In a general way, the proposition that humans can make finer distictions with pure tomes than complex ones is consistant with the practice of lossey encoding such as MP3. Even at it's highest resolution version, MP3 takes advantage of of the fact that under certain conditions uch as sudden peaks the brain lets some sounds totally mask others. This type of compression tries to identify the masked tones and and drop them from the encoding. I've listened to a couple of demonstrations of this and it works regardless of how much gold plate one has on ones ears.
A DIY example. Play an LP and burn or record it at 16/44 or better (i.e., standard Redbook or CD). Find a few "pops" or "clicks" to test. If you apply a pop or click filter to this, much of the time it completely goes away, however sometimes the "pop" is replaced by a softer "thuup" sound that is just as annoying. The filter whether soft ware or hardware is failing right? Maybe. . Use something like "Cool Edit" or "Goldwave" that lets you magnify the pop untill you can see the actual waveform before and after filtering. You should see the "thuuup" as several cycles of ringing following the initial impulse. The filter actually did what was asked of it, it cut out the spike. It left the ringing un changed. The only thing is that you could not hear the ringing or "thuup" so long as it was preceeeded by the initial impulse. In short you very own brain was engaged in lossey signal processing triggered by the close juxtaposition (in time) of two dissimilar sounds.
Conclusion: Substantial eveidence that I didn't get from a book or learned article that human hearing is LESS precise when listening to complex signals.
A DIY example. Play an LP and burn or record it at 16/44 or better (i.e., standard Redbook or CD). Find a few "pops" or "clicks" to test. If you apply a pop or click filter to this, much of the time it completely goes away, however sometimes the "pop" is replaced by a softer "thuup" sound that is just as annoying. The filter whether soft ware or hardware is failing right? Maybe. . Use something like "Cool Edit" or "Goldwave" that lets you magnify the pop untill you can see the actual waveform before and after filtering. You should see the "thuuup" as several cycles of ringing following the initial impulse. The filter actually did what was asked of it, it cut out the spike. It left the ringing un changed. The only thing is that you could not hear the ringing or "thuup" so long as it was preceeeded by the initial impulse. In short you very own brain was engaged in lossey signal processing triggered by the close juxtaposition (in time) of two dissimilar sounds.
Conclusion: Substantial eveidence that I didn't get from a book or learned article that human hearing is LESS precise when listening to complex signals.
Yes, I agree that sonic phenomenon can be identified more precisely under “laboratory” conditions and using pure sinewaves, then it can by listening to music. This is not to say that these phenomenons cannot be heard when listening to music. IMO they can be, and are heard.
The link I included in post #54 shows that ultrasonic frequencies can and do combine to produce audible sound. Again, if we were dealing only with monophonic reproduction, this would be a moot point. However, when we talk about stereo reproduction and multi-mic and multi-track recordings, this can, and IMO, does make a difference in the sound of the final output of the speakers.
Don’t get me wrong, I love technology as much as the next guy. Most of the aspects of CD technology are a giant leap above vinyl music reproduction. The 22kHz brick-wall filter is not one of these.
The link I included in post #54 shows that ultrasonic frequencies can and do combine to produce audible sound. Again, if we were dealing only with monophonic reproduction, this would be a moot point. However, when we talk about stereo reproduction and multi-mic and multi-track recordings, this can, and IMO, does make a difference in the sound of the final output of the speakers.
Don’t get me wrong, I love technology as much as the next guy. Most of the aspects of CD technology are a giant leap above vinyl music reproduction. The 22kHz brick-wall filter is not one of these.
roddyama said:
Actually my comment regarding pure tones/noise versus music wasn't to do with the precision of identification but rather just the simple detection of a difference.
For example, detection of a difference in say frequency response is more sensitive when using noise than when using music.
se
Charles (aka ("Phase Accurate"),
I should have printed an image of the wave form, then I might still have it around. It was a single posative spike. My guess from memory was that it was between the equivalent of 1kHz to 5kHz (keeping in mind it was a single half cycle) so that the time would havebeen between 1/1000/2= .5mSec and 1/5000/2=.1mSec. (Check my arithmetic). The ringing that follows was about 10 cycles. They totally swamped the music content. Itried mually cutting them out of the wave file but the result often sounded as bad as the orginal.
Before there were CDs I had an SAE5000 clicp & pop gadget in my tape loop. It was very frustrating as sometimes the cure was worse than the disease. When I was transfering som LPs to CD a year or so ago my curiosity was raised and I posed this question: How is it that both an analog process and digital prosess appear to have the same audible anology. Namely the "'thuuup" after the click was removed. That's how I figured out that neither process created the anpmoly, they just let me hear it. If I were more seriously interested in transfering LPs to CD I would find a standard pop and play it with a variety of cartiges/styluses. I bet the ringing has a mechanical origin and that choice of stylus and presure an influence (minimaize?) it.
I should have printed an image of the wave form, then I might still have it around. It was a single posative spike. My guess from memory was that it was between the equivalent of 1kHz to 5kHz (keeping in mind it was a single half cycle) so that the time would havebeen between 1/1000/2= .5mSec and 1/5000/2=.1mSec. (Check my arithmetic). The ringing that follows was about 10 cycles. They totally swamped the music content. Itried mually cutting them out of the wave file but the result often sounded as bad as the orginal.
Before there were CDs I had an SAE5000 clicp & pop gadget in my tape loop. It was very frustrating as sometimes the cure was worse than the disease. When I was transfering som LPs to CD a year or so ago my curiosity was raised and I posed this question: How is it that both an analog process and digital prosess appear to have the same audible anology. Namely the "'thuuup" after the click was removed. That's how I figured out that neither process created the anpmoly, they just let me hear it. If I were more seriously interested in transfering LPs to CD I would find a standard pop and play it with a variety of cartiges/styluses. I bet the ringing has a mechanical origin and that choice of stylus and presure an influence (minimaize?) it.
To Ian (post #100)
I think I've designed enough amps to know the difference between the input LP filter and slew rate limit.
However, any cirsuit with a feed-back loop can oscillate, provided that there is more than unity gain at resonance frequency. This does not depend on input LF filtering, but solely on the nature of phase lags and feed-back loops. Hence, I disregard the input LF filter, as most (proper) designs will be able to do without it. If your gear is well designed, you will not apply HF signals ( >> 20 kHz) to an amp input, which makes the input LF filter somewhat academic if the rest is designed appropriately).
(For reference/comparisaon, a tone generator can be made very reliable by adjusting values of a conventional negative feed-back amplifier, simply by adjusting the loop gain at its resonance frequency. This is even without any signal at the input.)
Jennice
I think I've designed enough amps to know the difference between the input LP filter and slew rate limit.
However, any cirsuit with a feed-back loop can oscillate, provided that there is more than unity gain at resonance frequency. This does not depend on input LF filtering, but solely on the nature of phase lags and feed-back loops. Hence, I disregard the input LF filter, as most (proper) designs will be able to do without it. If your gear is well designed, you will not apply HF signals ( >> 20 kHz) to an amp input, which makes the input LF filter somewhat academic if the rest is designed appropriately).
(For reference/comparisaon, a tone generator can be made very reliable by adjusting values of a conventional negative feed-back amplifier, simply by adjusting the loop gain at its resonance frequency. This is even without any signal at the input.)
Jennice
Yes Steve, I agree there are differences in the ability to detect various sonic phenomenon based on signal source and conditions. It is certainly not a black and white issue, but one of degree and one that relies heavily on human perception. To me, this is a moot point. Under certain conditions, certain sonic phenomenon are detectable and under other conditions they are less detectable.Steve Eddy said:
There seems to be a general position in this thread against the roll of ultrasonic sound as a contributor to the overall character of reproduced music. The point I want to make is that by applying the 22kHz Nyquist filter, you will lose the “music” that exists in the range above 22kHz. The studies investigate very specific frequencies under very specific conditions. Music exists in a full range frequencies and under a multitude conditions. One study of one set of frequencies, under specific conditions is like sticking your toe in the ocean in Florida and concluding that the oceans are warm.
The manufacturers of CDs and CD players imposed this limit on us back in the 80’s. It was set by the cost of the technology at the time without regard for the ultrasonic contribution to the character music/recording. The question they had to answer was “What level of technology is the average consumer willing to pay for?” This is an artificial (digital) limit being placed on a real world (analog) phenomenon.
roddyama said:
Certainly.
Yes. But I don't think at this point anyway the question has been confidently answered as to whether we actually need content much above 20kHz on the reproduction side of the equation.
I mean, is it truly the lack of content above 22kHz that limits 44.1kHz CD or is it the sub-22kHz artifacts of the brick wall anti-aliasing filter? Or perhaps both.
se
Or perhaps neither
The brick wall stuff has always been easy to deal with; if the Red Book had standardized an anti-imaging filter (brick wall would be fine), we could have avoided all the oversampling and weird phase correction schemes. Instead, we have a mishMASH, and the lack of standards means we get it wrong most of the time.
I've been following this thread with some interest. First, I'd like to say that I read the article about the "hypersonic effect." I don't know exactly what to conclude from that article. Occipital alpha activity is a normal response in an awake individual with his eyes closed. I don't knoe how they measured increased alpha in experiment 2 in which the eyes were open "to enhance vigilence." Eye opening should extinguish alpha waves. Furthermore, EEG is very nonspecific. In addition, I would be wary of someone who is touting their own wares, "Most of the conventional audio systems that have been used to present sound for determining sound quality were found to be unsuitable for this particular study....The speaker system was designed by one of the authors (T. Oohashi) and manufactured by Pioneer Co., Ltd. (Tokyo, Japan). This sound reproduction system had a flat frequency response of over 100 kHz." Human physiology is geared more for detecting differences than absolutes.
The question I have from the standpoint of trying to design amps is should there be a high frequency limit? When listening to music, we hear often a summation of many different waveforms. We hear a flute playing over a violin. At some point the peak waves may be within 1/20,000 of a second away from each other. We don't hear a different tone higher than 20K but the two individual instruments with clarity if the system is able to reproduce the signals with enough definition. I feel that the higher the bandwidth of a particular system, the more definition one experiences. That's probably why vinyl arguably has more definition that digital. That's my $0.02.
The question I have from the standpoint of trying to design amps is should there be a high frequency limit? When listening to music, we hear often a summation of many different waveforms. We hear a flute playing over a violin. At some point the peak waves may be within 1/20,000 of a second away from each other. We don't hear a different tone higher than 20K but the two individual instruments with clarity if the system is able to reproduce the signals with enough definition. I feel that the higher the bandwidth of a particular system, the more definition one experiences. That's probably why vinyl arguably has more definition that digital. That's my $0.02.
"There's Life Above 20kHz"Steve Eddy said:
http://www.cco.caltech.edu/~boyk/spectra/spectra.htm
Steve Eddy said:
I think since the 80's, the industry has been so busy trying to figure out how to do it (cheaply), that nobody (with a loud enough voice) is stopping to ask if it's right.SY said:
Or perhaps neither
The brick wall stuff has always been easy to deal with; if the Red Book had standardized an anti-imaging filter (brick wall would be fine), we could have avoided all the oversampling and weird phase correction schemes. Instead, we have a mishMASH, and the lack of standards means we get it wrong most of the time.
The technical limitations are the digital sampling rate and the BW capability of the record/playback string. I believe the technology is already available at a reasonable cost to extend the range another octave to 44kHz.
roddyama said:
Yes, I'm familiar with Boyk's piece. I didn't say there was nothing up there. What I'm saying is that I don't know whether it actually matters to the listener and therefore necessarily needs to be recorded and reproduced. The research I've read so far doesn't seem to be quite definitive on that point.
Ultimately the loudest voice out there is the one with the most money. And that's the mass-market consumer.
Sure. And it's already coming to fruition in the form of DVD-A and SACD.
se
Just the death-rattle of a dying medium. The good news is that we'll be able to choose our formats on a custom basis.
Steve Eddy said:
Yes. The reason most equipment is crappy is that most mass-market consumers (MMC) are crappy listeners. That is not to say that they cannot hear above 20kHz or that they cannot resolve fine details or discern differences from one playback to another... but that they may not have honed their ear-brain coordination or care enough about the details of the music to warrant learning how.
However, I personally feel that there is something about the way we (on this board) can hear that makes having better equipment a necessity. Maybe we can hear >20kHz, or maybe we can discern the diffractive effects of incoming waves, or maybe we can feel it in our facial hairs or maybe we place value on the individual notes as well as the poetry of the lyrics. Unfortunately, none of this matters. Whatever the difference, there are not enough of us to force changes in the mass-market and so we either get crappy equipment, spend a lot of money or DIY.
Yeah, Jensen, so what...
Well, it doesn't matter much to me if the reason is better top end FR or if it is the way I hear wave interference or IMD. I know something about my hearing is better than average. If a medium like DVD-A or SACD delivers that extra "information" to my ears for translation by my brain, then sign me up for one. (Which I've done.)
What I don't want to have happen is for these mediums to die out just because most consumers can't hear their benefits. If some studies on human hearing "prove" that humans discern audio information >20kHz, then that might provoke the big fabs into making equipment good enough to resolve it. Just like the inkjet printer pixel race, a frequency response race might just put good equipment within reach of the mortals who can appreciate it.
ensen.Thats one of reasons for going ahead. After all, the technology behind the CD is already a quarter of a century old !!
At a demo I once attended they played a recording with a choir from CD and SACD. The latter indeed offered a clear (though not large) improvement in ambiance and "being there" experience.
Regards
Charles
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.