Questions re:  GaNfet technology vs other designs.

@theaudioamp I see you dropped the idea that a GaNFET amplifier by virtue of GaNFETs must have load dependent frequency response. That's good- Occam's Razor wasn't supporting that idea at all. 

@atmasphere , I never ever said this. Not even close.

We seem to be seeing two messages from you- that harmonic distortion is and at the same time is not audible.

Also a gross misrepresentation of what I said.  I have been very clear.  Distortion is audible when a high enough level, end of story. I have also been clear that distortion is less detectable with music. I have also been clear that frequency plays into distortion audibility. There is no contradiction in anything I have said.

If  you will set aside the handwaving for the time being, the ear's sensitivity to higher ordered harmonics has been known for a very long time-

And the reasons also known, with perfect clarity for a very long time. Is there a point?

Regarding the effects of insufficient feedback, I refer you to Norman Crowhurst, who wrote about how feedback adds distortion of its own (including IMD) back in the late 1950s.

Do you have a point. You bring out this tired argument, it appears time and time again. No one disputes this is the case. However, you fail to provide evidence to support this is at all an issue. As noted, a quick perusal of Stereophile reviews and ASR shows most solid state amplifiers from the last several decades and certainly today having very low distortion out to 20KHz. You also ignored my very salient point that masking of lower order harmonics coupled with hearing degradation and the slope of Fletcher Munson curves means out past 4KHz, if not lower, higher order harmonic distortion becomes less of an issue.

w.r.t. Bruno, I will link his article here:  https://linearaudio.net/sites/linearaudio.net/files/volume1bp.pdf  which is a good dissertation on distortion and why negative feedback is good, which I am certainly not disputing, and which most SS amplifier vendors use enough of (as noted in measurements by Stereophile and ASR). Bruno does not talk a lot about audibility of distortion in this article. Mainly hand waving. He did make an off the cuff comment comparing linear amplifiers to Class-D and typical input stages off linear amps. This is why as engineers off the cuff statements should be avoided as most Purifi amps are front ended by just those same type of circuits buried in op-amps with enough feedback not to be an issue.

You know where Bruno spends a lot of time talking about distortion? Speaker drivers. it is also where he has stated he will spend his time, on speakers, specifically complete amp/speaker combinations. He acknowledges this is where most distortion lies.

I saw stated, and please correct me if I am wrong, that your OTL have output impedance in the 2-4 ohm range?  Unless you have a speaker with a near flat impedance over frequency, it is impossible for that to sound neutral compared to a high damping factor amplifier. It is simply impossible. With typical dynamics speakers at 8 ohms and less, that is going to make a very noticeable change in the frequency response. I personally do not consider that neutral.

Distortion is audible when a high enough level, end of story. I have also been clear that distortion is less detectable with music. I have also been clear that frequency plays into distortion audibility. There is no contradiction in anything I have said.

@theaudioamp Hm. On that basis we’re on the same page. Since you seem to think we are not, I’m disposed to think that you’ve not stated your position all that clearly in the past. There is also the issue of what is meant by ’when a high enough level’. IME that means well over 105dB down. Most amps can’t do that. The bit that I think gets ignored here is the fact that the ear uses the higher ordered harmonics to sense sound pressure, and due to a +130dB range is far more sensitive to them than most people think.

Bruno Putzeys points out that we’ve been hearing the brightness of solid state since its inception. So if we go back about 20 years or so its safe to say that at that point no amp prior employed enough feedback as that was the basis of his statement; in fact largely the reason he wrote the article!

No one disputes this is the case. However, you fail to provide evidence to support this is at all an issue.

😁 If that were the case Crowhurst might have spent more time at golf or something. Do you see how this statement contradicts itself?

You also ignored my very salient point that masking of lower order harmonics coupled with hearing degradation and the slope of Fletcher Munson curves means out past 4KHz, if not lower, higher order harmonic distortion becomes less of an issue.

You can’t be serious! IME/IMO you got it backwards. Making sure the harmonic content is correct in the Fetcher Muson range is crucial. Not only do you have a lot of higher ordered harmonics in that range but its also the region at which the ear is most sensitive!

w.r.t. Bruno, I will link his article here: https://linearaudio.net/sites/linearaudio.net/files/volume1bp.pdf which is a good dissertation on distortion and why negative feedback is good, which I am certainly not disputing, and which most SS amplifier vendors use enough of (as noted in measurements by Stereophile and ASR).

OMG... Bruno is essentially pointing out that most of those amps don’t have enough feedback, and shows just exactly why. Did you not read the article??

This is why as engineers off the cuff statements should be avoided as most Purifi amps are front ended by just those same type of circuits buried in op-amps with enough feedback not to be an issue.

’just those same type of circuits buried in op-amps with enough feedback’ as you put it does not seem to make sense. With any of his modules you simply need an opamp circuit that can drive the input of the comparitor; you don’t need that ’same type of circuits’ thing (which seems to suggest more than just opamps) you came up with. Its hard to interpret what what you meant there; FWIW Bruno spent a bit of energy in that article explaining why a certain capacitor was used in conjunction with degenerative feedback in the differential input amplifier of conventional amplifiers, none of which would be used in an input buffer to one of his modules. That’s the only interpretation I can come up with for your remark; if I misinterpreted that I apologize. Otherwise I agree that if a purely opamp circuit is used as a buffer, the feedback on them renders them entirely neutral. But we are not talking about Bruno’s amps, FWIW...

Unless you have a speaker with a near flat impedance over frequency, it is impossible for that to sound neutral compared to a high damping factor amplifier.

This statement is simply false- we’ve compared exactly that side by side. I concede that you do have to be careful about the speaker that is being used. But it does not have to have a flat impedance curve! As I pointed out in the article to which I previously linked, the speaker needs to fall into the Power Paradigm of design rules (an example is a speaker designed to work with SETs, which have a similar output impedance to our OTLs).

@theaudioamp Hm. On that basis we’re on the same page. Since you seem to think we are not, I’m disposed to think that you’ve not stated your position all that clearly in the past. There is also the issue of what is meant by ’when a high enough level’. IME that means well over 105dB down. Most amps can’t do that. The bit that I think gets ignored here is the fact that the ear uses the higher ordered harmonics to sense sound pressure, and due to a +130dB range is far more sensitive to them than most people think.

Now you are just making things up. This is as technically ludicrous as the CD demagnetizer someone reminded me of.

Let us unpack this fully. You are claiming that harmonic distortion 105 db below the fundamental is audible. Let us take an example of a 90db/watt efficient speaker, with 200 w/channel, what most would consider pretty loud. That is 90db/watt at the speaker. Let’s say 8 feet distance in a typical room.

• 6 db gain for 2 speakers
• 8db loss for listening distance
• 3db gain from reflections

Let’s ball park that at 110db peak at the listener and 113 db at the speaker. Sure you can get more efficient speakers and more wattage (be careful of real efficiency versus in-room), but most people would consider 200W/chan with 90db/watt as a setup loud.

Most people’s listening rooms that are not dedicated are probably 40db background, with some lucky to be 35db. Dedicated rooms you may get down to 25db with reasonable sound deadening. Keep in mind simply breathing is 10-20db. Now I realize this is broadband and the spectral energy is spread out and we can detect tones below the noise floor, but it still provides perspective. 105db on top of 25 is 130db (and we only have 110 at our disposal). Realistically you can probably hear down to a 0db tone around 3Khz in a really quiet room. That is a tone, in a really quiet room. I have to have at least a 105db primary tone, at say 300Hz, and that is going to mask a 0db tone at 3KHz.

But let’s go back to where we started. I have a primary, that is limited by my speakers to 110db, at the listening position, but you are claiming that a 5db harmonic is buried in there, that at 110db listening level is going to be heard. Really now? Do you want to walk that claim back a little? A harmonic 105db down is 0.0005% distortion from that harmonic. And you think this is audible in music? I am curious what playback device (speaker or headphone) you managed to do a test where you showed 0.0005% added distortion of a high level harmonics, in music, was audible, when the peak level was 110db to start. Why do I ask that? Because I want to buy those speakers or headphones as they are probably the lowest distorting ever made for home use. Even horn speakers at those levels have audible distortion in controlled circumstances, not to mention the masking and IM distortion induced by real music.

I would not expect anyone to take my word for this. Simply create a file with a single tone at any frequency you want, probably 200-300 Hz is best, then add in another tone 105db down anywhere you want and see if you can hear it. I would suggest about 2-3Khz.

Unless you have a speaker with a near flat impedance over frequency, it is impossible for that to sound neutral compared to a high damping factor amplifier.

This statement is simply false- we’ve compared exactly that side by side. I concede that you do have to be careful about the speaker that is being used. But it does not have to have a flat impedance curve! As I pointed out in the article to which I previously linked, the speaker needs to fall into the Power Paradigm of design rules (an example is a speaker designed to work with SETs, which have a similar output impedance to our OTLs).

My statement is absolutely true. Yours is not. I said it is impossible for your OTL amplifier to sound neutral compared to a high damping factor amplifier with any speaker that does not have a flat impedance curve.

Whether the speaker is designed to sound best (flattest) with a constant power, constant voltage of constant current does not make my statement incorrect.  A speaker designed for constant power delivery (uncommon these days) may sound better with your amplifier, but it will not sound the same as a high damping factor amplifier with that speaker. The two will sound much different and have a much different system level frequency response.  You will note I qualified with dynamic speakers of 8 ohms or less. A high impedance speaker will be less impacted by the output impedance of the amplifier obviously.

At a basic level, with your high output impedance OTL amp and a high damping factor amp, with any speaker with a typical varying input impedance, no matter the design rules, it will become almost impossible to level match such that you can even do a proper A/B comparison as the frequency response of the system will be so different.  We are dealing with basic electrical parameters here that cannot be simply had waved away.

My statement is absolutely true. Yours is not. I said it is impossible for your OTL amplifier to sound neutral compared to a high damping factor amplifier with any speaker that does not have a flat impedance curve.

And the fact of the matter that given the right speaker which may or may not have a flat impedance curve the OTLs can sound perfectly neutral (and to be clear those speakers are the exception rather than the rule). Just FWIW, I have the advantage of you on this point- we make an amplifier that has an output impedance that's difficult to measure and we can compare it to our OTLs and do so on a daily basis. Clearly you didn't read the article at the link I posted. Maybe you should. Or not... Tempest in a teapot, that sort of thing 😁

Let us unpack this fully. You are claiming that harmonic distortion 105 db below the fundamental is audible. Let us take an example of a 90db/watt efficient speaker, with 200 w/channel, what most would consider pretty loud. That is 90db/watt at the speaker. Let’s say 8 feet distance in a typical room.

Ever hear a system that sounds loud? If you work to really get rid of those pesky higher orders, you'll find that even at high sound pressure levels on a meter that the system no longer sounds 'loud'.

Put another way the sign of a good system is that it does not sound loud at any volume- it will always have a relaxed character. What you are forgetting is the ear converts distortion to tonality and it also perceived higher orders as loudness. This is why SETs can sound so 'dynamic' and its the source of brightness in most transistor amps. Easy to demo, FWIW... and is also why the right measurements on paper have a directly line to our listening experience (and likely why most of the time they are either not made or not published; I think the industry really wouldn't like people knowing what equipment sounds like by looking at the measurements...)

Its funny how so many people think the ear stops using its perceptual rules beyond an arbitrary point on a bit of paper.