Absolutely true, and it is a pity that so far we have no method of measuring amplifier, that would correlate with listening experience. Moreover, I almost lost hope that this method will appear during our lifetime. My usual practice is to regularly clean the contacts of connectors being in the signal chain, as soon as I start to notice a change in the sound (slight loss of vocal emotions). After contacts wiping, the emotions are restored. It creates a reason for sadness: what signal measurements can monitor the status of the contact surfaces coated with gold or rhodium?
And, of course, action of GNFB is recognizable, even with fully-class-A SE solid-state amp. However, I guess, this is not a reason that all buyers of audio equipment would give definite preferance to that or another amp. Discussing of the GNFB artefacts is more a subject for some "aficionados". After all, now I listen to GM70 SE tube amp, and it can serve as a reference for recognizing GNFB artefacts.
Last edited:
I think JC is once again playing with us as he knows full well what the additional tests are that can and should be done...
... especially since he owns IIRC an SR-1. I have however not seen ANY measurements from him with this super duper equipment.
Dust it off JC, and do something with it. Explore for instance multi-tone measurements! 😉
Jan
One overall comment I'd like to make is that 'listening tests' are measurements, and that they can have just as much validity as mechanical and technical measurements, if they're done properly.
My complaint is that listening tests are tedious and error prone, and fail to result in a 'better - worse' ranking in the way that numeric tests can do. So, they're difficult to perform, and they don't help out when one is trying to incrementally optimize a design.
Still, yes, at the end of the day, a listening test is the ultimate arbiter, but they're so hard to do without falling victim to bias and preconception. Further, numeric tests won't always show what 'sounds better', but they can be used to get a better insight into how a circuit works, which parameters matter and which don't, and whether the results of willful effort actually effected a physically noticeable change in the circuit.
Obviously, a combination of both types of tests are needed, but again, my point is that a good listening test really is a measurement, just as valid and reliable as a technical measurement. And, the errors and uncertainty of both types of measurements need to be considered in order to draw any valid conclusions. Sloppy technical measurements are just as invalid as sloppily conducted listening tests, and only a fool follows either blindly.
My complaint is that listening tests are tedious and error prone, and fail to result in a 'better - worse' ranking in the way that numeric tests can do. So, they're difficult to perform, and they don't help out when one is trying to incrementally optimize a design.
Still, yes, at the end of the day, a listening test is the ultimate arbiter, but they're so hard to do without falling victim to bias and preconception. Further, numeric tests won't always show what 'sounds better', but they can be used to get a better insight into how a circuit works, which parameters matter and which don't, and whether the results of willful effort actually effected a physically noticeable change in the circuit.
Obviously, a combination of both types of tests are needed, but again, my point is that a good listening test really is a measurement, just as valid and reliable as a technical measurement. And, the errors and uncertainty of both types of measurements need to be considered in order to draw any valid conclusions. Sloppy technical measurements are just as invalid as sloppily conducted listening tests, and only a fool follows either blindly.
I would be very interested to know what SS/D amplifier does not produce higher ordered harmonics (5th and above)!
I run a recording studio so I deal with this all the time. When I play back recordings I have made on various systems I can tell if they are bright nor not right away. It is helpful to make your own recordings and then produce them so you indeed do have a metric. But just in case you are not aware of this fact, the single most often made complaint against solid state amplifiers in general is that they are bright.
Text in bold: from your comment only, it seems that you are not understanding how the human ear/brain system detects sound pressure. In a nutshell the ear is more sensitive to higher ordered harmonics than most test equipment can winnow out. So there is no 'negligible amount'.
Since we as an industry have not been able to solve this problem, the inevitable result is that amplifiers have tailored distortion characteristics, taking advantage of the ear's masking principle to hide the presence of higher ordered harmonics, or in some cases simply eschewing GNFB altogether to avoid their generation (BTW, there are SS amps that fall in this category).
IOW, you can produce an amplifier with a fair amount of distortion that is less audible than if it produces very little but all of it is higher orders. And don't get me wrong here: I am not a fan of 2nd order harmonics. I'm just being pragmatic.
Well its easy enough to hear 0.005%... the reasons why were covered earlier in this thread. As to the harmonics: the 5th and above. However, in the real world its more complex than that as IM plays a role (interpreted by the ear as brightness and graininess) as do inharmonic distortions.
In a nutshell, as soon as a person thinks they know everything they are also playing the role of a fool. The simple fact is that if we can understand how we perceive sound we can engineer better equipment. Most of our engineering is based on our understanding of sound perception from 50 years ago; we've learned a lot since then but very little is applied.
In case the anti-feedback comment is aimed at me, you are mistaken in thinking that is what I am. I may sound anti-feedback to some and perhaps also anti-SS; both assumptions are quite mistaken. I want it to work is all and I see a lot of poor designs in the marketplace that seem to be accepted as excellent. Is it just that I have higher standards??
Ah! Thank-you for this!! +1 JC and I seem to be on the same page; I have noticed that I have agreed with many of his posts. To expand on JC's comment we have known this for at least 50 years- Norman Crowhurst was writing about this in 1959.
Text in bold: (with the exception of the word 'your') this is a common conflation of personal taste vs human perceptual rules and as such is wildly inaccurate. 'Warmth' in tube amps and brightness and hardness in SS amps are both manifestations of distortion. What I find that most SS advocates seem to misunderstand is how trace amounts of higher ordered harmonics are extremely audible to the human ear (but so far when I encounter that, universally they don't understand much in the way of human hearing perceptual rules either). I don't have any trouble understanding that SS advocates don't like the warmth of tubes. They hear it (human hearing rules) and don't like it (taste).
IOW the warmth of tubes is not something experienced by some and not others. Its not personal to you- I hear it too (and just FWIW, as you alluded its possible to design a tube amplifier that does not exhibit this property; I do it by avoiding the use of single-ended circuitry).
I can understand though that someone might consider a bright amplifier with low distortion and lacking warmth to be more neutral. That is an example of taste. Do you see the distinction?
I see neutral as something else- no manifestation of warmth, no manifestation of brightness either, even though its obvious that the amplifier has speed and extended bandwidth.
flat frequency response over audio then is not "neutral" in your worldview?
can we put your ears in a vault, take them out to calibrate secondary standards every few decades?
show me a study indicating human listening discrimination of distortions below human perceptual noise floor in quiet with appropriate critical band frequency bandwidth - while listening to music that is north of 100 dB SPL to be creating even that ppm distortion for properly sized amp to speaker in room sensitivity
Last edited:
Not sure what your threshold is, but how about the attached?
Just a random example that springs to mind.
Jan
Hi Jan
I wonder what`s hiding beyond 20Khz on that`s NC1200 from bottom three IMD charts ?
well the quoted 0.005% is -86dB from wanted. Plenty of amplifiers out there with the 5th and above well below that level as well as the 2nd 3rd and 4th!
The idea is to look at the IMD product from 18 & 19 kHz that folds down to 1kHz. This is a standard test that is sensitive to distortion at the top end of the audio band.
Anyway, Ralph asked about the 5th, so I pulled up these graphs for him.
Jan
Thanks for your answer Jan !
Anyway personally I still wonder what`s hiding beyond 20Khz ?
Well I was hoping that Ralph would proudly show his amplifier doing much better - after all, if he feels the 5th is so important I would expect his own products to do (much) better. But maybe he's just out on an errand and will report back later.
Jan
Guy`s ,
In the mid-time let`s see & hear what BHK have to say on our Amps & GNFB subject
:
https://www.youtube.com/watch?v=toJMJ2VDMBo
In the mid-time let`s see & hear what BHK have to say on our Amps & GNFB subject
: https://www.youtube.com/watch?v=toJMJ2VDMBo
let`s see & hear what BHK have to say on our Amps & GNFB subject
I see that we share the same type of soldering iron.
😎