Only thing that can be said is that THD figures correlate very poorly with listener satisfaction.
Low order harmonics are a lot less sonically harmful than high order sonics. Back in the 1950s, Norman Crowhurst proposed a weighting system that would discount the importance of low order harmonics, and emphasize the high order harmonics. His proposal was soundly rejected.
Mooly,
Well, I'll be..... it does sound wonderful. It's possible the darl 'decoupler' is compensating for the non-linearities of the VAS pn junction, your thoughts?
Miles,
Interesting point about the LTP. In my experience unbalancing them is not so bad. Consider the S shaped transfer function of the LTP; it compresses the error signal output with increasing diff input signal because of the increase/decrease of Vbe as more/less current flows. This compression affect high amplitude signals, particularly bass, and it's not good distortion - all odd order.
Unbalancing changes the balance of odd v. even and seems to make it a bit more musical.
Crowhurst was well ahead of his time, and in fact Earl is recovering much of this ground today, and it's still rejected - this time by entrenched marketing interests, who exploit the 'psychology of numerical appraisal', even though it doesn't correlate well, it still gets the consumer in.
Bonsai,
Yes, you are right, but can you draw conclusions; AC is merely fast moving DC, so considering the DC conditions in our slow, human piecemeal way is one way to achieve understanding. Can you comment on why in an LTSpice distortion analysis the even and odd order distortions are always phase shifted by strange amounts? That might help understanding what's going on in the dynamic sense.
The npn darl is certainly an emitter follower, and their Zout is 26/Ie, which is very low, and reducing with increasing current draw. It's a sort of 'anti-inductor' because of reducing Zout with increasing current draw. It's also a cap multiplier, of course. But since it is located on one arm of the LTP, it is subject to varying current, and this would certainly change the sonics.
Cheers,
Hugh
Well, I'll be..... it does sound wonderful. It's possible the darl 'decoupler' is compensating for the non-linearities of the VAS pn junction, your thoughts?
Miles,
Interesting point about the LTP. In my experience unbalancing them is not so bad. Consider the S shaped transfer function of the LTP; it compresses the error signal output with increasing diff input signal because of the increase/decrease of Vbe as more/less current flows. This compression affect high amplitude signals, particularly bass, and it's not good distortion - all odd order.
Unbalancing changes the balance of odd v. even and seems to make it a bit more musical.
Crowhurst was well ahead of his time, and in fact Earl is recovering much of this ground today, and it's still rejected - this time by entrenched marketing interests, who exploit the 'psychology of numerical appraisal', even though it doesn't correlate well, it still gets the consumer in.
Bonsai,
Yes, you are right, but can you draw conclusions; AC is merely fast moving DC, so considering the DC conditions in our slow, human piecemeal way is one way to achieve understanding. Can you comment on why in an LTSpice distortion analysis the even and odd order distortions are always phase shifted by strange amounts? That might help understanding what's going on in the dynamic sense.
The npn darl is certainly an emitter follower, and their Zout is 26/Ie, which is very low, and reducing with increasing current draw. It's a sort of 'anti-inductor' because of reducing Zout with increasing current draw. It's also a cap multiplier, of course. But since it is located on one arm of the LTP, it is subject to varying current, and this would certainly change the sonics.
Cheers,
Hugh
Last edited:
Which leaves me wondering, why limit to complimentary and quasi-complimentary? Why not add straight push-pull to the options?
se
Well my quasi - comp Crown DC300A does not sound better than my thresholds, s500.. for sure .... not even close.
Personally i have never heard of a quasi amp sounding better than a fully comp ... This is very interesting as so many have said otherwise..
Hmmm maybe worth another look ...
Personally i have never heard of a quasi amp sounding better than a fully comp ... This is very interesting as so many have said otherwise..
Hmmm maybe worth another look ...
Last edited:
Before this turns into a train wreck, one should not compare unless otherwise identical circuits are actually compared with same source, speaks and power supply.
Any other way is apples and pears, and thus wholly invalid.
BUT, to offer some technical explanation.
Full complementary will create a much more symmetrical handover at the crossover, with some gm doubling but very closely controlled current exchange between the two sides. Thd will be very low, but with odds around 5-10dB above the evens.
Quasi will create a relatively messy handover, across a wider voltage range, and a markedly different Zout between half cycles due to one device being in common collector and the other in common emitter. This will create much higher thd, but it will be dominated by the even order distortions formed by the markedly asymmetrical circuit conditions.
So, you have to assess the electrical and acoustic differences, which immediately makes this a subjective choice. We know such choices are controversial, with name calling etc.
I wonder if there is a half way point between full complementary and quasi? Steve's straight push pull, like a Class AB tube amp using devices of one gender?
Hugh
Any other way is apples and pears, and thus wholly invalid.
BUT, to offer some technical explanation.
Full complementary will create a much more symmetrical handover at the crossover, with some gm doubling but very closely controlled current exchange between the two sides. Thd will be very low, but with odds around 5-10dB above the evens.
Quasi will create a relatively messy handover, across a wider voltage range, and a markedly different Zout between half cycles due to one device being in common collector and the other in common emitter. This will create much higher thd, but it will be dominated by the even order distortions formed by the markedly asymmetrical circuit conditions.
So, you have to assess the electrical and acoustic differences, which immediately makes this a subjective choice. We know such choices are controversial, with name calling etc.
I wonder if there is a half way point between full complementary and quasi? Steve's straight push pull, like a Class AB tube amp using devices of one gender?
Hugh
Hugh... "My thoughts" Hmmm... well I honestly don't know... could this be the basis of a future project (for me
). Perhaps a quasi lateral FET output. Has anyone simulated it (check my signature lol) to see what the spread of harmonics actually is.
Don't we just
To me "subjectivity" has to be the priority... having bought in the past so called "perfect" amps, and having constructed "blameless" amps, you soon realise that what matters more than anything is whether you actually like listening to the thing. It's no good telling and demonstrating to everyone that the amp is perfect, if you and they know that something else actually sounds far better.
So subjectivity tempered with measurement and good design is the order of things for me.
I have just completed design and prototyping of an amp with a weird asymmetrical output stage, very low parts count, and it very clearly has the 'magic' factor. It actually simulates very well too, returns low distortion, but it breaks all the symmetry rules. Yet this amp gets feet tapping and fingers clicking quicker than anything I have done until now.
I find myself haplessly ignoring the specs in the face of this sort of evidence. What is more, while the technocrats argue the figures and measurements, the punter is forming his own opinions and voting with his pocket.
Is this right? Hell no, but it is the commercial reality. Do you design what people damn well should have (take that!!), or do you build what they like?
I think there is a great deal in ideas of acoustic masking. I have never heard a 'perfect' amp yet which did not sound rather mechanical and unmusical, whatever that means. After all, logic is not strictly valid if the original premise is wrong......
Flame suit on??
Hugh
I find myself haplessly ignoring the specs in the face of this sort of evidence. What is more, while the technocrats argue the figures and measurements, the punter is forming his own opinions and voting with his pocket.
Is this right? Hell no, but it is the commercial reality. Do you design what people damn well should have (take that!!), or do you build what they like?
I think there is a great deal in ideas of acoustic masking. I have never heard a 'perfect' amp yet which did not sound rather mechanical and unmusical, whatever that means. After all, logic is not strictly valid if the original premise is wrong......
Flame suit on??
Hugh
We mostly refer to our past experience and this is a very shaky reference at best.
Who can tell for sure that the system that you heard 10 years ago was more pleasing than the one you are listening to now.
There are songs that I loved when I was younger and now hate for no particular reason regardless of the system that I use to listen to it.
There is music in my collection that I enjoy on one day but do not the next day. We are trying to convince ourselves and others that one thing is better than another purely based on what we think we remembered we felt at the time, but then create it into a referenced standard.
There is absolutely now way that one can claim that quasi sound better or worse than complimentary. They may sound different, but it would be hard to tell whether different is better unless you switch between two virtually identical amplifiers thought the same music chain in the same environment and have some sort of pleasure indicating meter monitoring your brain activity relating to pleasure.
One may as well ask, what is the best sexual experience, quasi or complimentary? Has anyone simulated this and correlated the result with some actual data. This would make for an equally interesting discussion because it is as nebulous as defining what sounds better in equally well designed audio amplifiers.
Besides what is the definition of better, referenced to what. The same experience may be better for one person but worse for another. If you have only eaten poultry in your life, is steak better or different? Even if you had poultry and steak all of your life, is better a preference or a known standard.
I have designed and purchased audio equipment since the late sixties, and I must admit that I cannot say which of all the equipment I had sounded better. For sure they sounded different, and I may have had some preferences at times, but even times changed and so has my preferences.
It would be interesting to see what arguments support quasi and what supports complimentary. I remember a few threads ago that Carlos made a statement that complimentary was cr@p and suddenly there was an outcry that he never heard a properly designed complimentary amp.
Kindest regards
Nico
Who can tell for sure that the system that you heard 10 years ago was more pleasing than the one you are listening to now.
There are songs that I loved when I was younger and now hate for no particular reason regardless of the system that I use to listen to it.
There is music in my collection that I enjoy on one day but do not the next day. We are trying to convince ourselves and others that one thing is better than another purely based on what we think we remembered we felt at the time, but then create it into a referenced standard.
There is absolutely now way that one can claim that quasi sound better or worse than complimentary. They may sound different, but it would be hard to tell whether different is better unless you switch between two virtually identical amplifiers thought the same music chain in the same environment and have some sort of pleasure indicating meter monitoring your brain activity relating to pleasure.
One may as well ask, what is the best sexual experience, quasi or complimentary? Has anyone simulated this and correlated the result with some actual data. This would make for an equally interesting discussion because it is as nebulous as defining what sounds better in equally well designed audio amplifiers.
Besides what is the definition of better, referenced to what. The same experience may be better for one person but worse for another. If you have only eaten poultry in your life, is steak better or different? Even if you had poultry and steak all of your life, is better a preference or a known standard.
I have designed and purchased audio equipment since the late sixties, and I must admit that I cannot say which of all the equipment I had sounded better. For sure they sounded different, and I may have had some preferences at times, but even times changed and so has my preferences.
It would be interesting to see what arguments support quasi and what supports complimentary. I remember a few threads ago that Carlos made a statement that complimentary was cr@p and suddenly there was an outcry that he never heard a properly designed complimentary amp.
Kindest regards
Nico
Last edited:
Nico,
Since the human perceptual experience is so unreliable, we can draw no conclusions from people's likes and dislikes, since they are neither verifiable nor are they consistent.
Heck, people won't even listen to the results of a properly conducted double blind test!
Therefore, we have no choice but to revert to our mathematical, scrupulously measured and constructed models. They tell the truth, no question.
But hang on, where does that leave us? Ignoring what we 'think' we like, and preferring instead the components we know measure well and have all the technically relevant aspects ticked off?
I would suggest that in the face of a well run marketing campaign, we buy on hunches subliminally crafted by others anyway, and just about anything we have learned goes out the window. The same, incidentally, is largely true of the way we choose our mate in life. And we should realise that the copy will be written in such a way that subliminally we will think that we made an 'intelligent, rational choice' anyway.
Not for one moment do I believe that a product is successful if it sells well. Often the opposite is true. But sales do help the bank account, I guess, and that is one measureable hallmark of success. I've just watched the story of JK Rowling and come away with the revelation that she has had a miserable, impoverished life and the great art she has created is as much the result of broken dreams and unrelenting poverty as it is personal talent.
My message is that the specs will always be there, but their correlation with the sound quality is specious at best and people will soon know if it sounds any good and buy or not buy accordingly.
Good to see you in form,
Hugh
Since the human perceptual experience is so unreliable, we can draw no conclusions from people's likes and dislikes, since they are neither verifiable nor are they consistent.
Heck, people won't even listen to the results of a properly conducted double blind test!
Therefore, we have no choice but to revert to our mathematical, scrupulously measured and constructed models. They tell the truth, no question.
But hang on, where does that leave us? Ignoring what we 'think' we like, and preferring instead the components we know measure well and have all the technically relevant aspects ticked off?
I would suggest that in the face of a well run marketing campaign, we buy on hunches subliminally crafted by others anyway, and just about anything we have learned goes out the window. The same, incidentally, is largely true of the way we choose our mate in life. And we should realise that the copy will be written in such a way that subliminally we will think that we made an 'intelligent, rational choice' anyway.
Not for one moment do I believe that a product is successful if it sells well. Often the opposite is true. But sales do help the bank account, I guess, and that is one measureable hallmark of success. I've just watched the story of JK Rowling and come away with the revelation that she has had a miserable, impoverished life and the great art she has created is as much the result of broken dreams and unrelenting poverty as it is personal talent.
My message is that the specs will always be there, but their correlation with the sound quality is specious at best and people will soon know if it sounds any good and buy or not buy accordingly.
Good to see you in form,
Hugh
Last edited:
I really want to do this, otherwise, i'll always keep the doubt
I want to bring a couple of dishes on the table.
First is class A (much less problems for a quasi)
Then, what about push-pull with transistors (again, class A)?