening room with proper acoustic treatments. It's also limited to a few seats within one row but is stereo/multichannel anything more than a one seat solution?"
2-3m is NOT near field. Near field subs have to be placed within about 2ft (60cm).
Ample? Objective?
I think that these measurements are just some building blocks, but by no means a "house of wisdom". If pictures look to ones eyes like what he hears with his ears, we are talking about religion, but not about science. We need to know the exact physical reasons, why and from what those perceptional differences arise. Before that is achieved we are just guessing imho.
Rudolf
#551
This is way off topic; both monopole and dipole speakers are capable of setting up LF standing waves in appropriate sized rooms.
But perhaps some clarification:
"....gradients between both ears...." and "minimizing standing wave issues and therefore steep gradients....." is confusing language when describing ears as straddling a standing wave nodal point. Particle velocity is zero, and smoothly increases to maximum at antinode 1/4 wavelength away. For 65Hz this is roughly 132cm.
Hair cells appear to have a zero crossing function (and as I recall it is for negative going pressure at eardrum), and neighboring hair cells thus tend to phase lock. This behavior is exhibited below about 600Hz.
With ears straddling the node, particle velocities are 180 degrees out of phase, and so are hair cell firings of left and right ears.
With one ear effectively placed at node, it receives no signal while the other ear does.
In both cases a localization cue is generated.
For typical stereo speaker placement in typical room, generation of the requisite lateral standing wave formation seems highly unlikely.
#555:
What are you missing from my post #444? Both frequency response and transient responses may be equalized to be the same for monopole, dipole, and cardioid radiators.
If "there is that slight overhang in sealed units" it means they haven't been correctly equalized.
Regards,
Andrew
This is way off topic; both monopole and dipole speakers are capable of setting up LF standing waves in appropriate sized rooms.
But perhaps some clarification:
"....gradients between both ears...." and "minimizing standing wave issues and therefore steep gradients....." is confusing language when describing ears as straddling a standing wave nodal point. Particle velocity is zero, and smoothly increases to maximum at antinode 1/4 wavelength away. For 65Hz this is roughly 132cm.
Hair cells appear to have a zero crossing function (and as I recall it is for negative going pressure at eardrum), and neighboring hair cells thus tend to phase lock. This behavior is exhibited below about 600Hz.
With ears straddling the node, particle velocities are 180 degrees out of phase, and so are hair cell firings of left and right ears.
With one ear effectively placed at node, it receives no signal while the other ear does.
In both cases a localization cue is generated.
For typical stereo speaker placement in typical room, generation of the requisite lateral standing wave formation seems highly unlikely.
#555:
What are you missing from my post #444? Both frequency response and transient responses may be equalized to be the same for monopole, dipole, and cardioid radiators.
If "there is that slight overhang in sealed units" it means they haven't been correctly equalized.
Regards,
Andrew
BarleyWater
You make very good points, unfortunately the tight coupling between the frequency response and the impulse response is not always appreciated by many.
I had some doubts about your #444 post at first, but then after I figured out what you had actually done, it became clear. I think that what got missed by many (even myself at first) is that you EQ'd for a linear phase band-pass system and not a minimum phase system. You could have done either correct? Or better said, either is possible. Had you done the minimum phase we would have seen the kinds of impulse responses that we are used to seeing.
You are completly correct about the zero crossing firings below 600 Hz, which is precisely why the ear resolution falls below this frequency - fewer and fewer firings occur in a given time resulting is less and less information to the brain for it to process. This is also why we see our hearing sensitivity falls below 600 Hz.
I completely agree that "For typical stereo speaker placement in typical room, generation of the requisite lateral standing wave formation seems highly unlikely."
You make very good points, unfortunately the tight coupling between the frequency response and the impulse response is not always appreciated by many.
I had some doubts about your #444 post at first, but then after I figured out what you had actually done, it became clear. I think that what got missed by many (even myself at first) is that you EQ'd for a linear phase band-pass system and not a minimum phase system. You could have done either correct? Or better said, either is possible. Had you done the minimum phase we would have seen the kinds of impulse responses that we are used to seeing.
You are completly correct about the zero crossing firings below 600 Hz, which is precisely why the ear resolution falls below this frequency - fewer and fewer firings occur in a given time resulting is less and less information to the brain for it to process. This is also why we see our hearing sensitivity falls below 600 Hz.
I completely agree that "For typical stereo speaker placement in typical room, generation of the requisite lateral standing wave formation seems highly unlikely."
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.