The Audiophoolery Thread.

[Craig Sawyers]

One of the trickiest things to get right with an AB amp is induction distortion. Because the power rails take half cycle currents, they can couple easily into low level circuits. That means that at low frequency the distortion is low - which is why most amp manufacturers selectively quote distortion at 1kHz. However induction distortion is frequency dependent, and leads to anomalously high distortion at upper audio frequencies.

This is not a problem with class A (because the power supply current follows the audio signal) or class D because of its operating principle.

But Halcro and Benchmark have clearly found solutions to that problem; I'd like to see inside Benchmark's case if only to look at their power supply routing!

Craig

[Craig Sawyers]

Originally Posted by stevehertz

Undoubtedly, I am very cynical when it comes to the ways in which many hifi manufacturers and retailers now go about their businesses, making and selling components based on the 'new wave' of hifi ethics and beliefs; audiophoolery. Basically, it's stuff that new people drawn into hifi swallow as being true, when more often than not we're talking about totally nebulous items that capitalise on their lack of knowledge to extract money from them. So the audiophool genre grows and it gathers momentum, and we, the 'feet on the ground' objectifiers are seen as the phools - by them! That's not lost on me, though the difference is, my money is used on stuff that is 'real' using tried and tested techniques (not just hifi) whereas the audiophool is spending thousands of pounds on cables and fanciful upgrade widgets that more often than not do nothing to improve their hifi system. It's all in the mind and it's an atrocious rip off. Yes, I'm cynical!

Well, I hired the guy, I know him well, and he most definitely is not and never has been a snake oil merchant. He is one of the best audio engineers I have met.

You want to rubbish his achievements, well that is up to you. But I'm not going to play that game.

[ajgriff]

Originally Posted by stevehertz

Yes, I'm cynical!

So am I. Distortion might be measurable with sophisticated equipment but can anyone actually hear it?

Alan

[stevehertz]

Originally Posted by Craig Sawyers

One of the trickiest things to get right with an AB amp is induction distortion. Because the power rails take half cycle currents, they can couple easily into low level circuits. That means that at low frequency the distortion is low - which is why most amp manufacturers selectively quote distortion at 1kHz. However induction distortion is frequency dependent, and leads to anomalously high distortion at upper audio frequencies.

This is not a problem with class A (because the power supply current follows the audio signal) or class D because of its operating principle.

But Halcro and Benchmark have clearly found solutions to that problem; I'd like to see inside Benchmark's case if only to look at their power supply routing!

Craig

As an aside to that, subjectively we are able to detect distortion easier in the mid frequencies due to millions of years of us needing to decipher speech from background noise. We are 'tuned in' to mid frequencies. At both higher and lower frequencies that is not the case, and on top of that, most people are unable to hear much difference between (say) cymbal sounds and their harmonics or at the other end of the spectrum, kick drums and low bass guitar notes; they're there and they're doing their job, but people don't notice a small amount of distortion there, not so with a human voice. In recent decades cymbals have even been replaced in dance records by white noise; they're not even cymbals. Though yes, at least to me, that is noticeable!

[stevehertz]

Originally Posted by Craig Sawyers

Well, I hired the guy, I know him well, and he most definitely is not and never has been a snake oil merchant. He is one of the best audio engineers I have met.

You want to rubbish his achievements, well that is up to you. But I'm not going to play that game.

I wasn't specifically rubbishing that guy's achievements or ability, more about procedures and ways of doing business etc.

[Radio Wrangler]

Originally Posted by Craig Sawyers

One of the trickiest things to get right with an AB amp is induction distortion. Because the power rails take half cycle currents, they can couple easily into low level circuits. That means that at low frequency the distortion is low - which is why most amp manufacturers selectively quote distortion at 1kHz. However induction distortion is frequency dependent, and leads to anomalously high distortion at upper audio frequencies.

The main compensation pole for the feedback loop design is usually well below the upper limit of the audible range, so the open loop gain is falling beyond 1kHz, and so the beneficial effects of the feedback - lower output impedance, reduced distortion etc are scaling down with increasing frequency. This also applies to the amplifier's ability to combat unwanted signal ingress within the loop - eg output currents inducing voltages in sensitive input areas.

Inductive coupling is a two-part affair. Careful layout of the high-current source of the unwanted stuff can minimise the effective loop area and hence the field strength. Careful layout of the victim region can minimise its effective loop area or arrange net cancellation, reducing the induced voltage still further.

Like any screening/interference job, it's wise to win whatever advantage you can get wherever the opportunity arises. So don't look only at the power current loops, look also at input signal loops. They're just as important.

David

[Craig Sawyers]

Absolutely agree David. I'm just saying that it is one of the toughest things to eliminate.

I've been looking at the Benchmark specs with a more critical eye - and sure enough they only quote at 1kHz, with either a 20kHz or 80kHz bandwidth. On the main page the do say that 3ppm is met from 20Hz to 20kHz for the following table of output conditions, but don't quote the measurement bandwidth.

And in the manual, the AP plots of distortion vs output level are at an unspecified frequency and measurement bandwidth.

So even they are very carefully stepping around some of the real world issues and being economical with the numbers.

But is it a bargain for USD3000? Indeed it is. Quoting our friends from across the pond, it is a lot of bang for the buck.

Craig

high power amplifier to pull that stunt.

Stunt is right, less than 1% distortion is inaudible by you average listener, 0.01% easily done. Nice bit of engineering though.

[stevehertz]

When did the world of hifi amplification take a major step forwards in being able to produce another octave of bass? Probably back in the 1920s?..

[stevehertz]

Originally Posted by merlinmaxwell

Stunt is right, less than 1% distortion is inaudible by you average listener, 0.01% easily done. Nice bit of engineering though.

But surely, it's not just a question of being able to hear 1% or even 0.01% distortion (whatever), it's the ability of an amp to exhibit clarity in complex passages? That can be affected by distortion levels that may not manifest themselves as 'pure distortion', but make for a subtle muddling of clarity.

[Ted Kendall]

Originally Posted by stevehertz

When did the world of hifi amplification take a major step forwards in being able to produce another octave of bass? Probably back in the 1920s?..

Quite apart from the fact that the blighter can't spell...the word is "eke".

[Radio Wrangler]

Easy to do an amplifier right down to DC = 0Hz

How many octaves down is that? well, theoretically infinity!

Realistically it depends on how long it's been turned on for and the listener's life expectancy.

Suitable speakers might be a bit of a problem.

David

[Ted Kendall]

Originally Posted by stevehertz

But surely, it's not just a question of being able to hear 1% or even 0.01% distortion (whatever), it's the ability of an amp to exhibit clarity in complex passages? That can be affected by distortion levels that may not manifest themselves as 'pure distortion', but make for a subtle muddling of clarity.

THD figures don't directly address intermodulation distortion, which causes most "muddying", and should be well-controlled in any respectable amplifier, but sometimes isn't. Peter Walker once said you could design an amplifier which would pass the routine tests but on which you couldn't recognise a tune. I don't know whether he ever did it to prove the point...

[GrimJosef]

Originally Posted by stevehertz

But surely, it's not just a question of being able to hear 1% or even 0.01% distortion (whatever), it's the ability of an amp to exhibit clarity in complex passages? That can be affected by distortion levels that may not manifest themselves as 'pure distortion', but make for a subtle muddling of clarity.

I'm not clear (no pun intended) what you're getting at here.

Distortion (if you also include noise) is, at its root, any difference between an amp's output signal and its input signal other than the obvious one resulting from the amp's gain. If the distortion is low then, well, the output signal is the input signal, only louder. I can't see how the 'clarity' can be compromised in this case.

When distortion percentage is quoted it commonly refers to harmonic distortion measured with a single-frequency sine input. Sometimes a percentage figure for intermodulation distortion, measured with two simultaneous input frequencies, is also given. It is theoretically possible for an amp to have low levels of these two types of distortion but still to misbehave when it's fed music, with its very broad range of frequencies and occasional rapid transients. But it's rare I think, isn't it ?

Cheers,

GJ

[turretslug]

Originally Posted by Ted Kendall

Quite apart from the fact that the blighter can't spell...the word is "eke".

As one of the fabled effects of heavy infrasonics is involuntary bowel movement, maybe they really did mean "eek"....

[stevehertz]

Originally Posted by Radio Wrangler

Easy to do an amplifier right down to DC = 0Hz

How many octaves down is that? well, theoretically infinity!

Realistically it depends on how long it's been turned on for and the listener's life expectancy.

Suitable speakers might be a bit of a problem.

David

Indeed, it kinda beggars belief that someone should claim that a new amp is able to reproduce another octave lower than other amps. But of course, totally in alignment with modern audiophool writings and musings. Outrageous, nothing short of.

[mole42uk]

I wonder, if all the energy that has been put into this debate were used to amplify a musical signal, just how much louder would be the output?

[mark_in_manc]

Originally Posted by Radio Wrangler

Realistically it depends on how long it's been turned on for and the listener's life expectancy.

David

That made me smile - thanks! (Memories of teaching basics of FFT and windowing functions to reluctant 2nd years).

[Radio Wrangler]

My favourite Fourier amusement is entirely encapsulated in the phrase "Fast bass" it always makes me smile. Trying to explain Fourier transforms and the Q of a resonance to anyone who would use that phrase might make your reluctant 2nd years look easy

David

[Craig Sawyers]

As applied to any half way decent electronics, "fast bass" is of course nonsense.

As applied to loudspeakers, it is much more meaningful. In particular bass reflex and ABR speakers can indeed sound bloated and muddy at the bottom end.

However, even there it is subject to analysis, as Richard Small and Neville Thiele showed in their seminal papers on low frequency speakers in the early 70's.