Sorry, but this is a popular misconception and simply wrong. There is absolutely zero difference between quantisation (done properly with dithering) and the noise inherent in the vinyl medium.
Taking a sample, dithering it and converting it to a number with 16 bit accuracy is exactly the same as recording that sample onto an analogue medium with a 96 dB noise floor (note we're talking here just about the quantisation to 16 bit, not the sampling rate). Vinyl has a noise floor of 75 dB. This "digitization error" of which you speak is noise, pure and simple.
There's nothing wrong with push-pull. However, the first solid-state amplifiers, whilst they did have better THD numbers than valve amplifiers, used "quasi-complimentary" output stages using all-NPN output devices, rather than NPN/PNP matched pair. This is because no-one at the time was manufacturing PNP power transistors with the necessary high-frequency performance (fT). Quasi-complimentary output stages have quite horrible cross-over regions. This means that they produce many high-order harmonics in contrast to valve amplifiers which generate mainly second harmonic distortion. Second harmonic distortion generally adds "warmth" to a signal and therefore sounds more pleasant than an objectively "more accurate" amplifier producing high-order harmonics.
If an amplifier distorts a signal with a discontinuity as you state, it must necessarily produce harmonic distortion. The only way of having "anharmonic distortion" is by the amplifier literally generating its own signals. If you've got a link to a proper technical paper on this, I'd love to see it.
Gee, too bad I didn't get to your post before posting mine. I could have shortened it.
I agree with what you've written here, as I'm sure you could tell from my preceeding post.
This section of John Watkinson's book "The Art of Sound Reproduction" seems to disagree with this and shows how anharmonic distortion can arise in low-level sound in digital recording, too. You learn something new every day!
I'm familiar with that book, as is anyone who has worked in audio.
But, what you are missing here is his conclusion, which everyone agrees eliminates quantization error as being relevant.
Quote:
There is little point in studying the adverse effects further as they should be and can be eliminated completely in practical equipment by the user of dither.
(a form of the convolution I mentioned earlier in my other post)
This is all well understood and characterized. Except for the first recordings done in the first year or so, no recordings have been released without proper dithering. It's likely that the first ones that had no dithering were dithered as more copies were made.
Quote:
Here's an article dealing with how anharmonic distortion arises when an amp is challenged with a chord instead of the pure tones most are tested with. Undoubtedly sum and difference frequencies result which are not harmonics of any of the components. (He finds that single-ended amps are worse than push-pull in this aspect.)
This is well known.
Quote:
I got roped into doing something tonight, but I'll keep looking for references. I'm sure I must have read this in Stereo Review or Audiophile many, many years ago, but it stands to reason that a singularity in any function can lead to bizarre, unpredictable results. That's why Stephen Hawking postulated the "Cosmic Censorship Conjecture" to keep them confined inside the event horizons of black holes!
You mean Audio? I don't recall a magazine just called Audiophile.
We'll just have to agree to disagree on a lot of this, but this has been a very interesting discussion.
Quote:
Originally Posted by melgross
We certainly haven't evolved to hear record noise, harmonic and IM distortion, etc. It's very obvious when playing records, and when listen to certain tube equipment.
I think noise is very much a part of the natural environment, and unless you live in an anechoic chamber, any sound is going to vibrate other objects in the environment and create harmonics. So, yes, I think our hearing has evolved to allow for this and to some extent tune it out.
Quote:
Originally Posted by melgross
"Push Pull" was invented in the late '40's by Williamson. It was first used for tube amps, because solid state wasn't around then. In honor, that amp design is called the "Williamson Amp".
Yeah, my Williamson-circuit monoblocs I built 35 years ago sound more musical and pleasant (even if their specs don't bear that out) than any other amp I've owned until the Carver Magnetic Field Amplifier came along. (Of course, it's not really an amplifier, but a switching power supply that tracks the input.)
Quote:
Originally Posted by melgross
It is the first modern amp design. Before that, all amps were of the type called "SLT", or "Single Ended Tube". All amplification was done with a positive going signal. But that kind of amp has many problems, and died out quickly after the "push pull' came out. More recently, it has made a comeback amongst a certain group of audiophiles who don't mind its high distortion, poor frequency response, and very low power. They claim that it has other qualities that more modern tube and SST amps lack.
I freely admit that a lot of self-described audiophiles have some pretty weird ideas, including (probably) me.
Quote:
Originally Posted by melgross
"Brick Wall filters have their problems. The biggest problem was that of "ringing". At the filter point, the frequency response is unstable, and "rings". That is, it varies about the cutoff point. Some people claim to hear it, while others don't. The actual frequency is about 21.5 KHz, which is above the hearing limits of most younger adults (about 23 and up, if they are lucky, if not, 16 and up)).
I think the problem was that the Chebyshev filter they used caused phase distortion way down the audio range where people can hear it. It's my understanding that they chose it over the...Pratt (?) filter, because it caused oscillations (ringing) in the frequency response, which they considered undesirable, while they thought the human ear couldn't distinguish phase distortion. Apparently they were wrong, because the early CD players were roundly panned. It's all handles in the digital domain now, anyway, so that's moot.
Quote:
Originally Posted by melgross
I happen to know Bob Carver from way back. His ideas were well ahead of their time back then, but are still being used today.
That's interesting! I've been following his work since Phase Linear was operating out of the basement of the A&P store in Edmonds. In fact, a buddy and I went up there pretending to be looking for jobs and got a tour of the place. Unfortunately, Bob wasn't there, so I've never met him. If anybody from around here should be a billionaire, it's him and not Bill Gates!
Quote:
Originally Posted by melgross
The biggest problem Bob had with it, was that he was trying to do something in the analog domain that was better done in the digital domain. The problem was that there WAS no digital domain when he was working with this! It was amazing that it worked at all, much less as well as it did.
All I can say is that Sonic Holography works. The stereo imaging when properly placed is unbelievable, and it can make sounds appear to come from beyond the confines of the two speakers. Dolby Surround, which is so common now, addresses the opposite problem of keeping sound from being "pinned" to the closest speaker.
Quote:
Originally Posted by melgross
I'm always happy for someone when they tell me that the equipment they now have is enabling them to enjoy their music more than ever before, even if that equipment make me want to cringe.
This is the bottom line. What I don't enjoy is being told by kids who have never heard good sound that nobody can tell the difference between a quality recording played over a good sound system and 128,000 bps "sound" played through a $5 audio circuit into 89-cent earbuds. Unfortunately, you and I are both dinosaurs, and that's the future. True quality in audio is such a niche market now that it's out of my price range.
You mean Audio? I don't recall a magazine just called Audiophile.
Sorry, I got Audio mixed up with Stereophile, which is a little too far towards the lunatic fringe of the audiophile spectrum for me.
Quote:
Originally Posted by melgross
I'm assuming that last was a joke?
Not at all, a singularity in any mathematical function (like the characteristic curve of an amplifier) can produce totally unexpected results. That's why they don't let you divide by zero.
We'll just have to agree to disagree on a lot of this, but this has been a very interesting discussion.
I think noise is very much a part of the natural environment, and unless you live in an anechoic chamber, any sound is going to vibrate other objects in the environment and create harmonics. So, yes, I think our hearing has evolved to allow for this and to some extent tune it out.
Noise in the environment is very different from the steady state noise we hear in audio equipment, mostly Lp's, cassettes, and tubes.
We evolved to hear natural sounds, not manmade ones.
It's interesting that just earlier today, I was looking through one of my science journals, Science, that I am going to throw out, as I keep them for a month or two to get to finish reading them. This almost missed our discussion.
The issue is Vol. 321 page 189. The article is "Major European Cities Are Quietly Missing Antinoise Deadline"
I think you can tell what it's about.
A quote:
Quote:
People are far more tolerant of sound levels depending on the context and source, researchers noted at the meeting. Relatively loud natural sounds from birds and water, for example, can put people at ease, whereas quieter sources, such as an electrical buzz, cause stress.
I hope that helps.
Quote:
Yeah, my Williamson-circuit monoblocs I built 35 years ago sound more musical and pleasant (even if their specs don't bear that out) than any other amp I've owned until the Carver Magnetic Field Amplifier came along. (Of course, it's not really an amplifier, but a switching power supply that tracks the input.)
There's really no such thing as a power AMP. All that's happening, really, is that the power from the supply is being modulated by the front end of the circuit, and the output tubes, or transistors. They just allow more or less of the power that's already coming in from the supply to be passed to the speakers.
Quote:
I freely admit that a lot of self-described audiophiles have some pretty weird ideas, including (probably) me.
Sadly, that's become part of the audiophile requirement. Mostly on the High End side.
Quote:
I think the problem was that the Chebyshev filter they used caused phase distortion way down the audio range where people can hear it. It's my understanding that they chose it over the...Pratt (?) filter, because it caused oscillations (ringing) in the frequency response, which they considered undesirable, while they thought the human ear couldn't distinguish phase distortion. Apparently they were wrong, because the early CD players were roundly panned. It's all handles in the digital domain now, anyway, so that's moot.
All brick wall filters have ringing, that was the main problem. They went to slower filters that passed through over a frequency range, rather than doing it suddenly. But, then they went to digital filters, and the problem became moot.
Quote:
That's interesting! I've been following his work since Phase Linear was operating out of the basement of the A&P store in Edmonds. In fact, a buddy and I went up there pretending to be looking for jobs and got a tour of the place. Unfortunately, Bob wasn't there, so I've never met him. If anybody from around here should be a billionaire, it's him and not Bill Gates!
Oh, if you ask him, he will happily tell you that he should be a billionaire.
He's been playing around a lot lately. Here's his latest "hobby" stuff:
All I can say is that Sonic Holography works. The stereo imaging when properly placed is unbelievable, and it can make sounds appear to come from beyond the confines of the two speakers. Dolby Surround, which is so common now, addresses the opposite problem of keeping sound from being "pinned" to the closest speaker.
It does work. It would work much better doing it digitally. My friend Ralph has been doing work somewhat smilar to this for a while. It's called Ambiophonics.
This is the bottom line. What I don't enjoy is being told by kids who have never heard good sound that nobody can tell the difference between a quality recording played over a good sound system and 128,000 bps "sound" played through a $5 audio circuit into 89-cent earbuds. Unfortunately, you and I are both dinosaurs, and that's the future. True quality in audio is such a niche market now that it's out of my price range.
Totally agree with you there, but I do have hope! My system is in the low five figure range (though I've poked around with it over the years, of course). I've got Carver Platinums for the front channels, that I've modified. When I called Bob to tell him what I had done, he had a hissy fit for several minutes. Then, as usual, he stopped, and said, Hey! That could work! Too bad they were already discontinued.
Sorry, I got Audio mixed up with Stereophile, which is a little too far towards the lunatic fringe of the audiophile spectrum for me.
Stereophile is a lot of fun. I know most of the guys, and call them up, or send an e-mail, when they really say something overboard.
I do like John's testing though. He's about the only one that really does it anymore, and it's valuable.
Quote:
Not at all, a singularity in any mathematical function (like the characteristic curve of an amplifier) can produce totally unexpected results. That's why they don't let you divide by zero.
I usually think of a singularity the way we do in math, or physics, which is to say, an infinity, an impossibility. It's usually meant, in literature, and some popular semi science books as a "coming together", such as in Ray Kurzweil's The Singularity.
This section of John Watkinson's book "The Art of Sound Reproduction" seems to disagree with this and shows how anharmonic distortion can arise in low-level sound in digital recording, too. You learn something new every day!
Thanks for that link, looks pretty good. However, you need to keep reading to the "dither" section. As I said, quantisation is exactly the same as noise, if done properly with dither. Without dither, the problems in the section you linked to do indeed occur.
Quote:
Originally Posted by Mac-sochist
Here's an article dealing with how anharmonic distortion arises when an amp is challenged with a chord instead of the pure tones
Ah, you were talking about intermodulation distortion. Fortunately a well-designed modern amplifier has extremely low IMD (much lower than a valve amplifier).
Comments
Sorry, but this is a popular misconception and simply wrong. There is absolutely zero difference between quantisation (done properly with dithering) and the noise inherent in the vinyl medium.
Taking a sample, dithering it and converting it to a number with 16 bit accuracy is exactly the same as recording that sample onto an analogue medium with a 96 dB noise floor (note we're talking here just about the quantisation to 16 bit, not the sampling rate). Vinyl has a noise floor of 75 dB. This "digitization error" of which you speak is noise, pure and simple.
There's nothing wrong with push-pull. However, the first solid-state amplifiers, whilst they did have better THD numbers than valve amplifiers, used "quasi-complimentary" output stages using all-NPN output devices, rather than NPN/PNP matched pair. This is because no-one at the time was manufacturing PNP power transistors with the necessary high-frequency performance (fT). Quasi-complimentary output stages have quite horrible cross-over regions. This means that they produce many high-order harmonics in contrast to valve amplifiers which generate mainly second harmonic distortion. Second harmonic distortion generally adds "warmth" to a signal and therefore sounds more pleasant than an objectively "more accurate" amplifier producing high-order harmonics.
If an amplifier distorts a signal with a discontinuity as you state, it must necessarily produce harmonic distortion. The only way of having "anharmonic distortion" is by the amplifier literally generating its own signals. If you've got a link to a proper technical paper on this, I'd love to see it.
Gee, too bad I didn't get to your post before posting mine. I could have shortened it.
I agree with what you've written here, as I'm sure you could tell from my preceeding post.
This section of John Watkinson's book "The Art of Sound Reproduction" seems to disagree with this and shows how anharmonic distortion can arise in low-level sound in digital recording, too. You learn something new every day!
I'm familiar with that book, as is anyone who has worked in audio.
But, what you are missing here is his conclusion, which everyone agrees eliminates quantization error as being relevant.
There is little point in studying the adverse effects further as they should be and can be eliminated completely in practical equipment by the user of dither.
(a form of the convolution I mentioned earlier in my other post)
This is all well understood and characterized. Except for the first recordings done in the first year or so, no recordings have been released without proper dithering. It's likely that the first ones that had no dithering were dithered as more copies were made.
Here's an article dealing with how anharmonic distortion arises when an amp is challenged with a chord instead of the pure tones most are tested with. Undoubtedly sum and difference frequencies result which are not harmonics of any of the components. (He finds that single-ended amps are worse than push-pull in this aspect.)
This is well known.
I got roped into doing something tonight, but I'll keep looking for references. I'm sure I must have read this in Stereo Review or Audiophile many, many years ago, but it stands to reason that a singularity in any function can lead to bizarre, unpredictable results. That's why Stephen Hawking postulated the "Cosmic Censorship Conjecture" to keep them confined inside the event horizons of black holes!
You mean Audio? I don't recall a magazine just called Audiophile.
I'm assuming that last was a joke?
We certainly haven't evolved to hear record noise, harmonic and IM distortion, etc. It's very obvious when playing records, and when listen to certain tube equipment.
I think noise is very much a part of the natural environment, and unless you live in an anechoic chamber, any sound is going to vibrate other objects in the environment and create harmonics. So, yes, I think our hearing has evolved to allow for this and to some extent tune it out.
"Push Pull" was invented in the late '40's by Williamson. It was first used for tube amps, because solid state wasn't around then. In honor, that amp design is called the "Williamson Amp".
Yeah, my Williamson-circuit monoblocs I built 35 years ago sound more musical and pleasant (even if their specs don't bear that out) than any other amp I've owned until the Carver Magnetic Field Amplifier came along. (Of course, it's not really an amplifier, but a switching power supply that tracks the input.)
It is the first modern amp design. Before that, all amps were of the type called "SLT", or "Single Ended Tube". All amplification was done with a positive going signal. But that kind of amp has many problems, and died out quickly after the "push pull' came out. More recently, it has made a comeback amongst a certain group of audiophiles who don't mind its high distortion, poor frequency response, and very low power. They claim that it has other qualities that more modern tube and SST amps lack.
I freely admit that a lot of self-described audiophiles have some pretty weird ideas, including (probably) me.
"Brick Wall filters have their problems. The biggest problem was that of "ringing". At the filter point, the frequency response is unstable, and "rings". That is, it varies about the cutoff point. Some people claim to hear it, while others don't. The actual frequency is about 21.5 KHz, which is above the hearing limits of most younger adults (about 23 and up, if they are lucky, if not, 16 and up)).
I think the problem was that the Chebyshev filter they used caused phase distortion way down the audio range where people can hear it. It's my understanding that they chose it over the...Pratt (?) filter, because it caused oscillations (ringing) in the frequency response, which they considered undesirable, while they thought the human ear couldn't distinguish phase distortion. Apparently they were wrong, because the early CD players were roundly panned. It's all handles in the digital domain now, anyway, so that's moot.
I happen to know Bob Carver from way back. His ideas were well ahead of their time back then, but are still being used today.
That's interesting! I've been following his work since Phase Linear was operating out of the basement of the A&P store in Edmonds. In fact, a buddy and I went up there pretending to be looking for jobs and got a tour of the place. Unfortunately, Bob wasn't there, so I've never met him. If anybody from around here should be a billionaire, it's him and not Bill Gates!
The biggest problem Bob had with it, was that he was trying to do something in the analog domain that was better done in the digital domain. The problem was that there WAS no digital domain when he was working with this! It was amazing that it worked at all, much less as well as it did.
All I can say is that Sonic Holography works. The stereo imaging when properly placed is unbelievable, and it can make sounds appear to come from beyond the confines of the two speakers. Dolby Surround, which is so common now, addresses the opposite problem of keeping sound from being "pinned" to the closest speaker.
I'm always happy for someone when they tell me that the equipment they now have is enabling them to enjoy their music more than ever before, even if that equipment make me want to cringe.
This is the bottom line. What I don't enjoy is being told by kids who have never heard good sound that nobody can tell the difference between a quality recording played over a good sound system and 128,000 bps "sound" played through a $5 audio circuit into 89-cent earbuds. Unfortunately, you and I are both dinosaurs, and that's the future. True quality in audio is such a niche market now that it's out of my price range.
You mean Audio? I don't recall a magazine just called Audiophile.
Sorry, I got Audio mixed up with Stereophile, which is a little too far towards the lunatic fringe of the audiophile spectrum for me.
I'm assuming that last was a joke?
Not at all, a singularity in any mathematical function (like the characteristic curve of an amplifier) can produce totally unexpected results. That's why they don't let you divide by zero.
We'll just have to agree to disagree on a lot of this, but this has been a very interesting discussion.
I think noise is very much a part of the natural environment, and unless you live in an anechoic chamber, any sound is going to vibrate other objects in the environment and create harmonics. So, yes, I think our hearing has evolved to allow for this and to some extent tune it out.
Noise in the environment is very different from the steady state noise we hear in audio equipment, mostly Lp's, cassettes, and tubes.
We evolved to hear natural sounds, not manmade ones.
It's interesting that just earlier today, I was looking through one of my science journals, Science, that I am going to throw out, as I keep them for a month or two to get to finish reading them. This almost missed our discussion.
The issue is Vol. 321 page 189. The article is "Major European Cities Are Quietly Missing Antinoise Deadline"
I think you can tell what it's about.
A quote:
People are far more tolerant of sound levels depending on the context and source, researchers noted at the meeting. Relatively loud natural sounds from birds and water, for example, can put people at ease, whereas quieter sources, such as an electrical buzz, cause stress.
I hope that helps.
Yeah, my Williamson-circuit monoblocs I built 35 years ago sound more musical and pleasant (even if their specs don't bear that out) than any other amp I've owned until the Carver Magnetic Field Amplifier came along. (Of course, it's not really an amplifier, but a switching power supply that tracks the input.)
There's really no such thing as a power AMP. All that's happening, really, is that the power from the supply is being modulated by the front end of the circuit, and the output tubes, or transistors. They just allow more or less of the power that's already coming in from the supply to be passed to the speakers.
I freely admit that a lot of self-described audiophiles have some pretty weird ideas, including (probably) me.
Sadly, that's become part of the audiophile requirement. Mostly on the High End side.
I think the problem was that the Chebyshev filter they used caused phase distortion way down the audio range where people can hear it. It's my understanding that they chose it over the...Pratt (?) filter, because it caused oscillations (ringing) in the frequency response, which they considered undesirable, while they thought the human ear couldn't distinguish phase distortion. Apparently they were wrong, because the early CD players were roundly panned. It's all handles in the digital domain now, anyway, so that's moot.
All brick wall filters have ringing, that was the main problem. They went to slower filters that passed through over a frequency range, rather than doing it suddenly. But, then they went to digital filters, and the problem became moot.
That's interesting! I've been following his work since Phase Linear was operating out of the basement of the A&P store in Edmonds. In fact, a buddy and I went up there pretending to be looking for jobs and got a tour of the place. Unfortunately, Bob wasn't there, so I've never met him. If anybody from around here should be a billionaire, it's him and not Bill Gates!
Oh, if you ask him, he will happily tell you that he should be a billionaire.
He's been playing around a lot lately. Here's his latest "hobby" stuff:
http://cgi.ebay.com/ws/eBayISAPI.dll...m=180275596145
All I can say is that Sonic Holography works. The stereo imaging when properly placed is unbelievable, and it can make sounds appear to come from beyond the confines of the two speakers. Dolby Surround, which is so common now, addresses the opposite problem of keeping sound from being "pinned" to the closest speaker.
It does work. It would work much better doing it digitally. My friend Ralph has been doing work somewhat smilar to this for a while. It's called Ambiophonics.
http://www.ambiophonics.org/
This is the bottom line. What I don't enjoy is being told by kids who have never heard good sound that nobody can tell the difference between a quality recording played over a good sound system and 128,000 bps "sound" played through a $5 audio circuit into 89-cent earbuds. Unfortunately, you and I are both dinosaurs, and that's the future. True quality in audio is such a niche market now that it's out of my price range.
Totally agree with you there, but I do have hope! My system is in the low five figure range (though I've poked around with it over the years, of course). I've got Carver Platinums for the front channels, that I've modified. When I called Bob to tell him what I had done, he had a hissy fit for several minutes. Then, as usual, he stopped, and said, Hey! That could work! Too bad they were already discontinued.
Sorry, I got Audio mixed up with Stereophile, which is a little too far towards the lunatic fringe of the audiophile spectrum for me.
Stereophile is a lot of fun. I know most of the guys, and call them up, or send an e-mail, when they really say something overboard.
I do like John's testing though. He's about the only one that really does it anymore, and it's valuable.
Not at all, a singularity in any mathematical function (like the characteristic curve of an amplifier) can produce totally unexpected results. That's why they don't let you divide by zero.
I usually think of a singularity the way we do in math, or physics, which is to say, an infinity, an impossibility. It's usually meant, in literature, and some popular semi science books as a "coming together", such as in Ray Kurzweil's The Singularity.
This section of John Watkinson's book "The Art of Sound Reproduction" seems to disagree with this and shows how anharmonic distortion can arise in low-level sound in digital recording, too. You learn something new every day!
Thanks for that link, looks pretty good. However, you need to keep reading to the "dither" section. As I said, quantisation is exactly the same as noise, if done properly with dither. Without dither, the problems in the section you linked to do indeed occur.
Here's an article dealing with how anharmonic distortion arises when an amp is challenged with a chord instead of the pure tones
Ah, you were talking about intermodulation distortion. Fortunately a well-designed modern amplifier has extremely low IMD (much lower than a valve amplifier).