Your experience (of listening the bass coming from your speaker) is not quite rigorous because you know where the speaker is located. Your subconscious is suggesting that the sound is coming from that direction, and you might believe that you can actually hear that.
Well, actually I closed my eyes and kept spinning myself around to lose all sense of direction (I still hope nobody saw me that day). It was by no means a rigorous scientific experiment, but it was good enough for myself to know for certain that under those circumstances the detect-ability of the sub's relative direction was real.
Our posts crossed each other and i was actually writing in response to the OP, sorry about that. The point is that we already have more info on the original experiment, as we know from Pano what signal he used and what sound setup (low HD). Can i suggest you elaborate on that as well?
Localization
of low frequencies by ear iws certainly possible, and it has been documented in scientific studies in Denmark many years ago by Henrik Moller, an exoert in low frequency noise.
A number of students were participating in the study, that consisted of a number of subwoofers (I believe as much as 12, but I am not sure) placed in a semi-arc in an anechoic chamber.
It was found (again, from my recollection) that the subjects were albe to pinpoint the low-frequency sound-source within a few degrees.
However, when a loudspeaker is placed in a normal (living-) room the room itself has a very large effect on the emitted sound: reflections from the floor, wall and ceiling are dominating at low frequencies, and this make loalization much more difficult.
But in high-quality set-up's it's certainly possible.
Steen
of low frequencies by ear iws certainly possible, and it has been documented in scientific studies in Denmark many years ago by Henrik Moller, an exoert in low frequency noise.
A number of students were participating in the study, that consisted of a number of subwoofers (I believe as much as 12, but I am not sure) placed in a semi-arc in an anechoic chamber.
It was found (again, from my recollection) that the subjects were albe to pinpoint the low-frequency sound-source within a few degrees.
However, when a loudspeaker is placed in a normal (living-) room the room itself has a very large effect on the emitted sound: reflections from the floor, wall and ceiling are dominating at low frequencies, and this make loalization much more difficult.
But in high-quality set-up's it's certainly possible.
Steen
Yes, near standing wave node is where the 'localization' in a small room comes from (insert caveats about high order distorsion, noises etc.). It doesn't depend on source placement (another caveat: placement dictates how a mode is excited). As dantheman observed, smoothing the modes with multiple subs gets rid of this artifact.
Actually in a large room, where the modes are so dense at low freq., low freq. 'localization' is very real, but manifests in the form of envelopment.
How great would be to have this same phenomenon in a small room? But due to sparse modes IACC is too high for this to happen. How to excite the soundfied (in a small room) in the low frequencies to emulate this effect? James Johnston thinks he has an answer, involving different recording techniques as the norm. Same thinks Griesinger. Others remain unconvinced. AES has a Spatial Audio commitee who tries to tackle this problem. AES Technical Committee Spatial Audio
Actually in a large room, where the modes are so dense at low freq., low freq. 'localization' is very real, but manifests in the form of envelopment.
How great would be to have this same phenomenon in a small room? But due to sparse modes IACC is too high for this to happen. How to excite the soundfied (in a small room) in the low frequencies to emulate this effect? James Johnston thinks he has an answer, involving different recording techniques as the norm. Same thinks Griesinger. Others remain unconvinced. AES has a Spatial Audio commitee who tries to tackle this problem. AES Technical Committee Spatial Audio
Fun little test. 🙂
I could hear a difference in left-right location easily down to 120Hz, although even as high as 320Hz the precise location is somewhat indistinct - "somewhere to the left" or "somewhere to the right" rather than appearing to originate from the speakers specifically.
At 80Hz I could just detect a difference, not so much in perceived location, but a sense of "more pressure" on one side than the other which was enough for me to reliably identify left or right. (I think this may be partly due to imperfectly balanced standing wave patterns in the room)
At 60Hz and below I was completely unable to tell a difference in location, tonal balance or anything else for that matter - even though I was sometimes cheating by watching a level meter to see which channel was playing I could not hear any difference between the two channels.
I've noticed this before - standing wave notches can be so deep and precisely located that one ear can be almost completely notched at some bass frequency from one speaker while the other ear still hears something, and vica versa, allowing you to identify which speaker is playing even as low as 60Hz or so, in fact I'm pretty sure thats what's happening in my listening test above at 80Hz as I know that is a problem area for standing waves in this room.
If you sit centrally aligned between left and right room boundaries and play bass from only left or right speaker, you will find one or more spot frequencies where the bass is mostly notched out due to room modes, but what is left appears to be "out of phase" between the ears, which sounds very odd and disconcerting. In actual fact what each ear receives is 180 degrees out of phase because each ear is on opposite sides of the cancellation node.
It's surprising how many people insist that its not possible to have significant amplitude or phase differences at bass frequencies between left and right ears - because of room modes and boundary cancellation, it is possible, in fact its hard to avoid at one or more spot frequencies without multiple distributed bass sources.
This is one reason why I've always been against systems that cross over two main speakers to a single asymmetrically located sub below say 80Hz - this is exactly where you're going to run into this problem and get weird left/right ear asymmetry at some bass frequency.
Last edited:
I can see a big hole in your methodology - ambient outdoor sounds.
If you visually saw where the subwoofer was located before the test, and there was even the slightest outdoor noise such as birds chirping in a tree, traffic noise in the distance, your brain would record the location of the subwoofer in its "mental picture" of the auditory scene relative to these other noises.
Spin yourself around on a chair or whatever, and when you stop despite keeping your eyes closed you will be able to determine your orientation from the other ambient noises and your brain will "fill in" the blank and identify where the woofer is from memory. Even steady wind on your face will give you an orientation clue relative to the world with your eyes closed.
Don't underestimate the power of the brain to continue to "believe" previously noted locations of items within the acoustic scene. This effect is real and extremely powerful.
To eliminate this effect you could try a few different approaches. One would be to have someone else choose a random location for the subwoofer each time, lead you out blindfolded to the listening position, then switch on the sound for you. Because you have never had a visual reference for where the subwoofer was relative to your orientation in space (as determined by ambient sounds) there would be no prior cue for the brain to fall back on.
Another possible approach would be to have 3 or more subwoofers set up in a circle and adjusted for equal SPL at the listening position. An assistant (or computer) then chooses one to play at random.
In both tests it would be important for the tone to be turned up slowly so that the higher frequency components from the switch on transient couldn't be used to locate the source.
Unfortunately I don't think you can conclude anything from your test simply because you knew where the subwoofer was before the test, and despite spinning around blindfolded you would still be able to determine your orientation from other sounds and thus subconsciously deduce where the subwoofer is.
Last edited:
I absolutely concur with the phenomonon of standing waves creating nulls in one ear or the other, but because of that I find the location of the bass sources impossible to figure out. A null in the left ear swings the apparent source hard right (and vice versa). In general, sine waves in a live room will be hard to place at any frequency.
We need the leading edge of transient waveforms to consistantly place objects. Thats why telephones (used to) ring with the "packeted" waveforms used: it created lots of wavefronts to figure out direction with.
David S.
Yes, and leading edge of transient waveform is a high frequency component of the signal, hence our ability to discern its direction. Making deliberately the (signal) zero-crossing different (in time) between ears at low-frequencies in a small room is not so easy to do.
No you don't localize sources, you localize 'something somewhere', not depending on real sources. It is an unwanted artifact of the small room. Effect only depends on listener/source placement, not so much on the source material. If the modes get sufficiently dense, as in near/above Schroeder, then it's a wanted thing - you get some spaciousness from the room.(and from the recording/performance) (unless you'r a padded-cell type) and can reliably localize.DBMandrake said:
And try the test not in the steady state. Use a 2-3 ms LP filtered signal - much more like music - and see if you can detect its location - blind.
The whole discussion started from a comment that the ear begins to loose it abilities below about 500 Hz and falls to almost nothing at the very lowest frequencies. Its not a black and white situation like I can or cannot localize LFs. Its a continuum where your ability is lowest at the lowest frequencies.
The whole discussion started from a comment that the ear begins to loose it abilities below about 500 Hz and falls to almost nothing at the very lowest frequencies. Its not a black and white situation like I can or cannot localize LFs. Its a continuum where your ability is lowest at the lowest frequencies.
If you read carefully I didn't say I was able to localize the source direction at low frequencies, as in accurately pointing to the direction the sound was coming from, I said I was able to identify which speaker was playing. (Down to 80Hz in this particular situation, and only just)
In fact I even said that as high as the 320Hz test the source location was somewhat vague, being "to the right" or "to the left" at most. I was able to tell which speaker was playing with confidence, but could not have picked the exact direction if the speakers had been hidden behind an opaque curtain. Accurate pin-pointing of the lateral location of a sound to a few degrees requires frequencies much higher than 320Hz.
Whether its an "unwanted artefact" of a small room or not, differences between bass produced by a left and right speaker can sometimes be heard to fairly low frequencies precisely because of the effects of the room and its standing wave patterns, something that is important to be aware of rather than just believing that "any" source of bass at frequencies below 80Hz cannot give a localisation cue, false or otherwise.
The larger the room the lower these standing wave induced false localisation cues can be generated, and thus the lower in frequency a difference can be detected between left and right speakers.
There have been several statements made on this forum (and others) that you can't locate bass below 100Hz. Well, you can. In my tests sometime better than sounds above 6KHz.
But no argument that it does get harder the lower it goes, at least in most listening rooms.
In answer as to whether I could locate the sounds because I know where the speakers are I would say "Close, but no cigar". There were some test tones - notably 120Hz and 40Hz where I could not tell. I still knew where the speaker were, but could not tell which it came from. That was clear, not ambiguous.
I originally did files that panned slowly left-right, right-left. In some cases I'd hear a flip, were the sound seemed to be coming from the speaker it shouldn't be. I suspect room nodes, just like when I couldn't tell the direction.
Interestingly, even on headphones the location can be kinda ambiguous below ~50Hz. Maybe it's leakage of the open back phones?
That said, even if we CAN locate very low sounds in a typical listening room, is it desirable? Maybe we simply enjoy enveloping bass more. That's what the Lexicon papers tell us.
Thanks everyone for the great comments. All good food for thought!
But no argument that it does get harder the lower it goes, at least in most listening rooms.
In answer as to whether I could locate the sounds because I know where the speakers are I would say "Close, but no cigar". There were some test tones - notably 120Hz and 40Hz where I could not tell. I still knew where the speaker were, but could not tell which it came from. That was clear, not ambiguous.
I originally did files that panned slowly left-right, right-left. In some cases I'd hear a flip, were the sound seemed to be coming from the speaker it shouldn't be. I suspect room nodes, just like when I couldn't tell the direction.
Interestingly, even on headphones the location can be kinda ambiguous below ~50Hz. Maybe it's leakage of the open back phones?
That said, even if we CAN locate very low sounds in a typical listening room, is it desirable? Maybe we simply enjoy enveloping bass more. That's what the Lexicon papers tell us.
Thanks everyone for the great comments. All good food for thought!
I don't think it is, when that localisation is artificially coming from standing wave patterns in the room. Most music has bass panned fairly centrally so I don't normally notice it too much, but I do have a few songs that have bass either panned hard to one side or deliberately out of phase, and the severe changes with listening location sounds rather odd and unpleasant.
Achieve adequate modal smoothing with multiple subs around the room and this problem goes away. Smooth "enveloping" bass that doesn't change much as you move around slightly is much more enjoyable IMO.
Not sure what signal you are suggesting here. My test signals (which are posted in this thread) are noise that is low pass filtered. I used noise because with steady state sine waves I can't locate nothin!
What would the content of the 2-3mS low passed signal be? I'm game to try.
Whatever signal you want inside of the envelope. Noise, sines, whatever. Just gate it at 2 s (sorry I meant seconds the first time not ms.) and THEN LP filter it - not the other way arround that isn't correct. It will be a slowly rising and falling signal, much like bass in music.
Steady state is just far too easy a test. Music is never steady state. There also must not be any discontinuities in the signal - those are HFs and even the slightest presence of them will be sufficient clues to localize. Even port noise, which is broadband. (I have to laugh when people claim to localize on "LF sounds" which clearly have HF content. )
Thanks for the info. Coincidentally, that's basically what I have. It's ~2S of white noise with a logarithmic fade in and fade out to avoid clicks or pops. It was then low pass filtered at 60dB/octave for each band. Each low passed file was made from the full white noise, not from any other band. It's a pretty smooth rumble. The 2 second noise alternates left and right.
This wouldn't work on my system because of the multiple subs all arround the room. Maybe the inability to localize the bass helps make multiple subs sound so good.
Earl,
From your post Post #23
"The whole discussion started from a comment that the ear begins to loose it abilities below about 500 Hz and falls to almost nothing at the very lowest frequencies. Its not a black and white situation like I can or cannot localize LFs. Its a continuum where your ability is lowest at the lowest frequencies."
As Pano mentioned in post #38, the Rumble pingpong zip file in #14 is ~2S of white noise with a logarithmic fade in and fade out to avoid clicks or pops.
As I stated in #12, high order filter ringing and white noise rather than pink may contribute to localization of those tracks, but at least Pano provided a usable test source to compare.
Although a 2 second gate is similar to some music, I find a more rapid source with varying attack and delay (like filtered kick drum) easier to locate than a more “slowly rising and falling” source.
Legato LF passages are more difficult to locate in space than percussive, sine waves are far more difficult to locate, source material (not just HF content) makes a big difference in localization.
I first listened to some of the low tones Pano provided on Sony MDR-7506 headphones, even the 32 Hz track (the lowest) clearly panned back and forth between my ears, no surprise to me, though I was surprised Pano did not hear that way with his headphone test.
Listening on my stereo speakers, the upper tones from 120-60 Hz were noticeable as left/right discreet sources, while below 60 Hz where my main speakers naturally roll off to the point where they are not audible over the sub I could only ascertain the sub direction.
Listening was done at a level that THD was 40 or more dB below the fundamental, even at 12 inches from the speakers there were no identifiable HF noises.
I also did several spins in place till dizzy and pointed to where I heard the noise. I always pointed towards the sub, though with the lower frequencies (32 and 40) pointed only within about 60 degrees.
I agree with your statement that localization ability is a continuum that is lowest at the lowest frequencies.
With that continuum, and room anomalies, there is no doubt that statistically a high percentage of people may not to discern sound source location below around 100 Hz, given a 24 dB per octave acoustic crossover and a legato source in a small room.
100 Hz or lower with a high order crossover avoids the most objectionable (to me) sub problem, that of low vocal notes apparent location moving from L/R to the sub location.
At any rate, if someone claims they can’t tell where an asymmetrically placed sub crossed at 100 Hz is located, fine with me, but if they think I can’t place it in the rooms I listen in, they are mistaken.
Art Welter
- Status
- Not open for further replies.