There are references to slew limiting distortions in the 2nd and possibly the 1st edition of the Radiotron Designers Handbook: https://archive.org/details/bitsavers_rcaRadiotr1954_94958503 Its not new or limited to IC's or solid state at all. At the time (1935-1955) the terms were different so you won't see TIM or SID but the description matches.
I had read about that, too. Because no title or theory was proposed, I couldn't reference it. However, concepts, practices, and discoveries often predate the latter articles and individuals of whom became famous for them. It's fascinating to realize this.
Usually by those whose perspective lives within a small box. They don't know what they don't know. And assume they do! 🙂
When I was a teenager, I was fooled by guys like Julian Hirsch and bought into the whole low THD story. After owning an AR Integrated in '72 (which had horrible low level resolution) and a Crown D-150 in '74 (which was powerful, but hard and closed in at the top), I began to understand the Emperor was completely naked. I was introduced to Audio Research gear shortly thereafter that was markedly better - yet measured "worse".
I remain amazed there are those today who still believe within some nebulous qualification that all amplifiers essentially sound the same. There are no qualitative benefits to using more linear active devices, caps, quieter resistors, stiffer power supplies, more EMI/RFI resistant circuits, etc. Who cares about the distortion spectra under dynamic conditions, right?
sounds like you haven't read any of AP's references, Cabot's review articles, want to take the usual "stupid conventional engineers", "THD only" strawmen out for a another run
some engineers here have been following both conventional professional literature and "tweak audiophile" claims for 30+ years, and took in few demos, tried ABX testing, added some reading on psychoacoustics
you should for all your puffery know the real claims - address the mature arguments quit wasting our time on reruns
some engineers here have been following both conventional professional literature and "tweak audiophile" claims for 30+ years, and took in few demos, tried ABX testing, added some reading on psychoacoustics
you should for all your puffery know the real claims - address the mature arguments quit wasting our time on reruns
Last edited:
It may be possible that if someone was taught, or, otherwise introduced at an earlier age to an idea that all electronics sound the same, that they will develope an expectation bias that alters their perception. There may be some research into that, but I don't see the greater value in referencing a link to such paper right now.
What is of value is that some parties do measure various aspects of performance under dynamic conditions. I was reading a very good article on the subject a few days ago. A bright and seemingly young but experienced engineer produced a paper based on his experience in design and discussed a number of distortions that I'd never heard of before, but all checked out. Based on what I read, I do think that it's important for myself to stand back and look at the larger picture like he did; to understand that a THD is quite a bit more complex but still smaller part of the whole sound when all the others are weighed in. In the article, two amplifiers could mearure the same on an FFT and be very different. It was an eye opener to say the least. Then, there were a dozen or more often factors we never see measured- all real behaviors of electronics.
I do think that if two things measure the same in each of the 10-20 types of distortion classes, they will sound the same. But, they have to measure the same in each or at least to below the hearing threshold. I don't believe there are any hidden unmeasurable factors in this context. I think that is quite reasonable, and scientifically supported by the law of conservation of energy. How the sound is read by the brain is outside that scope, because if the ear drums are struct with the same timbre and sound, it is the same sound and in the same timeframe would be interpreted as such, unless the auditory system was acted upon by another factor. This was a great point the author brought up. Being human, there are many factors, but they are after the fact. Quite reasonable all around.
What is of value is that some parties do measure various aspects of performance under dynamic conditions. I was reading a very good article on the subject a few days ago. A bright and seemingly young but experienced engineer produced a paper based on his experience in design and discussed a number of distortions that I'd never heard of before, but all checked out. Based on what I read, I do think that it's important for myself to stand back and look at the larger picture like he did; to understand that a THD is quite a bit more complex but still smaller part of the whole sound when all the others are weighed in. In the article, two amplifiers could mearure the same on an FFT and be very different. It was an eye opener to say the least. Then, there were a dozen or more often factors we never see measured- all real behaviors of electronics.
I do think that if two things measure the same in each of the 10-20 types of distortion classes, they will sound the same. But, they have to measure the same in each or at least to below the hearing threshold. I don't believe there are any hidden unmeasurable factors in this context. I think that is quite reasonable, and scientifically supported by the law of conservation of energy. How the sound is read by the brain is outside that scope, because if the ear drums are struct with the same timbre and sound, it is the same sound and in the same timeframe would be interpreted as such, unless the auditory system was acted upon by another factor. This was a great point the author brought up. Being human, there are many factors, but they are after the fact. Quite reasonable all around.
Last edited:
And yet, there are large numbers of folks who still believe that. Isn't that incredible? Take Siegfried Linkwitz for example. While quite the speaker designer, he has no problem putting a dozen op amps in series into the signal path with his Orion active xover and thinks that generic amps from companies like ATI offer state of the art performance. Imagine that!
Exactly. The usual complement of measurements you find with most any audio review are insufficient to fully characterize the performance envelope of any piece of gear.
Exactly which metrics do you believe are relevant to a component's dynamic behavior? Let's forget sine waves since - I don't know about you - but I never listen to them.
Last edited:
Before you even look at nonlinear distortion you need to confirm that no other mechanism is influencing the sound. Output impedance even a different size or shape for an output network can affect things. input impedance interacting with your source. Grounding anomalies, marginal stability, clipping behavior (if you play loud) and many more can have as much or more real known sonic impact than small amounts of higher order harmonics.
Looking at the distortion spectrum of a SET amp with no feedback may not show as much as the different voltage/frequency response with level compared to a decent solid state amp. Even a small change in response vs. level would swamp everything else and its far from unknown.
I suspect making two amps with identical qualities except for distortion would require an interstage distortion generator inside one. Otherwise many of the qualities I mentioned will force a congruence of linear and non-linear behavior in two amps.
Looking at the distortion spectrum of a SET amp with no feedback may not show as much as the different voltage/frequency response with level compared to a decent solid state amp. Even a small change in response vs. level would swamp everything else and its far from unknown.
I suspect making two amps with identical qualities except for distortion would require an interstage distortion generator inside one. Otherwise many of the qualities I mentioned will force a congruence of linear and non-linear behavior in two amps.
While Linkwitz is a capable designer, there are a hundred or more that can accomplish similar or better performance on a daily basis, and may or may not be recognized for their efforts. As to him using generic or mixed brands of chipsets, there is certainly more than one side to the auditioners: one group may fall under the category that believes if its build by X famous designer or brand, then it must be perfect without exception and cannot be surpassed. That would be a very biased viewpoint, but that is a part of the human element that they can be conditioned.
PS. Any amplifier with a non-zero resistive or impedance emitter, cathode, or source load is by definition using feedback. This includes every triode schematic I've ever seen. Operating the triode without feedback makes every aspect of operation worsen. If there is an emitter resistor to the ground plane, then that amplifier has source feedback because of a non-zero Rs (FETs), Re (BJT) or Rk (Triode). Without feedback, a circuit only produces even-ordered sums and differences of the source signals, as per the quadratic law (for BJT it is also a square law, although exponential, and BJT's actually often produce more distortion with FB by their nature. The misinterpretation of this unique device was what gave FETs, transistors and Mosfets a bad name in audio circles). Adding feedback creates all orders of harmonics and intermodulation stepping down in tiers, in addition to the increased frequency band. However, there is a flip side. By adding feedback, semiconductors actually mimic the harmonic spectrum of tubes without feedback, with the distinction that the upper harmonics in semiconductors are much, much lower in amplitude. Contrary to popular belief, semiconductors do not primarily create odd ordered harmoncs, and the introduction of such on the output is wholly design related. Without feedback, both tubes and semiconductors are largely unusable and hardly listenable by high fidelity standards. It's a great marketing tool, though.
PS. Any amplifier with a non-zero resistive or impedance emitter, cathode, or source load is by definition using feedback. This includes every triode schematic I've ever seen. Operating the triode without feedback makes every aspect of operation worsen. If there is an emitter resistor to the ground plane, then that amplifier has source feedback because of a non-zero Rs (FETs), Re (BJT) or Rk (Triode). Without feedback, a circuit only produces even-ordered sums and differences of the source signals, as per the quadratic law (for BJT it is also a square law, although exponential, and BJT's actually often produce more distortion with FB by their nature. The misinterpretation of this unique device was what gave FETs, transistors and Mosfets a bad name in audio circles). Adding feedback creates all orders of harmonics and intermodulation stepping down in tiers, in addition to the increased frequency band. However, there is a flip side. By adding feedback, semiconductors actually mimic the harmonic spectrum of tubes without feedback, with the distinction that the upper harmonics in semiconductors are much, much lower in amplitude. Contrary to popular belief, semiconductors do not primarily create odd ordered harmoncs, and the introduction of such on the output is wholly design related. Without feedback, both tubes and semiconductors are largely unusable and hardly listenable by high fidelity standards. It's a great marketing tool, though.
Last edited:
Bless you. These are all factors with demonstrated audibility and many of them affect frequency response and noise, to which human ears are extremely sensitive.
Harmonic distortion (no modern engineer simplifies this to THD, they look at spectra) is a very useful design tool. As a consumer spec, not very meaningful unless it's very large, but for the designer, it's absolutely critical to know in order to optimize linearity under various conditions and understand what his circuit is doing. The "lower THD is always better" canard is generally stated by the faith-based audio community as a way to sneer at that which they don't really understand. I've never run across an actual engineer who makes that claim.
Sorry, apparently I didn't make my point clear. Using a dozen op amps (of any type) in series will seriously degrade the audible result. Siggy is one of many who believes that conventional distortion metrics relate to audibility.
If certain criteria happen to be important to you - as opposed to how it sounds.
Now, that's a novel viewpoint.
Yet another novel viewpoint.
in the dominant voltage drive scheme for audio amp/speaker interface we expect low output Z from amps - which really isn't feasible without feedback
it may be local/internal but negative feedback is the only practical method
Triode plate R is a consequence of internal voltage feedback, and of course followers of any flavor have 100% negative voltage feedback
Hi and thanks and you know what ... i agree completely.
The people who likes distortion they have not been exposed to the music without it
I have an experience ... but with drivers
I had a quite good speaker with a dynaudio dome mid ... nice midrange i mean
Then i went to a friend house ... and he showed me a JBL 375 compression driver ... it looked unbelievable
But nothing compared to the listening ... 😱
There was everything ... tone, dynamic, transparency, complete realism
Just one little thing was missing .... distortion .... zero distortion
Well after that experience i think i could live without distortion very very well
But anyway ... it is always nice to talk about distortion
A little here ... a little there ...
Thanks and regards, gino
gino, not so much zero distortion, there's always some there - but subjectively it was inaudible, it didn't get in the way of your listening to the music. Compression drivers make it a lot easier to get there, because they take the load off the amplifier having to work hard - a lot of the disturbing distortion artifacts in conventional hifi sound are due to the amplifiers having to push the accelerator down, and in so doing drawing strong spikes of current from the mains, to get the job done - and the impact of that ripples in subtle ways through the audio system. Hence, sound that lacks at least one of " tone, dynamic, transparency, complete realism ...zero distortion" ... 🙂
Last edited:
Hi and well but ... so we have found a way ... compression drivers
And maybe we should stop worrying about amps ... seriously
This could be a revolution ... a change of prospective
Thanks and regards, gino
That is a very strong absolute claim to make. Siegfried is a very capable and talented engineer. I have spent time with him listening to his stuff and understanding what he is doing. The analog crossover is a special case of multiple amp stages where the devices do not even pass the full audio spectrum. Relating to that fact your claim doesn't carry meaning because there is no pathway through the box that does carry the full audio signal. By definition the audio is massively altered. But going further in that specific case the options for splitting the frequency spectrum are passive crossovers which have serious compromises, active crossover which can be split into discrete circuits, opamp based circuits and DSP based digital circuits. Properly applied any of which can yield better results than any passive crossover.
Meanwhile, Sy said in his article that a series string of 6 opamps were audible to him. Fewer were not audible if I understand what he wrote. Since finding a way to make them not audible was not part of the effort he moved on. However that does not mean that a different approach could not overcome the limitation. Lower noise opamps (the noise will build up by the number of unity gain opamps) possibly and inverting circuits (gets into the absolute phase issue) may well overcome the issues. Even audibility of absolute phase seems to depend on the speaker, and not because a better speaker can elucidate the subtle differences but rather that the one showing the difference is not necessarily symmetrical in its motion. It seems very few transducers are.
Finding controls so you know what your actually testing becomes really challenging. I think that's why you really don't see good work outside of academia unless real money is at risk (like codec development).
I hate to call you on this but I have never seen those "spikes" of current on the mains and I have been looking with the right stuff to see them. The fastest ramp up in current is at turn on and even there its about 16 mS before the current hits max. I have never seen the equivalent or even near on a functioning amp. You can see the power sag a little under load but by that point the SPL is usually so high that other factors enter, like the local constabulary.
More efficient speakers tend to sound better simply due to things like more linear magnet structures and less power required keeping the load out of the area where the amp may be losing supply rails. Compression drivers can be very nonlinear if they are not properly loaded. Even a small mismatch in the throat coupling can get a lot of nonlinearity, not to mention horn resonances, something that humans are highly adapted to recognize and dynamic nonlinearity. Essentially there is no easy and ideal answer.
Yes, yes, yes, and yes, in order. They were configured as unity gain buffers, a different matter than Linkwitz's application, and eliminating the absolute polarity variable. I was a bit coy in the article about what I was able to hear as the difference (it was not distortion), but you seem to have guessed it and qualify for the promised prize. 😀
- Status
- Not open for further replies.