Lumanauw,
you have to work hard to get results. Hard working does not mean posts like http://www.diyaudio.com/forums/showthread.php?postid=702560#post702560 this might be hard "pen work" probably.
This says and solves absolutely nothing. Design, build, listen, compare, visit acoustical live concerts. Than advise the engineers what to do.
And I would add - everyone has ears and everyone can listen in case he wants to do it. This is no special merit and no special qualification to advise and judge the others. Being an engineer does not automatically equal to be deaf and blind.
you have to work hard to get results. Hard working does not mean posts like http://www.diyaudio.com/forums/showthread.php?postid=702560#post702560 this might be hard "pen work" probably.
This says and solves absolutely nothing. Design, build, listen, compare, visit acoustical live concerts. Than advise the engineers what to do.
And I would add - everyone has ears and everyone can listen in case he wants to do it. This is no special merit and no special qualification to advise and judge the others. Being an engineer does not automatically equal to be deaf and blind.
Above 22khz
It an alltogether different thing if you can hear it!
I have measured the hearing of a great many clients from varied backgrounds. They are all above thier mid 30's. I have yet to find someone that can hear about 18khz. That includes women to!
Mark
It an alltogether different thing if you can hear it!
I have measured the hearing of a great many clients from varied backgrounds. They are all above thier mid 30's. I have yet to find someone that can hear about 18khz. That includes women to!
Mark
Hi, PMA,
I'm sorry if anyone (including you) is offended by that post. Offcourse I don't intend to look down to engineers, infact I respect them (offcourse they are not blind and deaf). I'm an engineer myself on mechanical engineering, and I'm not deaf nor blind😀
I will work harder, thanks to your advice.
It is 1 to 1.000.000 level of difficulty between making a comment (being a commentator) and ones who actually in the field. If you watch soccer game or any game, usually the commentator easily pointed the weakness of every player. Commentator usually said : This player should be doing this.....that one should be doing that......The defence tactic should be like this......The striking tactic should be like that......etc.
Just place that commentator in the real game, maybe he cannot run at all😀
To me a successfull fashion designer is also an engineer. He can make a fashion trend that the customer will follow, his products will be bought so much by the customer.
Do you agree on this? Or an "engineer" is a guy who is good with math and physics?
I'm sorry if anyone (including you) is offended by that post. Offcourse I don't intend to look down to engineers, infact I respect them (offcourse they are not blind and deaf). I'm an engineer myself on mechanical engineering, and I'm not deaf nor blind😀
I will work harder, thanks to your advice.
It is 1 to 1.000.000 level of difficulty between making a comment (being a commentator) and ones who actually in the field. If you watch soccer game or any game, usually the commentator easily pointed the weakness of every player. Commentator usually said : This player should be doing this.....that one should be doing that......The defence tactic should be like this......The striking tactic should be like that......etc.
Just place that commentator in the real game, maybe he cannot run at all😀
To me a successfull fashion designer is also an engineer. He can make a fashion trend that the customer will follow, his products will be bought so much by the customer.
Do you agree on this? Or an "engineer" is a guy who is good with math and physics?
Hi, PMA,
I re-read that post. Yes, it sounds discrediting for engineers. Sorry for this, I don't intended nothing of this kind to any member here.
You're right. That post didn't solve anything. Anyone has the very right to make what he believe is supposed to be. And everyone have right to disagree with something, that's natural life.
I believe God is fair to each person in this world. Each one will be given one thing/one ability that better than others, so he can be proud and happy with his life. Maybe one is given the ability to make money, one is given very bright brain, etc. Each person will be having the same life happines, not depending on he is rich or not. Good with math or not. He must have "something" better than others. I know rich people who cannot sleep at night, stressed. Depended on stress pill. Some jump from high building. If I am to make choice to be rich with that condition, I would not choose that 😀
I'm poor in math, I admit that. I admire those who has excellent ability in math, in analyzing cct. Wish I had the same abilities too😀
About SIM. Why is that everytime I see SIM result (I have a friend who can use SIM), I automaticly wanted to have "as clean as possible" FFT, without knowing what the sound will be like. Trying to modifying the CCT so the FFT is very clean. When no harmonics occur, I'm happy (Eventhough it's just modifying the cct in SIM). This happens automaticly in me. Is this happens to anyone else or just me?
I re-read that post. Yes, it sounds discrediting for engineers. Sorry for this, I don't intended nothing of this kind to any member here.
You're right. That post didn't solve anything. Anyone has the very right to make what he believe is supposed to be. And everyone have right to disagree with something, that's natural life.
I believe God is fair to each person in this world. Each one will be given one thing/one ability that better than others, so he can be proud and happy with his life. Maybe one is given the ability to make money, one is given very bright brain, etc. Each person will be having the same life happines, not depending on he is rich or not. Good with math or not. He must have "something" better than others. I know rich people who cannot sleep at night, stressed. Depended on stress pill. Some jump from high building. If I am to make choice to be rich with that condition, I would not choose that 😀
I'm poor in math, I admit that. I admire those who has excellent ability in math, in analyzing cct. Wish I had the same abilities too😀
About SIM. Why is that everytime I see SIM result (I have a friend who can use SIM), I automaticly wanted to have "as clean as possible" FFT, without knowing what the sound will be like. Trying to modifying the CCT so the FFT is very clean. When no harmonics occur, I'm happy (Eventhough it's just modifying the cct in SIM). This happens automaticly in me. Is this happens to anyone else or just me?
Re: Above 22khz
Hello, Mark
Out of curiosity:
A long time ago I worked with a guy in his mid thirties that could listen higher freq than any other of the peple in the lab (I don't remember how high - but more than anyone else).
He was an asthmatic. Do you know if is there some correlation?
Cheers,
Jorge
mwmkravchenko said:
Hello, Mark
Out of curiosity:
A long time ago I worked with a guy in his mid thirties that could listen higher freq than any other of the peple in the lab (I don't remember how high - but more than anyone else).
He was an asthmatic. Do you know if is there some correlation?
Cheers,
Jorge
Hi PMA,
interesting data, both regarding the FR extension of vinyl itself, and about the spectral content beyond 20 kHz. In your data it really looks quite significant.
Mark,
I see the point of the 20 kHz+ content in the time domain, not the frequency domain. That is, most people can't hear a 20 kHz sine. I can't even really hear 15 kHz! But, we can hear transients that depend on HF content. The digital sampling rate and by extension, the frequency response, will determine the accuracy of transient timing, and the "color" of the transients themselves too of course, if you can hear it. And this from all I know is crucial for music. In fact the timing of the harmonics produced by a single instrument determines a whole lot of its character. This is why I mentioned PRAT ;-)
FWIW in the analog domain (amps etc) the rule of thumb seems to be to have bandwidth at least 2-3 octaves beyond the targeted frequency range. CD clearly doesn't achieve that.
BTW Art Ludwig on his website recounts how he implemented a 12 kHz lowpass (knowing that he could only hear up to 13.5 kHz) - and found that transients "limped out" rather than jumped out. An easy experiment to do - on a PC with headphones, first check how high you can hear. Then, using even a primitive EQ that comes with most soundcards, noth out everything 1/3 or 1/2 octave above your just determined ear-cutoff frequency (no, not van Gogh). And at least for me, removing >16 kHz does sound different, even though I can'
t hear a 16 kHz sine unless insanely loud.
re: engineers, I don't think anyone wants to disparage engineers here... certainly not me either, just for the record. My take on Lumanauws post is that , if we are not so sure about the target specifications (i.e., specific human sensitivity to specific "distortion windows" , to avoid, and relative insensitivity to others, lower priority to avoid), the engineer does not have a clear task at hand to work with. No wonder then that he may just do the best he can to meet certain standard measurement targets, even if those may not correlate well to the human sensitivity.
interesting data, both regarding the FR extension of vinyl itself, and about the spectral content beyond 20 kHz. In your data it really looks quite significant.
Mark,
I see the point of the 20 kHz+ content in the time domain, not the frequency domain. That is, most people can't hear a 20 kHz sine. I can't even really hear 15 kHz! But, we can hear transients that depend on HF content. The digital sampling rate and by extension, the frequency response, will determine the accuracy of transient timing, and the "color" of the transients themselves too of course, if you can hear it. And this from all I know is crucial for music. In fact the timing of the harmonics produced by a single instrument determines a whole lot of its character. This is why I mentioned PRAT ;-)
FWIW in the analog domain (amps etc) the rule of thumb seems to be to have bandwidth at least 2-3 octaves beyond the targeted frequency range. CD clearly doesn't achieve that.
BTW Art Ludwig on his website recounts how he implemented a 12 kHz lowpass (knowing that he could only hear up to 13.5 kHz) - and found that transients "limped out" rather than jumped out. An easy experiment to do - on a PC with headphones, first check how high you can hear. Then, using even a primitive EQ that comes with most soundcards, noth out everything 1/3 or 1/2 octave above your just determined ear-cutoff frequency (no, not van Gogh). And at least for me, removing >16 kHz does sound different, even though I can'
t hear a 16 kHz sine unless insanely loud.
re: engineers, I don't think anyone wants to disparage engineers here... certainly not me either, just for the record. My take on Lumanauws post is that , if we are not so sure about the target specifications (i.e., specific human sensitivity to specific "distortion windows" , to avoid, and relative insensitivity to others, lower priority to avoid), the engineer does not have a clear task at hand to work with. No wonder then that he may just do the best he can to meet certain standard measurement targets, even if those may not correlate well to the human sensitivity.
Hi MBK !
I believe, that for a real test you can't use EQ or lowpassfilters,
they all can't cut off above a specific frequency rather than having
at most ~24db/octave rolloff. This means if you want to get rid off
every frequency above 16khz, these filters would audibly change
also signals ~2octaves below 16khz ! Of course you can hear that.
For a real test you would need a FFT, i think i will try that...
Even when properly cutted, intermodulations in the further reproduction
would make these signals above 16khz again audible.
Mike
I believe, that for a real test you can't use EQ or lowpassfilters,
they all can't cut off above a specific frequency rather than having
at most ~24db/octave rolloff. This means if you want to get rid off
every frequency above 16khz, these filters would audibly change
also signals ~2octaves below 16khz ! Of course you can hear that.
For a real test you would need a FFT, i think i will try that...
Even when properly cutted, intermodulations in the further reproduction
would make these signals above 16khz again audible.
Mike
Hi Mike,
I meant this test more as a quick and dirty assessement... plenty of ifs and but's here. Yet, the phase shift of any filter long before the cutoff frequency is part of the problem, no? As far as I know, you might get around this through oversampling in the D/A process but for all I know, you need a steep cutoff filter during A/D conversion as well, that is, if you sample at 16/44 you are creating the problem even before you enter the digital domain... or in other words, no matter your DAC, you can't retrieve the original phase relationship.
The point being, if you cut off frequencies sharply above the audible band , you may very well create audible effects within the audible band.
But fair enough, one could try to isolate purely the effects of missing, inaudible HF , on transient reproduction ("near HF" that is - not ultrasonics, I've seen studies on that, and they were a bit inconclusive, but the frequencies just bekow 20 kHz which most people can't hear as a pure sine). So that would mean a really steep cutoff just above the last frequency one's ears can hear, with only trivial phase shift and magnitude errors within the clearly audible band, and steep cutoff above. I assume one would do that in the digital domain - not trivial to implement.
If you could produce test files for that purpose (say, "with" and "without" 16-22 kHz, and that on program material containing sharp transients - cymbals, clicks, whatever), that would be very interesting.
I meant this test more as a quick and dirty assessement... plenty of ifs and but's here. Yet, the phase shift of any filter long before the cutoff frequency is part of the problem, no? As far as I know, you might get around this through oversampling in the D/A process but for all I know, you need a steep cutoff filter during A/D conversion as well, that is, if you sample at 16/44 you are creating the problem even before you enter the digital domain... or in other words, no matter your DAC, you can't retrieve the original phase relationship.
The point being, if you cut off frequencies sharply above the audible band , you may very well create audible effects within the audible band.
But fair enough, one could try to isolate purely the effects of missing, inaudible HF , on transient reproduction ("near HF" that is - not ultrasonics, I've seen studies on that, and they were a bit inconclusive, but the frequencies just bekow 20 kHz which most people can't hear as a pure sine). So that would mean a really steep cutoff just above the last frequency one's ears can hear, with only trivial phase shift and magnitude errors within the clearly audible band, and steep cutoff above. I assume one would do that in the digital domain - not trivial to implement.
If you could produce test files for that purpose (say, "with" and "without" 16-22 kHz, and that on program material containing sharp transients - cymbals, clicks, whatever), that would be very interesting.
Audiology tests show our ears are more sensitive to left/right channels phase shift diff rather to absolute phase shift value.MBK said:
anli said:
Anli, yep, i can definitely confirm that !
MBK, AFAIK there is no other filter than using fft to have a complete
precise frequency-cutoff...
I will try to make such material, but, as i mentioned above, material
that contains only frequencies above 16khz will still be audible
because of unavoidable intermodulations. These intermodulations
will also be generated in the speakers because of their immense
nonlinearity.
But checking material with cutted off frequencies above 16khz would
be interesting, at least in theory it should not be audible as the ear
also uses some kind of "fft"... The question is if unlinearity of the ear
will also create intermodulations, you will again be able to hear
frequencies far above 16khz, just not a plain sinewave... 😀
Have to try that, create a signal with 16khz+20khz, in theory
not audible, but might get audible through intermodulations...
To avoid intermodulation in the reproduction i could use left channel
for 16khz and right channel for 20khz ?
Weird world...
Mike
I'm one that positivelly cannot listen to more than 15kHz.
So, when I didn't have access to an analyser, I've tested an amp with 19 & 20K tones, LPF (RC) and a phone.
I could hear the inband beat freqs and used them as a guide to fiddle the amp....
So, when I didn't have access to an analyser, I've tested an amp with 19 & 20K tones, LPF (RC) and a phone.
I could hear the inband beat freqs and used them as a guide to fiddle the amp....
WIerd hearing
Yep humans are crazy when it comes to hearing. I personally can hear up to about 15khz. I have an older friend that can hear ultrasonic cleaners at 40khz. Go figure! He is legitimate in his claims and knows what he is talking about. He worked as a calibration tech at the NRC in Ottawa. Now retired. FOr low end I can hear all the way down to 16 hz. Neat!
As for the point about intermodulation wothin the high frequencies it is real and very valid. Usually when we hear the high end that is annoying that is what is bothering us.
Mark
Yep humans are crazy when it comes to hearing. I personally can hear up to about 15khz. I have an older friend that can hear ultrasonic cleaners at 40khz. Go figure! He is legitimate in his claims and knows what he is talking about. He worked as a calibration tech at the NRC in Ottawa. Now retired. FOr low end I can hear all the way down to 16 hz. Neat!
As for the point about intermodulation wothin the high frequencies it is real and very valid. Usually when we hear the high end that is annoying that is what is bothering us.
Mark
Re: Re: Above 22khz
Yes. There is a correlation.
From what I've been able to tell, there are three main factors that determine your hearing range:
(a) Exposure to harmful SPLs.
(b) Exposure to nicotine and other ototoxic compounds.
(c) Asthma.
The two former negatively impact hearing (one study indicates that exposure to nicotine through passive smoking is the only reason why hearing tends to decay with age in the West).
I personally have a light grade of asthma, and can hear frequencies up to 22KHz reliably. Above that, it varies throughout the day(s).
Other people I know with asthma also tend to have a higher upper cutoff frequency.
Given that Russian experiments show brain activity at up to 250KHz input, I think it would be stupid to dismiss >16KHz information as irrelevant out of hand.
However, we may bear in mind that our phase sensitivity drops with increasing frequency, and that the equal-loudness contour curves sharply upward at the high end. So I'd be more concerned with extension and level correctness (the contour becomes compressed, so smaller differences become more audible, like in the low end) than phase.
I like some of the work done by Steen Duelund (RIP). He has been a strong advocate of dropping the goal of -0dB @ 20KHz in favour of a more controlled high-end roll-off. -3dB at 20KHz is not only acceptable, it is required, if the ambience is to be correct. I cannot recall the number of the ISO specification, but there is one. The Vifa XT25 can be modified to give a -6dB point at 100KHz or so, with a gentle slope. That should cover the high end nicely. 😛
I think the modified XT25 may be available through the Duelund audio company (can't recall its name), run by his close associates.
Jorge said:
Yes. There is a correlation.
From what I've been able to tell, there are three main factors that determine your hearing range:
(a) Exposure to harmful SPLs.
(b) Exposure to nicotine and other ototoxic compounds.
(c) Asthma.
The two former negatively impact hearing (one study indicates that exposure to nicotine through passive smoking is the only reason why hearing tends to decay with age in the West).
I personally have a light grade of asthma, and can hear frequencies up to 22KHz reliably. Above that, it varies throughout the day(s).
Other people I know with asthma also tend to have a higher upper cutoff frequency.
Given that Russian experiments show brain activity at up to 250KHz input, I think it would be stupid to dismiss >16KHz information as irrelevant out of hand.
However, we may bear in mind that our phase sensitivity drops with increasing frequency, and that the equal-loudness contour curves sharply upward at the high end. So I'd be more concerned with extension and level correctness (the contour becomes compressed, so smaller differences become more audible, like in the low end) than phase.
I like some of the work done by Steen Duelund (RIP). He has been a strong advocate of dropping the goal of -0dB @ 20KHz in favour of a more controlled high-end roll-off. -3dB at 20KHz is not only acceptable, it is required, if the ambience is to be correct. I cannot recall the number of the ISO specification, but there is one. The Vifa XT25 can be modified to give a -6dB point at 100KHz or so, with a gentle slope. That should cover the high end nicely. 😛
I think the modified XT25 may be available through the Duelund audio company (can't recall its name), run by his close associates.
Okay, just listened to HF-content...
Here is a simple wave, fs=48khz, left channel is 20khz,right channel is 19khz.
http://www.lf-pro.net/mbittner/TestWave/beep.wav
The pure frequencies are not audible (for me), playing them both
results in a nicely audible 1khz-sinewave... (not subtile)
So much about HF-content not audible ? This seems to negate all
comments like "nobody hears distortions on 20khz signal" ?
If it's true that the ear is very able to "hear" that high frequencies,
keyparameters for amps should be quite different than they used to be ?
Mike
Here is a simple wave, fs=48khz, left channel is 20khz,right channel is 19khz.
http://www.lf-pro.net/mbittner/TestWave/beep.wav
The pure frequencies are not audible (for me), playing them both
results in a nicely audible 1khz-sinewave... (not subtile)
So much about HF-content not audible ? This seems to negate all
comments like "nobody hears distortions on 20khz signal" ?
If it's true that the ear is very able to "hear" that high frequencies,
keyparameters for amps should be quite different than they used to be ?
Mike
I hear nothing, even at fairly high volume. But then again in audiophile terms I am practically deaf.... 🙄
Hmm, that's funny !
I have to check my equipment, maybe my soundcard makes trash ?
When i turn the balance-knob the 1khz fades away when turning
to complete left or right.
How did you playback the wav ?
Mike
I have to check my equipment, maybe my soundcard makes trash ?
When i turn the balance-knob the 1khz fades away when turning
to complete left or right.
How did you playback the wav ?
Mike
Okay, damn it (again... 
)
Plugged the computer via spdif to my hifi, silence !
Even with downmixed to mono, nothing !
(At least the tweeter should have enough distortion...)
So, this is a good testsignal to verify quality of equipment...
Is it believable that some electronics are THAT bad ?
At least found an easy way to check equipment for IMD...
The result: HF-content is inaudible (or should)
And: Porksoda, you are NOT deaf !
And, the amp connected to my computer has more distortion than my tweeters !
(no, not any diy-amp)
sorry,
Mike
)Plugged the computer via spdif to my hifi, silence !
Even with downmixed to mono, nothing !
(At least the tweeter should have enough distortion...)
So, this is a good testsignal to verify quality of equipment...
Is it believable that some electronics are THAT bad ?
At least found an easy way to check equipment for IMD...
The result: HF-content is inaudible (or should)
And: Porksoda, you are NOT deaf !
And, the amp connected to my computer has more distortion than my tweeters !
(no, not any diy-amp)
sorry,
Mike
- Status
- Not open for further replies.