This is true, but it gets completely masked by the noise as the stimulus, by design. I'm 100% sure that raised cosine crossfade (of 100ms or so) will not change the percieved results. I nulled 6 adjacent samples in a noise signal and cannot hear it although it is a rather strong artifact.
I can quite believe that, as I said way back in the thread, I accidentally introduced a -2dB dip at 16Khz in one channel (about 1/6th octave, not sure what that would be in Q) and it was enough to skew the stereo image half way to one side and cause me to get up and start searching for what had gone wrong with the system.
For changes that small I doubt it would be possible to "remember" an absolute reference of what the pink noise should sound like and recognise it on its own, or after a long gap, as you discovered. As a relative change with no time delay, yes.
While the perceived spectrum of noise fairly quickly returns to sounding "normal" for a small error like that, I find the change in perceived imaging of music to be more absolute - if you played me fully isolated pink noise samples of the two responses above I would almost certainly not be able to recognise a difference, let alone identify which was which.
But let me hear the two responses applied to music that I'm familiar with and by the time I'd head at least 30 seconds of each I could tell you which was which based on how it affected the imaging.
I guess I'm the young whippersnapper of this thread at 36 then.
My conclusion is the same - I'm hearing the bottom end of the bell curve on the 18Khz test. Unlike you my hearing seems to stop dead at 17.5Khz - even 1Khz higher I can turn the level up 10-15dB and don't hear a thing, although I think I do possibly sense something that doesn't sound like a tone, but more like a presence. (In the look behind you, something is creeping up on you kind of way) My cut-off frequency also doesn't seem to have changed from when I first measured it in my early 20's - even then it was still only 17.5Khz.
Last edited:
I suspect that this paper was never published because of serious errors it makes in its assumptions. On the first page alone it is erroneously calculated that the sound wavelength would be X, far too long to exist in the cochlea. By that assumes a fluid in a rigid walled pipe acting in compression. In the cochlea the fluid is coupled to a membrane which is hardly rigid and the wave motion is not compresional but transverse. In this case the wave velocity is much much lower than what was calculated and their whole argument goes out the window.
I did not read the entire papre, but what I read was very sketchy. The work done on the mechanism of hearing is substantial and I doubt that everyone else has got ot wrong.
Earl, please read the full paper and check out the experiments, before coming to conclusions.
To me, there must be something to it, for example their theory makes clear why a sligtly detuned overtone (say 100Hz and 200.5Hz) is easily detected compared to a perfect 2nd (by the 2 second beating in timbre that the former does produce) inspite of poor pitch resolution for LF sines. It also explaines timbre of chords shift from allpass functions (excess phase), at LF and MF, etc
To me, there must be something to it, for example their theory makes clear why a sligtly detuned overtone (say 100Hz and 200.5Hz) is easily detected compared to a perfect 2nd (by the 2 second beating in timbre that the former does produce) inspite of poor pitch resolution for LF sines. It also explaines timbre of chords shift from allpass functions (excess phase), at LF and MF, etc
Last edited:
I notice the same thing but closer to 15.5k (I'm 43). Listening to tones at higher frequencies gives me the impression that my ears have lost some of their flexibility, almost as if they've become scarified or hardened through use.
Although I can hear much higher with enough SPL, it feels as if my ears are relying on the inner mechanism to sense them (I'm guessing here), as if the outer one is no longer contributing.
Simon & Allen,
My observations are similar, above 17k there is no sensations of tone anymore, but it's more like an annoying feeling that there "still is something there". And it needs some 20dB increase (at 20k) to get that feeling.
As for imaging, I also find there is a sweetspot of HF balance, too much HF and the treble keeps sticking to the speakers whereas the rest of the image depth is developed at about 1m behind them, too little HF and it all gets dull. All within a +-1dB range or so deviation from my "house curve" which is sloping down some 6dB from 100Hz to 20kHz at the listening position. This is a bit more than usual but I tend to dislike bright speakers in general, from my early days on. Might have to do with a slight upper midrange hearing loss coming from playing in loud rock bands without hearing protection when I was a teenager.
My observations are similar, above 17k there is no sensations of tone anymore, but it's more like an annoying feeling that there "still is something there". And it needs some 20dB increase (at 20k) to get that feeling.
As for imaging, I also find there is a sweetspot of HF balance, too much HF and the treble keeps sticking to the speakers whereas the rest of the image depth is developed at about 1m behind them, too little HF and it all gets dull. All within a +-1dB range or so deviation from my "house curve" which is sloping down some 6dB from 100Hz to 20kHz at the listening position. This is a bit more than usual but I tend to dislike bright speakers in general, from my early days on. Might have to do with a slight upper midrange hearing loss coming from playing in loud rock bands without hearing protection when I was a teenager.
I've always felt this way as well. For me, it betrays a disdain of poorly done mids. I had a college friend who hated the upper midrange (so he preferred 12" paper cones in two-way systems with widely spaced first order crossovers). It took me a while to realise I had the same issue, and much longer to get a handle on how deal with it
It is impossible to accurately reproduce the perceived tonality of musical instruments as they would be heard in your room unless all of the overtones are reproduced in the right proportions, arrive at the correct distribution of angles, and at the right times. There are no speakers that meet these criteria. To reproduce them as they'd be heard at a large venue, you must also reproduce the fact that over the course of the time the reverberation dies out the relationshp of the harmonics changes. Typically a concert hall has an RT of about 2 seconds at 1 khz but only 1 second at 8 khz. This alters the tone. If you don't get the reverberation right, the tone won't be right either. There are no sound systems that can capture and reproduce that except binaural sound system which means being heard through headphones. Too bad they make sound appear to come from inside your head. They have a different flaw, no directionality. You turn you head and the sound turns with it. Your brain quickly concludes the obvious. Play the sound through loudspeakers and the sound appears to have been recorded inside a tunnel with you on the outside because all of the reverberation comes from the same direction as the source.
Sorry but I have not been saving the uncompressed files. They are too big to attach to the posts with their imposed limits.
Being just noise, I looked at the spectra both before and after mp3 conversion and didn't see any difference. Note that a filter of a defined Q has some width and you are just as likely to hear an effect from the lower skirt, if that reaches below your hearing limit. As I mentioned, I couldn't hear the pitch of the filter beyond 11kHz but could readily hear the 14k centered filter.
I don't think expectation bias is an issue since it is an off, then on, then off signal. If you are hearing the transition, I doubt you are imagining it.
Regards,
David
Huh? Have you ever listened to a binaural recording done with your ownears? You will be baffled by the realism. With such recordings directionality solely depends on room acoustics.
The phenomenon is very well documented and has been known since at least the early 1960s, probably a lot eariler. That's why it wasn't popular or adopted universally. It is not acceptable. It's also why headphone listening in general is not as satisfactory as loudspeaker listening. Perhaps if you keep your head in a vice so it can't turn it even slightly you might just convince yourself.
Psychoacoustics - Wikipedia, the free encyclopedia
"The brain utilizes subtle differences in intensity, spectral, and timing cues to allow us to localize sound sources."
This explantion is incorrect. Binaural recording can meet all these criteria but still doesn't work.
Dummy head recording - Wikipedia, the free encyclopedia
"Because each person's pinnae are unique, and because the filtering they impose on sound directionality is learned by each person from early childhood, the use of pinnae during recording that are not the same as the ultimate listener may lead to perceptual confusion"
Again the wrong explanation. Inverse filtering could correct for any FR differences between the dummy head and your own. Also a cast could be made of your own head and the dummy made of material that reflects and refracts sound exactly as your own ears do. Also since the ability to detect direction is not so critical of exact spectral content, changes to the FR of the sound field such as from substititing one speaker of the same type for another or a change in equalization of a sound system will not significantly change the perceived location of the sound. The concept is fatally flawed.
Figure out the right explanation and you will be well on your way to a far greater understanding of sound and hearing. You'll be where I was 38 years ago.
Here's a sample of a binaural recording demonstrating what's wrong with it. Use headphones.
Moonlight Beethoven in Binaural (USE HEADPHONES) - YouTube
and another;
Hendrik Andriessen - Ricercare (ajo) - binaural recording - YouTube
Pick anyone you like, they're pretty much the same. The musicians aren't in front of you and the imense space you hear at a live concert in a hall isn't there either. Everything is inside your head.
Moonlight Beethoven in Binaural (USE HEADPHONES) - YouTube
and another;
Hendrik Andriessen - Ricercare (ajo) - binaural recording - YouTube
Pick anyone you like, they're pretty much the same. The musicians aren't in front of you and the imense space you hear at a live concert in a hall isn't there either. Everything is inside your head.
I do own a Smyth Realiser and I also did my own binaural recordings, everything is externalized and there are no front-back reversals. I suppose I'm not at the same place you were 38 years ago - fortunately.
You don't need any kind of "realizer" to see that it doesn't work, just a pair of headphones. Here's another failed example.
Music from the Chronicles of Narnia - binaural recording - YouTube
Music from the Chronicles of Narnia - binaural recording - YouTube
You've never tried it yourself and yet you want to tell me what I hear? Pretty bold.
By the way, the examples you've posted don't work for me either. I had never good results with generalized HRTFs but my own recordings with in-ear mics work just fine. You should try it and hear for yourself what you've missed the last 38 years.
I've just given you three samples of binaural recordings that proves it doesn't work because the sound turns with your head and what you hear is not what you would at a real musical performance. But don't take my word for it, here's an example of an attempt to solve this problem using a "head tracker" and multiple binaural recordings made at the same time using the head tracker to switch between them. (That's also an old idea and there are others.) The relevant part starts about 3:32 into the presentation.
AES Tutorial: Binaural Audio Technology?History, Current Practice, Emerging Trends by Bob Schulein - YouTube
AES Tutorial: Binaural Audio Technology?History, Current Practice, Emerging Trends by Bob Schulein - YouTube
By the way, here's a "generalized HRTF"-example that works better for me. Externalization is very good but the typical front-back reversal.
QSound Labs: Virtual Barber Shop (Long) -- Binaural Audio Demo
QSound Labs: Virtual Barber Shop (Long) -- Binaural Audio Demo
I've heard this before. It's about as effective demo as it can get from binaural recording. Very suggestive but if you listen carefully the sound is confined to a space from just beyond one ear to just beyond the other. Now show me one where a musical performance is in front of you and you hear the immense space of a concert hall or cathedral.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.