Talk to me about locating a mono sub...

[JRT]

2. I don't care how good someone's hearing is. The human ear can't localize tones below 80hz. The pinna isn't large enough and our ears are too close together to be able to detect phase differences. The way bass interacts with the room would make such phase detections impossible anyway.

3. So if you can localize your sub, you're probably not crossing it at 80hz or below.

4. If you can localize your sub and you ARE crossing it at 80hz or below, then your sub has a strong harmonic distortion component. You're not localizing 60hz; you're localizing the even order distortion at 120hz or 180hz.

5. Stereo subs are a MYTH. Why? See #2. If you are crossing over from satellites at some higher frequency like 120 or 150hz, okay, ya, I get the advantage of having two of them situated as near the satellites as possible. But that's a special case and please don't think it will add any imaging to the picture.

http://www.aes.org/e-lib/browse.cfm?elib=13041

"Results indicate that left-right discrimination occurs even at the lowest frequencies of human hearing – a finding consistent with other recent research."

I have not read that AES paper (circa 2005), just the abstract. And I am not one of the authors, so theirs is not a position I am taking up in point of argument. But the quoted snippet does seem to imply that their conclusions were significantly at odds with your assertions. Perhaps their work was flawed, and if so it may very well be the harmonic distortion that was audible (wouldn't rule that out without knowing more about the details of the experiments).


.


.

[Dirk]

I'd be very curious to know their test conditions, as well.

Anybody hanging out here with an AES membership?

[JRT]

I'd be very curious to know their test conditions, as well.

Anybody hanging out here with an AES membership?

If anybody bothers to look this one up, it would be good to look for foot notes to references associated with the phrase, "...a finding consistent with other recent research." It would be good to see that other work that backs up their findings.



.


.

[avaserfi]

The conference presentation in question has no real surprising findings with regard to low frequency localization. Essentially, what is found is that humans can localize low frequency information on a left-right basis. Also, There is some weak evidence found for front-back localization effects. Please note the ranges tested were from 25Hz-100Hz and it was found that there left-right localization was consistent in the full tested range, but as frequency decreased there was an increase in variance.

The authors speculate that such strong evidence for left-right localization can be attributed to binaural cues, specifically inter-aural time differences. This speculation seems to be based on the weak support for front-back localization.

This is one cited article that is claimed to have similar conclusions: http://www.aes.org/e-lib/browse.cfm?elib=12885

This is exactly why, for ideal music reproduction, dual subwoofers should be used and placed, ideally, within one quarter wavelength of Fc from the main speaker with one half wavelength being sufficient. Such placement will allow for proper integration such that each loudspeaker/subwoofer is able to act as a single point source.

[JRT]

The conference presentation in question has no real surprising findings with regard to low frequency localization. Essentially, what is found is that humans can localize low frequency information on a left-right basis. Also, There is some weak evidence found for front-back localization effects. Please note the ranges tested were from 25Hz-100Hz and it was found that there left-right localization was consistent in the full tested range, but as frequency decreased there was an increase in variance.

The authors speculate that such strong evidence for left-right localization can be attributed to binaural cues, specifically inter-aural time differences. This speculation seems to be based on the weak support for front-back localization.

This is one cited article that is claimed to have similar conclusions: http://www.aes.org/e-lib/browse.cfm?elib=12885

This is exactly why, for ideal music reproduction, dual subwoofers should be used and placed, ideally, within one quarter wavelength of Fc from the main speaker with one half wavelength being sufficient. Such placement will allow for proper integration such that each loudspeaker/subwoofer is able to act as a single point source.

Thanks for posting the info on the presentation and cited article. The question remains open as to whether or not some audible harmonic distortion products provided some cues for localization. Perhaps better subwoofers might have been more difficult to localize.

Compared to using a single subwoofer, your latter comments provide a soluton that better accomodates any localizable harmonic distortion components that might be generated within a subwoofer.

And your solution would also accomadate music that was well mastered on full range speakers. By the details in your comments I expect you aready well understand that mixing for 2.1 imposes a different EQ, but some others reading this may not. In your solution the two subs are each in close proximity to the satellites, but can be widely separated from each other. So radiation from those subs sums uncorrelated at the upper end of their range, and correlated at the lower end, as would be the case with a pair of full range speakers with similar separation. The same pair of subwoofers co-located midway between the satellites would sum correlated with each other over their full range, a 3dB difference in EQ at the upper end, affecting a broad range of frequencies; and a single sub at the same location would have the same EQ problem.

All that said, modal response will swamp most other considerations below the Schroeder frequency in many rooms.

[Dirk]

I'd also be curious to know what the room was for the test. Was the space acoustically large or small relative to the reproduced frequencies?

I suppose that's the same question as JRT's last sentence. Still curious.

[avaserfi]

From the article:

The test was conducted in an anechoic room – except that the anechoic criterion of this room is not satisfied below 250 Hz. The room dimensions (to the surface of the anechoic lining) are 3.6 m x 4.5 m, with a height of 2.5 m.

Reverberation time was measured as well, but I did not bother capturing an image of the graphs.

Harmonic distortions were attempted to be controlled for in the article as well via using moderate SPL levels (65dBA).

An image of the room from the article:


With regard to stereo subwoofers, of course, there will be room/subwoofer and subwoofer/subwoofer interactions. Proper placement, room treatment and if needed equalization can mitigate these issues.

[vasyachkin]

i see no reason whatsoever why it would even be difficult for a brain to localize a 25 hz source left to right.

a nerve can only fire something like 100 impulses per second ... and yet we can hear 20,000 hz - how do you explain THAT ?

in fact unless somebody screws your head to a support so you can't rotate it even 20 degrees i see no reason why a brain couldn't localize the position of a 25 hz source front-back as well.

based on frequency, and relative instantaneous spl in both ears the brain should be able to calculate the ANGLE from which the bass is coming. now as you turn the head only slightly this angle is either increasing or decreasing which immediately can tell the brain if the angle is in front or in the back.

what do you think harmonics are Dirk ? they aint nothing but the same fundamental wave only bent into a slightly more squarish or triangulish shape. do you really think the brain cares how pretty your sinewave looks for the purposes of telling where its coming from ?

pressure differential between the ears generated by the fundamental will be greater than that generated by harmonics anyway. the dipole nature of the differential will only have 6 decibels of attenuation ( relative to frequency of first harmonic ) while the harmonic itself will be something like 20 decibels down.

ultimately Dirk what it comes down to is you forgot the scientific method is. theory has to explain reality and if the two are in disagreement the theory is proven wrong. you seem to believe that theory can prove reality wrong.

further food for thought for you: two ears are not two separate microphones. the are coupled to each other acoustically through the head. in other words ears do not measure two pressures and then let the brain calculate a differential - they can measure this differential directly without the brain's involvement.

not only is there no need for a brain here, but yo don't even need two ears. one ear drum centered in the middle of your head would have measured precisely the pressure differential that you seem to believe two interconnected ears and a brain with a hundred billion neurons cannot detect.

face it Dirk - i am the best !

[Dirk]

a nerve can only fire something like 100 impulses per second ... and yet we can hear 20,000 hz - how do you explain THAT ?

The human ear is not a pulse-modulated device. The little protien "hairs" resonate with the frequency. That vibration is what's picked up by the nerves. I'm sure, like all resonant devices, they have a Q. If that Q creates a resonance longer than your proposed 1/100th of a second, why can't the frequency be picked up?

based on frequency, and relative instantaneous spl in both ears the brain should be able to calculate the ANGLE from which the bass is coming. now as you turn the head only slightly this angle is either increasing or decreasing which immediately can tell the brain if the angle is in front or in the back.

I will concede that in an anechoic chamber, free of heavy reflections, that may be. I'm still not convinced of how it plays out in a highly reverberant (acoustically small) room where everything is a pressure mode. I'm not sure how the precendence effect comes in to play here, either. I'm genuinely curious.

If I put my foot in my mouth, I'll willingly eat crow. But I'm not going to retract my entire swaggering statement when all that's needed is a qualifier.

what do you think harmonics are Dirk ? they aint nothing but the same fundamental wave only bent into a slightly more squarish or triangulish shape. do you really think the brain cares how pretty your sinewave looks for the purposes of telling where its coming from ?

Bent? Maybe it's a translation issue, but otherwise, uh, no. The sine wave of an even-order harmonic has twice as many (evenly spaced!) peaks as its fundamental. 60hz = 60 sine crests per second. 120hz = 120 sine crests per second.

pressure differential between the ears generated by the fundamental will be greater than that generated by harmonics anyway. the dipole nature of the differential will only have 6 decibels of attenuation ( relative to frequency of first harmonic ) while the harmonic itself will be something like 20 decibels down.

Armed with that logic, you're saying that if there's a 50hz tone playing at 70dB I can't hear a 100hz tone at 50dB?

ultimately Dirk what it comes down to is you forgot the scientific method is. theory has to explain reality and if the two are in disagreement the theory is proven wrong. you seem to believe that theory can prove reality wrong.

"Forgot scientific method", my hairy arse. No. Just the opposite. I am not yet convinved that theory HAS NOT proven reality wrong. The problem there is your definition of reality. If his test was done in an anechoic chamber, then yes, they have proven that the human ear CAN localize bass. But they have only proven it for the narrow confines of their test environment. An anechoic chamber is an artificial construct outside of the normal environment for which the human ear was designed (ie, a largely reverberant field). So the study does not tell me if the human ear can localize bass in a real room. That's "reality" and by that one small shred I maintain my ego!

further food for thought for you: two ears are not two separate microphones. the are coupled to each other acoustically through the head. in other words ears do not measure two pressures and then let the brain calculate a differential - they can measure this differential directly without the brain's involvement.

So how does that reinforce your case?

not only is there no need for a brain here, but yo don't even need two ears. one ear drum centered in the middle of your head would have measured precisely the pressure differential that you seem to believe two interconnected ears and a brain with a hundred billion neurons cannot detect.

Um, again, pretty sure that's not now sound works. That would only work if our brains knew ahead of time what signal was incoming. If the pressure wave biases that membrane more to the left than the right, that only means that the incoming signal has more crests than troughs. If then, after a short period of time, the signal inverts, and there are more troughs than peaks, does that mean that the sound will appear to be coming from the other side?

We localize sound specifically because we HAVE two ears. Binaural hearing. Stereo vision. Both sensory systems set up to be able to compare differing (yet corellated) inputs and from that triangulate a position in space.

face it Dirk - i am the best !

How could I EVAR argue that one?