|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: "Inaudible" HF sounds found to affect perception of sound posted by clarkjohnsen on August 25, 2003 at 13:47:37:
I think this is what high sampling rates would do, ie push the nasty sharp filters up out of the way so their effects are lessened down low.I am not led by this test, despite my attachment to starting transients and decays as part oif this, to conclude that we have to HEAR these VHF EH frequencies, it could also show that we need very wide bandwidths for digital if we are going to use steep antialiasing filters OR that we need gentle slopes at least lower down?
If we accept the primacy of starting transients (attacks) and decays - for each note - in our getting the music, then this could well be what is going on.
IE - we need to have a wide frequency response so that we don't have filter effects inn the audible range of the fine details of these parts of all notes.
I would expect that super tweeters are going to be of most value in sytems using analogue tape, MC's for vinyl, and systems with high sampling rate digital sources, and perhaps even resampling devices.
Testing for this, with and without the ST, for the wide band case could be revealing, I think.
It might also depend on the down band filter effects of the ordinary tweeter and its LP filter, when run without the ST. Let alone the matching of the ST and its HP - ie to the main tweeter's roll-out.
I mean Quad 57's get this stuff pretty right on vinyl, and yet they die pretty low down don't they?
Follow Ups:
I don't know if this has been mentioned - but one part of this which doesn't seem to be addressed if why we have to hear these "inaudible" frequencies to it to make a difference. Has there been any research into whether your ears are solely responsible for sound perception? Does your stereo sound different when you close your eyes?As an example, one's sense of taste in connected to ones sense of smell so taste is not purely the result of your tastebuds. Has there been any research as to the effect on manipulation of some of your other senses on your perception of sound? I'm not suggesting that inaudible frequencies cause some kind of nose-hair resonance which changes sonic perception (although I may be on to something ;-) but maybe the presence of those frequencies stimluate something entirely different which may or may not effect perception. The study cited above seeems to suggest there is some physiological response.
I know at some really loud concerts there have been bass notes thay I would describe more as being felt than heard. When you feel your internal organs moving around in such conditions if definitely affects your perception.
Another thing I would be interested in seeing is measurements in a real room with ultrasonics present to see if any resonance of room objects beating with audible frequencies cause spurious noise in the audible band. Has any study done this?
The statement, in the abstract at the URL, that this might be 'a previously unrecognised phenomenon' is a bit hard for me to take, there are heaps of such reports.This sort of intrepretation should and would be seen as an extension of the work on attacks and decays, if it were not for the fact that the continuous tone and its harmonics so strongly, and incorrectly, dominates our discourse on music and what thus might matter in audio.
If the primacy of starting transients (aka attacks) and decays in our getting of music were more widely understood, esp. instrument level attacks map to the dear old first arrival window, I think we might be more willing to test for what is happening.
The affect, for me of listening to many hi-fi systems, and to electric keyboards, before CD or digital LP's, and to CD a lot of the time - esp. in decade 1, is often the same. A growing sense of irritation, that I can't really hear what is going on, both from individual players and singers, AND between them.
It seems to me that this is almost certainly due to damage to the attacks and decays, OR the complete absence of appropriately modelled attacks from electronic keyboards.
After all, the piano, harpsichord and clavichord are 'percussion' instruments, and thus do not have ANY continuous tone anyway, just an attack, and a decay around the pitch. Viz for cymbals, drums etc
IF I am being unfair about "the continuous tone and its harmonics being 'the dominant paradigm' in audio", I would be happy to be corrected - with evidence.
The continuous tone on its own is as much use as guessing, in identifying an instrument by it's supposed 'charcateristic timbre'. IE remove the starting transient (attack) and we get a statistically significant and consistent degradation in identification, remove the decay, and we are just guessing.
Further, most of the expression is in the attack and decays.
Clearly, any tone(?) and expression from a percussionist, can only reside in attacks and decays.
HowdyIf you think about voice, the attack and decay hardly matter at all compared to the continuous tone.
Also, for example, almost everyone can tell a stringed instrument from brass even with the attack and decay removed.
Even with percussion don't take it too far: The spectrum of the tones on percussion do indeed differ based on the amount of energy put into their production, i.e. loud piano sounds different than quiet piano, even if you remove the attack and decay. With more energy in, in general, you have a richer tone, e.g. more energy in the higher harmonics, (tho with bells, etc. the 'harmonics' have strange freqs compared to the fundamental.)
I'm not disagreeing that attack (and decay) are important, but, for example, harmonic distortions are also noticeable and don't affect attack/decay much...
Beg to differ. Speech intelligibility is greatly enhanced when a recording is played in correct polarity. And polarity is *all* about attack.
HowdyI should have said the tonal character of vocal singing is hardly affected by attack and decay. The biggest difference between an Irish tenor and a baritone or soprano is not in the attack and decay.
Ted look this up,
Attacks and decays are shown to be critical to individual voice recognition, but can be preserved in narrow band 'phone networks.Timbo
I am aware of this as a chorister a good 'leader' (a purple ribbon and bronze medallion in Anglican cath.s) needs a distinctive and longer attack, helps everyone else 'come in' at the right time.
HowdyOK, I'll try one more time.
Where as with pairs of instruments, you can cut off the attack and decay and it becomes hard to tell which is which, this is much less true with human voice when singing.
I'm not talking about intelligibility, nor recognition of speech nor speaker. I'm just pointing out that the singing voice is clearly a counter example to a hypothesis about attack and decay being the most important feature. In some sense the singing voice is deemphasizing the fricatives and emphasizing the steady state part of the voice so this shouldn't be surprising.
This post is made possible by the generous support of people like you and our sponsors: