John Curl's Blowtorch preamplifier part II

[Juergen Knoop]

the CCIR/ITU people use a interesting ear weighting filter for noise measurement.
It shows rising sensivity with frequency and a maximum at 7kHz.
It does not need much creativity to apply the principle at distortion artefacts.
In fact there was a danish diploma thesis circuling on the net in which the authors tried it.
Afair they extracted the artefacts from the test signal and passed them through a ear weighting filter.
They claimed those numbers correlated with listening tests far much better, than simple THD specs alone.
Regards

[janneman]

Originally Posted by Joshua_G

The explanation is by psychoacoustics, not be engineering.

I would think so too. Juergen, is there actually an engineering reason?


jd

[Juergen Knoop]

Originally Posted by janneman

I would think so too. Juergen, is there actually an engineering reason?

I would rather call it physiological than psychoacoustical, but I confess my knowledge in this area is most poor.
Anyway, giving attention to human characteristics is most important for engineering decisions. From the beginning of telephone until now.
Regards

[SY]

hen the brain is sensitive to frequencies above 20 KHz, it doesn't matter to the brain whether those frequencies are produced by the musical instrument as harmonics, or whether they are distortion produced in the amp.

The paper doesn't say that. YOU say that. So when you said, "...scientific experiments show that the human ear is sensitive to distortion in frequencies far above 20 KHz, though not to fundamentals," you were just making things up? Or can you cite any scientific experiments which support that assertion?

[anatech]

Hi Joshua,

This isn’t accurate, I referred to limited budget on recording gear, not measuring equipment.

Please allow me to apologize to you then. What you told me earlier in the mists of time seemed to leave me with this impression.

So, okay. What instruments did / do you have access to on a regular basis?

Anyhow, what measurements, by themselves, do you think will be able to tell the sound quality of any amp?

Well, no. Of course not, and I have been very careful to always say that successful designs depend both on measurements and listening tests. On this I have been very consistent on. The fact you are asking means that you are either attempting to take an extreme view, or you've not read my posts in the past. Not that it matters from where I sit.

Now, to answer your question. Measurements can reveal problems that are audible. As you become more adept at measuring and knowing what to look at, measurements can often tell you if an amplifier has a chance of sounding good. You need experience to see this.

Without actual examples, this is meaningless.

in response to ...

Certainly, more well equipped technicians would have the ability to "see more".

Sorry Joshua, this is true and is used by most test and measurement companies in workshops to illustrate why you need certain minimum standards to do a particular job. For instance, you need about 5X the bandwidth of the signal of interest to view it with any confidence using a digital oscilloscope(to determine if you are looking at a square wave). Want proof? Go look at Agilent, Tektronix or any other T&M site. These statements are also enshrined in print.

and also this statement ...

Moreover, some people can look at a schematic, see the construction and have a pretty good idea what an amp may sound like. Not every time, but certainly noticeable and better than guessing.

You're now arguing a point that illustrates a lack of experience on your part. Why not do some research? I am pretty sure that John Curl can do this, Doug Self, Walt Jung and many others can probably do this also. If you can't, then you should begin listening to amplifiers while examining the schematics for starters.
Just because you can't do something has no bearing at all one anyone else's ability to do that same thing - within the realm of possibility of course!

Those numbers refer to best case listening environment.

Okay, you just threw in a qualifier that should have been included before. However, even using that qualifier, some signals will show THD much more easily than others. You've nailed the numbers down too tightly and there are far more variables at play than you seem to accept. Oh well, doesn't matter anyway. On one hand you want to say you can hear more than what can be measured, and yet you still want to claim that higher values of distortion that are easily measured (even with old tube gear!) have no sonic bearing. You really can not successfully argue a point from both sides Joshua. Pick a stance and stay there!

The 7th harmonics is unique, for it doesn't appear in nature. No natural musical instrument or sound producing device has 7th harmonics. Experience and experiments show that the human ear finds 7th harmonic extremely objectionable.

Huh? Is this based on the fact that it's a prime number? Perhaps we should say that no prime numbers are possible to appear in a harmonic series? I'll be kind and exclude 2, 3 and 5 'cause I'm nice.
You are not correct. Do some research on your own please.
The real facts would run something like this:
Higher order harmonics tend to sound less pleasant than lower order harmonics. The 7th harmonic is included in this group.
All of these higher harmonics are generally at low levels unless you are dealing with a poorly designed amplifier. Transducers are a whole 'nother ballgame.

Also, scientific experiments show that the human ear is sensitive to distortion in frequencies far above 20 KHz, though not to fundamentals.

Please tell me that you thought about this before you typed it in. I mean, you just have to be kidding us on this Joshua.
If you can hear something, it is audible (unless it's voices from your ancestors). There is no way that adding a lower frequency tone will make a tone above your hearing any more audible. In fact, the lower tone will tend to help MASK the higher tone you claim can be heard. This is true even if you really mean to say that the tone is sensed, rather than heard. The lower tone will tend to distract you if this is even remotely true.

Again, I am attempting to reconcile your claim that we can't hear harmonics that well. Remember that we are also talking about harmonics that are at least 0.002 X of the fundamental here. Example. You say that we can not hear a 26 KHz fundamental at ... say 0.5 acoustic watts, and yet we could if we had a 3.25 KHz tone at 0.5 acoustic watts along with 6.5 KHz, 13 KHz and 26 KHz (at a level of 0.001 acoustic watts). It's not the 7th harmonic, but I'm giving it a much higher likely power level than it would have at the 7th harmonic. If you really need to see the 7th harmonic, work it out for yourself. Then, try this in the studio or where ever you wish. You will find that the other harmonics will effectively mask the 7th (and probably 6th and 5th also). Even just a mix of 2nd and 3rd will be unpleasant enough sounding.

Not at all. Lower THD figures are easily achieved with NFB and GNFB. When large amount of NFB is applied without proper consideration of the appropriate Slew Rate, measurements of THD will be excellent, while the sound will be bad.

Wrong. The use of high feedback around a poor design will create horrid amounts of higher order distortion. More than a good design will. Besides, right now you are specifying an amp using high feedback when your original statement did not include any such limits. This is equivalent to changing the rules as you play.

BTW, slew rate limiting sounds really bad, and it certainly does show up in THD tests, but only if you are using quality test equipment.

Since when measurements affects tube amps differently than SS ones?

Since you can't apply near the feedback of a solid state amp to a tube amp using an output transformer, that's why. Also, the harmonic structure of a tube amp is different than a solid state amp. In many cases, a higher THD number with a tube amp is less objectionable than the same amount of THD in a solid state amplifier. All as a result of the harmonic structure differences between those technologies. This also means they sound different and people can tell between the two sometimes. Hey! That's true! Wow!

The bottom line is that the measurements most relevant to sound quality are never published. Hence, it's impossible to tell the sound quality of an amp by its' published measurements alone.

Actually, I'm going to bet that some companies do publish more. I have seen this over the years, but I'm not about to waste my time digging this stuff up. I have seen expanded specifications from time to time, even full graphs of THD vs frequency and the odd THD spectrum.

I don’t agree about published specifications being able to tell the sound quality of an amp. Please give examples to demonstrate your claim.

Okay, we disagree. That I can live with. However, coming from someone who can hear differences between cables, it would seem that you have not expended any effort in reading specs and listening to amplifiers. Given your interest and claims, I find this a curious omission in your life to date.
Also, please recognize that I have stated that some specs can give you an idea what some amps might sound like (some characteristics, not complete knowledge) and you are extending my statement to infer I'm suggesting that you can tell what (as in exactly) what an amplifier sounds like. I've been looking at specs and listening to amplifiers for over 30 years. I was a curious kid, what can I say?

There are wrong assumptions on your part. When I (and some others I know) compare pieces of audio gear, be it amps, or cables, or anything, we do it with double blind tests, so that brand and price are unknown as long as the test goes on.

Well, there you go again. Adding hidden information later on in time. Your answers can only be as accurate as the basis of your question and surrounding conditions when you ask. I've not made wrong assumptions. You have failed to disclose that your population of responders was both tiny and involved in a controlled experiment. That really changes the conditions to consider in order to answer your question / statement.

Since we don't know the experience and / or knowledge in your test group, no other things can be said. What has been discovered is that the better (more accurate) the equipment becomes, the more people prefer the better equipment. Also, not surprisingly, the better equipment gets, the more alike it begins to sound (by definition, actually). So your choice of equipment will affect the outcome of the data. As defects become more apparent, people tend to split into groups of what defect they can cope with more easily. This is more easily seen when testing loudspeakers.

As for integration, you are right. However, it's my experience that certain loudspeakers and certain amps sound much better than others in any audio setup.

Yes, there we completely agree.
Some loudspeaker loads can make an amplifier "less happy" (unstable) at certain frequencies and at certain loads. The phase angle of the load can really play havoc with the stability of an amplifier. It may even be as simple as the over current protection operating too early (for the music, it may be holding the amp together).

It’s impossible for me to listen to all well made equipment, in any price range.

Well, my point was that you should be listening to all good equipment you have access to. In North America, many wonderful amps from Europe are simply not available either. It's a five channel world over here.

That's not applicable to those designers who maintain that only measurements are relevant, while "refined listening" or "golden ears" are only marketing snakes oil.

I'm confused again. Where did this come in?
You know, a designer may believe only measurements matter, but if marketing hears the amp and doesn't like it, the amp gets redesigned and they listen again. In the end, the result is the same, but the designer is perhaps less happy.

Did I say anywhere I follow blindly the manufacturer's recommended load?

Actually, no. But then using the transformer as you are, who really knows what the loading actually is?

-Chris

[SY]

Originally Posted by Joshua_G

I don't trust commercial reviews, that is, reviews by magazines, or by professional reviewers. Moreover, it takes reading between the lines to note shortcomings not mentioned by the reviewer.

Please tell me what you read between the lines of this review.

"Here's one more thing some elder audiophiles won't want to hear: The Yamamoto A-08 was fun to have, fun to use, fun to look at, fun to swap tubes in and out of, and, most of all, fun to listen to. I'm quite sure the rest of my system thought it was fun to have around, too.

A great amp: The Yamamoto A-08 is a fine alternative to dense, boring, unmusical high-end audio products. It's a handcrafted consumer-electronics product—an animal that most people thought was extinct—yet it sells for less than the price of many comparable mass-produced products. In short, it's a howling bargain. The Yamamoto A-08 stands alongside the Lamms, the Fis, the Wavacs, the Komuros, the Wavelengths, and other handcrafted amplifiers in its aspiration to transmit the soul of music by embodying some of the soul of its maker. That's all there is to it, really: It has soul."

[SY]

Holy cow, Chris! I'm supposed to READ all that????

[anatech]

Hi SY,
No. Please don't.

I didn't even want to write it, but Joshua expected responses and it would have been impolite to ignore him. But it would have been easier for me.



-Chris

[anatech]

Hi Joshua,
I think you are missing the forest for the trees. Focusing on one issue hides it's relationship with the entire "production". This is also a common problem with some designers, where they focus on one aspect of performance to the detriment of all other aspects.

Remember. The 7th harmonic is accompanied by many others, then add IM distortion and perhaps some cutting edge worries. Every issue has a threshold where it becomes a concern. So until the other issues are dealt with, worry less about the 7th.

As a note, I tend to look at the spectrum up to the 10th harmonic. The 7th is no more special or threatening than the 8th or 6th or any others. Because of it's lower frequency, I'll wager that 5th and 6th are more important than 8th or 9th. Notice how I avoided the 7th? I don't want it stuck to the bottom of my shoes.

-Chris

[Joshua_G]

Originally Posted by anatech

Well, no. Of course not, and I have been very careful to always say that successful designs depend both on measurements and listening tests. On this I have been very consistent on. The fact you are asking means that you are either attempting to take an extreme view, or you've not read my posts in the past. Not that it matters from where I sit.

Now, to answer your question. Measurements can reveal problems that are audible. As you become more adept at measuring and knowing what to look at, measurements can often tell you if an amplifier has a chance of sounding good. You need experience to see this.

We are in full agreement here, we both agree that measurements can reveal problems that are audible, however, successful designs depend both on measurements and listening tests. The point I was trying to make initially, that started the present debate between us, is that measurements alone aren't enough to tell how an amp will sound, at least not published measurements. I may have chosen my words not carefully enough, however, that was my only point. Since both of us agree here, I see no point in continuing this debate.