Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Jon Hodgson]

maxdimario wrote on Sat, 26 November 2005 11:28

"Suggestion can have a HUGE effect on our perception, as has been shown many times, and not just in audio. That's why blind tests and numbers are vital, but here's how it should work..." you can't say that to a person who spent years comparing electronic circuits, like myself.

Yes I can. You may be experienced, but you are only human.

I've seen suggestion, preconception and misdirection convince people of all sorts of things, whether it's one cable sounding better than another, two identical files sounding different, making sick people feel better (the placebo effect) or that a losing ticket was a winning one (Derren Brown, watch him if you ever want to be freaked out by how readable and suggestable we all are).

As for you personally, despite your experience and passion about the subject, and the fact that I don't doubt you have good hearing and suspect you are a fine AE I've seen you make claims which I am CONFIDENT would not stand up in a properly administered blind test.

In fact when I warn of suggestion and preconceptions having a huge effect on our perception, you are EXACTLY the sort of person I talking to.

[Ronny]

Jon Hodgson wrote on Sat, 26 November 2005 11:11

maxdimario wrote on Sat, 26 November 2005 11:28 "Suggestion can have a HUGE effect on our perception, as has been shown many times, and not just in audio. That's why blind tests and numbers are vital, but here's how it should work..." you can't say that to a person who spent years comparing electronic circuits, like myself. Yes I can. You may be experienced, but you are only human. I've seen suggestion, preconception and misdirection convince people of all sorts of things, whether it's one cable sounding better than another, two identical files sounding different, making sick people feel better (the placebo effect) or that a losing ticket was a winning one (Derren Brown, watch him if you ever want to be freaked out by how readable and suggestable we all are). As for you personally, despite your experience and passion about the subject, and the fact that I don't doubt you have good hearing and suspect you are a fine AE I've seen you make claims which I am CONFIDENT would not stand up in a properly administered blind test. In fact when I warn of suggestion and preconceptions having a huge effect on our perception, you are EXACTLY the sort of person I talking to.

You'll be talking to a brick wall, until they start running blind tests on their own. Only than will they realize how really inconsistant the ear and auditory cortex are and how susceptible they are to non measurable phychoacoustics. A nice word to describe it would be psychosematic.  

Jon is correct, you must eliminate as many of the human elements as possible, expectations lead to imagination. Happens with all of our senses, not just auditory perception. No doubt this is why you hear all these claims from newbies on the internet and sometimes not so newbies, that don't follow the laws of physics. You should go into all audio evaluation tests armed with the knowledge that the weakest link in the audio evaluation chain is the auditory cortex. The mind is a huge sway factor. It's not that the ears are really poor, but the problems lie in assembling the data that the ear collects and coming up with consistency "every time". If you can't repeat the results every time, than there is either no difference, or the differences are too insignificant to worry about.

[maxdimario]

some things I am pretty sure about consistently.

consider I've been listening for the same things now for 15-20 years, so it might take me a minute, but I am pretty sure.

I used to kill myself for weeks trying to decide between one variable in a circuit and another, going back and forth.. trying to decide.

the senses can develop if one is nuts enough (dedicated?) to focus in.

the only exception being in extreme states of stress.. then I do find I lose my hearing... or after very loud sounds..

[danlavry]

maxdimario wrote on Mon, 28 November 2005 12:37

some things I am pretty sure about consistently. consider I've been listening for the same things now for 15-20 years, so it might take me a minute, but I am pretty sure. I used to kill myself for weeks trying to decide between one variable in a circuit and another, going back and forth.. trying to decide. the senses can develop if one is nuts enough (dedicated?) to focus in. the only exception being in extreme states of stress.. then I do find I lose my hearing... or after very loud sounds..

The ear can easily tell you the difference between musical notes (pitch) and some people can be trained to have absolute hearing. Mother nature intended us to respond to volume, pitch and much more of the content we find in air vibrations we call sound.

But going back and forth between say some electrolytic cap and a polyester, in a circuit made out of hundreds of parts (variables) is better left for test gear first, ear last.

True, a certain part that seem to perform very well (or poorly)in dozens of different applications and designs, is worth looking into. But more often then not, people come to a hasty conclusion based on limited understanding (or no understanding at all).

I have come across statements about OPamps (amplifiers), such as XXX is great (or YYY sounds bad). Say XXX could have been tested as inverting amp (where input common mode is a non issue), or XXX was tested with clean power supply rails (though it has a poor power supply rejection), or XXX load is very high... The same OPamp could be could bad news in a different circuit, such as non inverting (common mode problems), noisy and unsteady supply, heavy load... There are "endless" combinations out there.

Another set of example would be caps. Some came to believe in "no electrolytics in the signal path", a statement that can be very true or very false, depending on circuit architecture.
A statment that claims "no electrolytics in the signal path is a good thing" is clear indication for one's lack of understanding of circuit design!  

Some prefer poly caps everywhere, which is silly, given that polycarbonate is very different then polypropleyne, that polyester has it strong points, that some applications call for ceramics, tantalum, aluminum...

I have come across many statments about caps, resistors, transistors, wires... many of them are based on ironoius conclusions based on listening. The listeners were of course honest and serious, but the lack of understanding for the physical reasons that cause the observations, can lead to wrong generalizations.

Another way to put it - The voltmeter is the wrong tool to tell you about pitch, the ear is the wrong tool to tell you about voltage...

The ear can be trained to do many things (all sound related). The ear can tell you what sounds right, what it likes (or does not) and so on. The ear can not tell you that something sounds better because there is some 50KHz audio content in the air. The ear is not trained to hear 50KHz, in fact it can not react to 50KHz.

Last comment - there is too much emphasis regarding the importance of filters in how converters sound. Filters are important, but can be be designed well. There are other issues I consider much more important, from the basic approach to many circuit details.  

Regards
Dan Lavry
www.lavryengineering.com  

       

 

[maxdimario]

Quote:

The ear can be trained to do many things (all sound related). The ear can tell you what sounds right, what it likes (or does not) and so on. The ear can not tell you that something sounds better because there is some 50KHz audio content in the air. The ear is not trained to hear 50KHz, in fact it can not react to 50KHz.

you insist on this point and you are absolutely right I can't hear above 16 KHz.

but to balance out your argument, that does not change the fact that when somebody hears an improvement, given that he is capable of being objective, the improvement is real regardless.

why is another matter, less important to the end user, more important to the manufacturer, I suppose.

I feel that everyone should make it a point to listen attently and try and make their own observations about sound at an early stage.. some may find that they agree with convention.. others not.

the goal is to improve, and improvement can be slow and progressively difficult.

[danlavry]

maxdimario wrote on Mon, 28 November 2005 19:16

Quote: The ear can be trained to do many things (all sound related). The ear can tell you what sounds right, what it likes (or does not) and so on. The ear can not tell you that something sounds better because there is some 50KHz audio content in the air. The ear is not trained to hear 50KHz, in fact it can not react to 50KHz. you insist on this point and you are absolutely right I can't hear above 16 KHz. but to balance out your argument, that does not change the fact that when somebody hears an improvement, given that he is capable of being objective, the improvement is real regardless.

That is fine with me - if someone hears an improvement, if someone likes something, that is all cool with me.

My issue is with those that try to attribute the improvements to parameters that they do not know.

Example?

Some early claims that we need to go to 192KHz were based on "more dots is better", and when that fell apart, on "better time resolution", and when that argument fell apart, there was the wrong analog anti aliasing filter argument, and so on... and now, with no single technical argument, no physical explanation, nothing to back it up, it is all about "I like the sound", backed up by huge marketing advertising investment.

Yet whatever you hear is not due to the 96KHz audio bandwidth of 192KHz sampling. What you hear is due to many factors and design specific parameters.

What you like, and what sounds good to you is fine with me, as long as we do not build on it for far reaching wrong generalizations and conclusions...

Regards
Dan Lavry
www.lavryengineering.com  

[Schallfeldnebel]

And what about the quality of sound before it comes into the AD converter. When I am editing my organ recordings, my dog Laika, see avatar, often sleeps in the studio. As long as I have recorded the sharp loud high volume plenum from the organ with DPA omni microphones, my dog stays in the studio, and sleeps on. Whenever I used any other type of microphone outof my collection, the dog starts to turn the ears, gets nervous and finally after a minute is leaving the studio.

Dogs hear up to 100kHz. What I want to illustrate here is we think high frequency information will positively add to the reality of a recording, but we take for granted that the quality of this ultra sonic sound is as good as anything in the lower frequency bands, and I have my doubts about that. My conclusion, there are more microphones transducing crappy sounds above 16kHz than microphones doing that in a proper way. Organsound in that sense is one of the most challenging to record, and my dog's behaviour has prooven that to me what I already expected.

The sales of recording equipment has become a very trendy market, where decisions are not only made on technical basis, there is a lot of emotion going on, and manufacturers are appealing to that. I highly appreciate that Dan is not following trends and tries to educate his customers a bit, just to help them.

I know that letting a dog decide about the use of microphones would be silly, but when something gets painful for an animal having got the ability from nature to hear about three and a half octaves more than we can, there is something seriously wrong if that dog cannot stand that sound, while with a better microphone it seems to be no problem. Converters going up to a 50kHz of bandwidth do not help to let a microphone sound better, when that same microphone is transducing noise, frequency depending phase shift and distortion above 16kHz. I rather then prefer a roll off, and therefore it is not surprising so many of us like tube microphones with roll offs of more that 15 dB at 20 Khz.  

Erik Sikkema

[Ronny]

No problem, when you record Laika, just use a Sanken mic.

[Graham Jordan]

This illustrates a point that has been made before, but seems not to be mentioned enough, about the high freqs. 'People' talk about possible 'brain pleasure center activity', and other suggested interactions and responses, due to high frequencies that are above the audible range. But if you can't conciously hear them (or consistently perceive them) - how are you possibly going to produce the audio material properly in the first place? Assuming that you're not simply doing a straight record and playback, you're going to be mixing, EQ'ing, and generally processing the signal, but you can never hear the effects on these high frequencies. More disturbingly, even if you don't think these frequencies have an efect, or are useful, at the high sample rate (192kHz), without frequency analysis tools (FFT etc.) you don't know what's going on up there (but your dog clearly has an idea perhaps). With noise shaping, particularly if done on a lot of inputs subsequently mixed, or done repeatedly, it would seem that you can end up with quite an amount of energy up there. What effect does this have on processes that rely on making power level measurements across the entire 'valid' bandwidth (i./e. to fs/2)?

Anybody have any experience of, or know of good information sources, looking at what is practically, and actually, happening in the ultrasonic (>20kHz?) band up to fs/2 for 96kHz and 192kHz? Is it mostly low level noise? Introduced shaped noise? Possible microphone artifacts? What effect do these noise signals have on other processing/metering? Any 'real' HF signals? Clearly these could all vary depending on how particular material was recorded and produced.

[Ronny]

Graham Jordan wrote on Tue, 29 November 2005 13:37

This illustrates a point that has been made before, but seems not to be mentioned enough, about the high freqs. 'People' talk about possible 'brain pleasure center activity', and other suggested interactions and responses, due to high frequencies that are above the audible range. But if you can't conciously hear them (or consistently perceive them) - how are you possibly going to produce the audio material properly in the first place? Assuming that you're not simply doing a straight record and playback, you're going to be mixing, EQ'ing, and generally processing the signal, but you can never hear the effects on these high frequencies. More disturbingly, even if you don't think these frequencies have an efect, or are useful, at the high sample rate (192kHz), without frequency analysis tools (FFT etc.) you don't know what's going on up there (but your dog clearly has an idea perhaps). With noise shaping, particularly if done on a lot of inputs subsequently mixed, or done repeatedly, it would seem that you can end up with quite an amount of energy up there. What effect does this have on processes that rely on making power level measurements across the entire 'valid' bandwidth (i./e. to fs/2)? Anybody have any experience of, or know of good information sources, looking at what is practically, and actually, happening in the ultrasonic (>20kHz?) band up to fs/2 for 96kHz and 192kHz? Is it mostly low level noise? Introduced shaped noise? Possible microphone artifacts? What effect do these noise signals have on other processing/metering? Any 'real' HF signals? Clearly these could all vary depending on how particular material was recorded and produced.

Eric I received your PM, I was just joking about recording Laika, but you aren't creating music for dogs, you are creating music for humans and therefore, what Laika responds to is inconsequential to what your listeners are going to respond to.

In response to the last paragraph above by Graham. Higher order harmonics, cymbals can have over 100 harmonics that can be measured up to 102k, however, what many people leave out of the equation is frequency perception in humans falls way off above 16k. If you play a 2k tone, note it audibly in perceptive gain "to you", than play an octave higher at 4k, again at 8k and again at 16k with the exact same gain on the tone on all freq changes, you'll notice that 2k is the loudest, 4k is slightly lower but not much, 8k will be more noticably lower than 4k and 16k will be way down "perceptively" to humans than the 8k tone and way, way down compared to the 2k tone where many higher freq musical instrument fundamental tones reside. Each individual perceives frequencies at slightly different levels that follows the Fletcher-Munson curve, however the F-M curve is not finite to all humans, the smiley face will vary. With some people the smile will be wider, some more narrow, some the smile will be lower in the middle and with others higher in the middle, it's not exact, but it resembles human acuity in general. If you chart your own F-M curve and find out where your high freq's roll way off, you can equate the same curve to the ear at higher frequencies. The jury is still out on what frequency that the hair cells cease to detect and reproduce, I feel that just like the F-M curve being different that it's also different with different individuals, but let's say that we did have hair cells that could vibrate at 100k. The freq's at 100k due to the F-M curve will have to be boosted in gain far more than most tweeters can take to reproduce that frequency with enough gain for it to be audible. For example if you play the 2k tone and it's x dB in level human perceptive wise, the 16k tone would have to be boosted by +50dB (figuratively speaking) to equal the perception of the 2k freq. Using this logic, by the time that you get to 30k, the gain would have to be boosted by about +80dB to equal the perception of the 2k freq and as you go up in frequency, the human acuity falls ever faster. So in the real world, freq's that are most perceptive to humans that reside between 30 Hz and 8k will highly mask freq's that are 16k and above and the drop off in acuity compounds exponentially as freq's go higher. IOW, in the real world we don't hear cymbal harmonics at 100k, to hear them, we would have to have hair cells that vibrate that fast, which we don't and we'd have to boost the gain on the 100k tone by a couple of hundred dB to equal the 2k freq's without having them masked. Of course this is all theoretical as we can't hear that high, but if we could, due to the F-M curve the higher the frequency you go, the more you have to boost gain to reveal these freq's when they are playing along with the most perceptive freq's between 30Hz and 8k. ITR, freq's above 16k will always be attenuated and masked by the most perceptive freq's. That said, most recording mics max out at 20k, most playback systems don't even reproduce 20k flat. There is a reason that mics have been designed to only go to 20k and playback systems rarely go above 22k and cd's are filtered off above 22k, it's because they capture and reproduce "more than the typical human can perceive". 20k is 4,000 Hz above 16k, now go back and listen to your 16k signal and re-examine any information that you've gotten of the net about freq's above 20k having any significance to the energy below 16k. If we could hear the higher freq's, they would be masked regardless without artificially boosting them outside of the main energy freq's between 30Hz and 16k.

[Graham Jordan]

Graham Jordan wrote on Tue, 29 November 2005 10:37

Anybody have any experience of, or know of good information sources, looking at what is practically, and actually, happening in the ultrasonic (>20kHz?) band up to fs/2 for 96kHz and 192kHz? Is it mostly low level noise? Introduced shaped noise? Possible microphone artifacts? What effect do these noise signals have on other processing/metering? Any 'real' HF signals? Clearly these could all vary depending on how particular material was recorded and produced.

Thanks for that response Ronny. However, that's all stuff I'm aware of, and is not what I was trying to get at. I should have been clearer.

What signals/noise/energy is there >20kHz to fs/2 in someone's digital recording/production system, or final material, running at 96kHz or 192kHz sample rate? i.e. take someone's system, today, with it running at 96/192, throughout the system what is going on in the signals above hearing? What gets from the air vibrations into the digital realm? Did the ADC add to this with shaped noise? Is it then going through processing that adds it's own shaped noise/dither? Any non-linear processing adding signals up here? Any EQ cut/boost up here? What happens when you sum (mix) lots of signals like this together? Is there a situation where someone's set-up is adding/producing significant energy at higher freqs that might cause digital processors that use the power/peak measurements (compressors etc) to misbehave - assuming they look at this freq range? What about if you feed these signals into power amps/speakers - is there a chance that the energy/signals up here could upset them?
For example - what is actually up there on produced 96/192 material? This includes poorly produced material.

I'm wondering what is actually there - since you can't hear it it's not obvious during normal production I assume. Ultrasonic air vibrations are not the only source of signals/power up there (and may not even be the main contributor with 20-20k mics etc. if you then have noise shaping/dither, digital synths?).

Take an extreme case (don't know if anybody would really get this) - something causes -40dBFS (pulled out of the air, could be -60, -6, whatver - it's above the 'flat' noise floor significantly, or above a 'normal'noise shaped dither etc.) signal in ultrasonic range (e.g. digital synth, bad/amplified noise shaping, etc.) - you can't hear it, but maybe it causes problems? Have 96 channels of ADC with noise shaping all the same shape - what level of noise/power do you get up there, and could it cause problems? My thought is yes it can, but I haven't seen anything about this aspect to the high freqs. Hence the question. This is not about hearing - it's about what the devices produce/transmit and how they respond and what this results in in a real recording/prduction/SR environment through the production process to the final result.

Of course, after all that, going to CD with cut out all that HF stuff, but maybe some damage has already been done (e.g. confused processors making bad alterations to the audio - e.g. bad compressors? bad metering etc.), but if you're going to amps/speakers or 96/192 formats that will go to amps/speakers later? Maybe we 'don't care' what's up here - which of course says you maybe should be working at 48kHz.

Note that I am not in the recording/production side of audio - but an EE in the design/manufacture of some recording/producion/mixing equipment. So I am very aware of the fundamental 'science' stuff, it's the full span of real world usage that I'm less familiar with, and also less familiar with every type of processor/mixer/DAW out there (especially the 'bad' stuff) and how people actually use the gear (particularly those who use it incorrectly or non-optimally - but maybe can't hear the difference - I'm sure they exist).

[Schallfeldnebel]

Ronny Morris wrote:"Erik I received your PM, I was just joking about recording Laika, but you aren't creating music for dogs, you are creating music for humans and therefore, what Laika responds to is inconsequential to what your listeners are going to respond to."

You are right. But as you know, most women hear at least a couple of thousend herz more than men in the same age. Last week I showed my wife a newly made test recording, and my dog ran out. I asked my wife if she liked the recording, and she did not like it at all. In her observation the high frequencies were painful to her ears, although I thought it was pretty good for using cardioids on organ music.

Erik Sikkema

[maxdimario]

to me, personally speaking, the timing resolution argument did not totally fall apart. I don't wish to discuss it further because it pivots on arguments that are too difficult to prove.

as far as any real system increasing in accuracy with an increase of samples for a given slice of time, that would apply to any sampling system, once the invariable 'unknown element' is considered part of the equasion... but seemingly that does not apply to the world of audio converters.. oh well..

but my listening habits are quite different than 99.9999% of the population, so I may be focusing in on minutiae.

I do know that absolute precision in the attack transients in the time domain, relative to a fixed timeline (phase shift not included, phase distortion and ringing included) is something I definetly hear and I am so glad that i do because it adds to my listening enjoyment very much, especially with recordings done on low feedback tube equipment.

as for ultrasonic frequencies, my ears get very tired after 20 minutes of 78 records.

78 records are very fast and noisy, and I assume that they have noise that goes up over the 100KHz mark.

a lowpass filter at 12K does make them more tolerable.

even though our brains do not register over 20KHz, the immediate nervous system near the ear seemingly does.. otherwise ears would not get tired with ultrasonics.

[Ronny]

So what you are saying is that even though your ears don't hear over 20k, your mic doesn't capture over 20k, your speakers don't playback over 20k, you can tell that ultrasonic freq's that you can't capture, nor playback are making your ears tired.

How do you know what makes your ears tired, if you can't capture, reproduce or hear what you say are causing it?

[maxdimario]

78's (78 noise) , analog amps and speakers with tweeters go way over 20k.

tests have been performed where exposure to high levels of ultrasound caused ear fatigue and temporary tinnitus if I remember reading correctly.

didn't I read on this forum that rupert neve did a test such as this?

don't mix my two arguments together please.. the last phrase was just to say that: although we don't hear it, the nervous system is AFFECTED by the ultrasonics... how is a different story...