The Sound of Distortion (interesting PowerPoint presentation on the subject)

[tuga]

I have just come across this interesting presentation by a chap called Pete Millet:

The Sound of Distortion
http://www.pmillett.com/file_downloads/ThesoundofDistortion.ppt

One of the things I've learned is that single-ended circuits are asymmetrical so they tend to create more even harmonics and push-pull circuits being symmetrical tend to create more odd harmonics.

He also mentions that THD, unless excessive, provides little indication of what an amplifier will sound like, whilst the spectrum of that distortion, on the other hand, makes a great deal of difference in what an amplifier sounds like. (a good case for graphs vs. round numbers)

A worth read in my opinion.

 

[tuga]

 

[steve 57]

I have just come across this interesting presentation by a chap called Pete Millet:

The Sound of Distortion
http://www.pmillett.com/file_downloads/ThesoundofDistortion.ppt

One of the things I've learned is that single-ended circuits are asymmetrical so they tend to create more even harmonics and push-pull circuits being symmetrical tend to create more odd harmonics.

He also mentions that THD, unless excessive, provides little indication of what an amplifier will sound like, whilst the spectrum of that distortion, on the other hand, makes a great deal of difference in what an amplifier sounds like. (a good case for graphs vs. round numbers)

A worth read in my opinion.

I've read that a few years ago and my phone is not too keen to download it as it could not be safe..
But it's an interesting subject
My take take on amplifier topology is that small amounts of 2nd harmonics are benign, and if you use SE triodes to do that, you may well be keeping the original signal intact. But many don't as the powersupply becomes more important when there's no feed back.
Other topologies change the original signal to some extent. No one wants to admit that, but often all too easy to hear.
But as ever there are some very good and even better amplifiers in all topologies
But whats the best... I'm still searching
Bear in mind there are so many other aspects to good amplifier design.

 

[audio_PHIL_e]

A worth read in my opinion.

Interesting. Most (respected) guitar amplifiers are PP but are claimed to sound the way they do because of the "even order of harmonics" ... which contradicts this spiel. Who is right?

 

[steve 57]

Interesting. Most (respected) guitar amplifiers are PP but are claimed to sound the way they do because of the "even order of harmonics" ... which contradicts this spiel. Who is right?

Most valve guitar amps use single ended stages, and they are often well overdriven creating the distortion.
The output stage is push pull but to my knowledge the push pull stage should not cancel the incoming signal, just the distortion in the output stage as most tradition guitar valve amps don't use feedback.

 

[audio_PHIL_e]

Most valve guitar amps use single ended stages

Are you sure? Some of the lower output Fenders did, but AFAIK most of the bigger ones used PP. Likewise all the Marshalls I've come across. The only SE guitar amps I've got are the Epiphone Valve Junior, the Vox AC4TV and the Harley-Benton GA5 (which is an Epi VJ knock-off).

 

[steve 57]

Yes, its the pre amp & voltage amp stages I'm talking about, the output stage is most often push pull
But I always stand to be corrected !

 

[steve 57]

One of the things I've learned is that single-ended circuits are asymmetrical so they tend to create more even harmonics and push-pull circuits being symmetrical tend to create more odd harmonics.

My take is that musical signals are more than often asymmetrical ie guitars and voices, the overtones add to that asymmetrical waveform.
My understanding is s push pull stage will cancel out the differences making the waveform symmetrical...
Altering the intended signal.
But as always these amps work and measure perfectly fine with symmetrical test signals..
I don't know what other's views are on this subject?

 

[montesquieu]

My take: in theory SE should be the way to go, but in reality it's extremely rare (and when you do manage it, very expensive) to find a single ended amp whose high output impedance/minimal damping factor (pretty much a function of how they work) doesn't put a hefty limit on its performance at frequency extremes and its ability to get the best out of real world, moderately efficient multi-driver speakers.

Using SE amps with high efficiency speakers, by contrast, often raises specific issues relating to things like use of very lightweight cones with limited suspension excursion and output volume/frequency limitations, both associated with signal break-up (especially common with single drivers), interactions with power-sapping crossovers if multi-driver, and problems relating to driver/cabinet characteristics such as the fairly narrow optimal frequency ranges of horn driver-enclosure combinations requiring a large number of drivers with associated crossover complexity.

The results IMO is that, while SE amps often do some very nice things with easy music into moderately high-efficiency speakers, it all tends to fall apart when the going gets complicated. (Unless you spend an absolute shed-load and even then some limitations are hard to get around). Indeed their fans often start to restrict the music they listen to in order not to take things outside the system's comfort zone.

OTOH a really well designed push-pull amp with modest use of feedback and significantly lower output impedance/higher damping factor, while it may not have that designed-in euphony associated with SE even order harmonic distortion, can achieve useful/sensible real world power outputs, properly driving crossovers and controlling speaker cones in a more convincing way than even an expensive SE amp.

For me, having tried literally dozens of amps into my 93db Tannoys of all topologies, a relatively high power PP valve amp is the way to go. Again a good one can be expensive but relatively much cheaper than finding a SE amp that would come close to its capabilities even in comparatively restricted circumstances.

It sometimes seems to me that some people go for SE amps for theological reasons - avoiding splitting the signal MUST be better, right? - without really taking into account the end to end circumstances and opening their ears fully to the down-sides. (Exactly the same applies in the the passive pot vs active preamp debate).

All IMO of course.

 

[steve 57]

All IMO of course.

I can understand all what you've said,
And a pretty good sound can be achieved
But what gets you a better sound ?
Thats my goal.

 

[Nativebon]

You could read quote all that has even been... But does it sound like real music?. If not then it's just all theory which does not relate to the real world. That's my take on audio reproduction these days.

 

[montesquieu]

I can understand all what you've said,
And a pretty good sound can be achieved
But what gets you a better sound ?
Thats my goal.

My conclusion inw what I wrote was 100% clear, for all the faffing around with SE amps (my last pair listed at £11k though thankfully I didn't pay that for them) I came back to PP amps. For me, with 93db dual concentric Tannoys it was actually a no-brainer.

Ah you may ask but what do you mean by a better sound? Only you can answer that for yourself. What I am saying though is that I think some of the attachment to SE amps (and the reluctance in some quarters to let them go in situations where they are not the optimal choice) relates more to theology than to sound.

Oh and as for 'real music', I am a classical musician (in my younger days, professional), current instruments here in the house include a piano, a harpsichord, an 8 course Renaissance lute, 6 course Renaissance vihuela, four course Renaissance Guitar, 11 course baroque lute, four and five string banjos, a ukelele and a coupe of guitars. I don't frankly get the notion that somehow push-pull distorts the signal to the point where it doesn't sound like 'real music', no hifi sounds like real music, it all sounds like it was reproduced, and SE is no different to PP in that regard.

 

[tuga]

I agree with @montesquieu 's post above that low power SE amps require very sensitive (≥100dB) speakers and are best used band-passed in active topologies to avoid impedance interaction issues and high levels of not just harmonic but also intermodulation distortion (the two generally go hand in hand). If distortion levels are not kept low and the combined amp/speaker response is greatly affected then, as with single-driver speakers, the system will dictate which music one listens.

My view is that "euphonic distortion" is not universally pleasing: some like it and some don't.
For those who do, measurements can help shortlist equipment which may sound pleasing but not overcooked, and what amp/speaker combinations might work.

 

[steve 57]

I agree with @montesquieu 's post above that low power SE amps require very sensitive (≥100dB) speakers and are best used band-passed in active topologies to avoid impedance interaction issues and high levels of not just harmonic but also intermodulation distortion (the two generally go hand in hand). If distortion levels are not kept low and the combined amp/speaker response is greatly affected then, as with single-driver speakers, the system will dictate which music one listens.

My view is that "euphonic distortion" is not universally pleasing: some like it and some don't.
For those who do, measurements can help shortlist equipment which may sound pleasing but not overcooked, and what amp/speaker combinations might work.

https://digilander.libero.it/paeng/thd_cancellation_in_push.htmHalf way down this article it starts to get interesting regarding push pull issues.
Please bear in mind the asymmetrical nature of much of recorded music..

 

[tuga]

Please bear in mind the asymmetrical nature of much of recorded music..

What do you mean by that?
A signal is a signal as far as I know.

 

[montesquieu]

https://digilander.libero.it/paeng/thd_cancellation_in_push.htmHalf way down this article it starts to get interesting regarding push pull issues.
Please bear in mind the asymmetrical nature of much of recorded music..

All looks fairly irrelevant to me. Surely what matters is the relative amount of distortion of any sort when compared to the signal level? The specs of my current amp (Radford STA100) are set out in graphical form below. Distortion of any sort is very low (and when Will at Radford Revival restored it, he actually beat the official 1967 specs significantly).

In reality, any speaker on the planet is going to have distortion an order of magnitude higher than the amp specs below - the important thing is not the raw amplifier specs or some theoretical modelling of tiny asymmetries, but the real-world interaction of amp and speaker. Hence aspects like damping factor/output impedance become important.

 

[steve 57]

I don't know what other's views are on this subject?

Thanks all for your input..

A few months ago at my place we had quite a few single ended amps playing
Both valve and hybrid , with and without o/put transformers

The tracks played all sounded much the same through them all.
At the end of the day a valve push pull was put in the system, and the music changed, the bassline changed obviously with different emphasise but the rest changed too.
The ability to hear to the back of the recording demished.
I'm convinced that differences was not about euphoric distortion on the single ended amps, my speakers are over 100dbs so the amps hardly stressed, but there was a more going on than that.
A few years previously a chat with Morgan Jones at one of my diy meets had confirmed my views that there was push pull cancellation going on somewhere.

But where !

 

[montesquieu]

Thanks all for your input..

A few months ago at my place we had quite a few single ended amps playing
Both valve and hybrid , with and without o/put transformers

The tracks played all sounded much the same through them all.
At the end of the day a valve push pull was put in the system, and the music changed, the bassline changed obviously with different emphasise but the rest changed too.
The ability to hear to the back of the recording demished.
I'm convinced that differences was not about euphoric distortion on the single ended amps, my speakers are over 100dbs so the amps hardly stressed, but there was a more going on than that.
A few years previously a chat with Morgan Jones at one of my diy meets had confirmed my views that there was push pull cancellation going on somewhere.

But where !

It's simply not possible to generalise from 'a valve push pull' amp to 'all valve push pull amps'. I've had some pretty shit ones, some so-so ones (that probably should have been better given their pedigree) and some that are sublime and very 3D indeed. There are good and bad amps in all topologies and - most important - good and bad matches for individual speakers.

 

[rabski]

It's simply not possible to generalise from 'a valve push pull' amp to 'all valve push pull amps'. I've had some pretty shit ones, some so-so ones (that probably should have been better given their pedigree) and some that are sublime and very 3D indeed. There are good and bad amps in all topologies and - most important - good and bad matches for individual speakers.

This.

I've had some extremely good single-ended amplifiers here, and some extremely good push-pull ones. I've also had (or at least, borrowed) some utterly dreadful ones. The vast majority lie somewhere in between. They're fine, sound good, have tolerably low distortion, etc. They just somehow miss that sprinkling of magic that is the difference between hifi and music.

In my limited experience with them, Tannoys and single ended are not a perfect match. They can be very, very lovely, but something always seems just slightly off. Yet with (again extremely good) push-pull, they open up that last bit of the window to the music. On paper, the Living Voice that I'm still happily wedded to ought to be similar, if not more so, but they aren't. The opposite seems the case.

Either way, I don't believe it's distortion that's at play here. At least, not in the sense of harmonic distortion.

 

[rabski]

Anyway, it's Sunday night.

Anne is away at a conference in Chester for a couple of days, so I spent most of the day with a soldering iron and drill. I've spent most of the evening going through a range of things from Roger Waters, through Messiaen, Holst and Keith Jarrett and I'm now winding down with a bit of Oscar Peterson. Without, of course, any limits on SPL apart from my own tolerance levels.

My conclusion is that my 845 SET is going nowhere. Doubtless it's not perfect, but it does 'music' so realistically that I can't imagine bettering it within my abilities to either build or spend.