John Curl's Blowtorch preamplifier

Status
Not open for further replies.
I would agree, PMA. 50% measurements and 50% listening tests.
In fact, I rarely audition my designs in any significant way, today, before they are released. I DO measure them, and I rely on previous experience with the passive parts, active parts, and the circuit topologies from making dozens of different examples of essentially the same thing, e.g. an preamp, or a power amp, or an electronic crossover, studio board, etc. I gained much of my listening experience in earlier years, and I rely on it now, and IF I can, I use others to make the listening evaluation for me, if I think they are competent to do so. It is dangerous, I think, to evaluate your own designs exclusively. Too much tendency to get your own ego involved. Anyone should be able to understand that.
 
stinius said:
In my opinion it’s not very interesting to discuss if there are a Sousa band or a F1 racing car hidden in the noise, it’s in fact possible to hide everything in that noise.
The interesting issue is to look at the different parts in an amp that make a difference to the signal that we are trying to reproduce.

Cheers
Stinius

Stinius,

Tou are right that in the end it all about the sound, BUT wouldn't it be nice to have some kind of clear indication what it is we shouild worry about in parts we use? Anything that helps with that is welcome, and DiffMaker is a great tool for that.

Jan Didden
 
Hi Grey,
From a post way back there, 14654.
Yes, I am well intentioned with what I post. What I find disturbing are all the unnecessary personal digs that are thrown around. All of this tends to reflect back on the poster of these attacks rather than on the victim of same.


'Most audio engineers' don't post here.
That is fact simply by the sheer number of audio design engineers out there compared to the number here. However, the amount of intellectual talent that posts on this web site is extremely high. That fact should be humbling, even if ever so slightly, to the rest of us here. Another fact that is overlooked is that any engineer that processes signals for measurement purposes is doing the same thing that an "audio design engineer" does. The huge difference with them is that they must be accountable for their work and be able to prove through measurements that the equipment is accurate. Designing audio equipment is simply a specialized signal processing engineer. Therefore, the same principles apply with the exception that the audio equipment designer does not consider reliability of the equipment to the same extent that the wider field of signal processing does. My other point would be that hobbyists and technicians may have considerable experience with this subject also. They can not be dismissed out of hand. That seems to be one theme that consistently appears in the subjectivists camp.

Imagine that. People who are not really technical, have a lower level of understanding of circuits and who do not even design tests for anything are holding their experiences higher than those who do. Simply amazing. Now, that is beginning to sound like a religion or cult. Even the responses from the subjectivists camp to posts that question their beliefs are similar to the responses from religious leaders. I would even go so far as to say that the efforts of those in the subjectivists camp have created the equivalent of the "dark ages" starting around, say, 1976 and up to the present. It's a scary parallel if you ask me.

My own personal experience from dealing with real live engineers is that they watch the measurements carefully, using them as a useful metric. But they do listen to the results of their work as well. These people have also learned how to interpret the findings of those measurements so that they do have some correlation to what they are hearing. True, there are some who only look at the instruments, but for audio design, those people do not produce better sounding equipment as a rule.

Now, those in the subjectivists camp have a habit of disregarding what the measurements are telling them. They also tend to insult and disregard any opinions from people who take the time to make good measurements. Note that measuring something does not mean the results are valid. The test procedure has to be properly executed as well. The same ideas hold true for listening tests. If that test is not controlled in some way, the results are simply not valid.

Interestingly, people who do tend to take measurements on things also accept the uncertainty of those readings and the test procedure. In contrast, those in the subjectivists camp tend to hold a result as final and correct with no uncertainty. To me, that right there invalidates much of their arguments.

Now, when I do a listening test, it is usually an informal thing carried out over some time. I find that I may notice more things, good or bad, over normal listening. But those results are not the final word at all. What happens next is that the equipment goes back on the bench where I attempt to correlate the things I may have heard with what I can measure or see. In every case if I have been able to take a measurement that shows some of this effect, I can normally track it down. From there, the problem can be corrected, confirming once again with a listening test.

Listening tests are tools, just like any instrument you can buy. It's the setup of your testing and how you interpret the results that makes the difference. The point is that this partnership is necessary. Excluding on or the other handicaps the process.
Note that John has mentioned measurements many, many times, yet he's clearly in the listening camp.
Is this due to the quality of that test equipment, or a failure to interpret the results in a manner to be useful? Those would be forgivable situations. To ignore measurements due to a misguided belief that they do not represent useful information is a real shame.

There are two main problems with audio reviews.
At least! But we agree there are problems here.
First is the fact that print readership is declining across the board.
In my view, that had little to do with the quality of these equipment reviews. Having also worked as an audio salesman in the late 70s and early 08s, reviews and excerpts from these were part of my life. Most salespeople did not read the information handed out by the manufacturer's sales department. Incoming customers often knew far more about the product that most other salespeople. Sad. Therefore, manufacturer's had to rely on what they could get directly into their customers hands in the form of equipment reviews.

What a perfect weapon! What was for some time an accurate and factual information source for buyers became a vehicle to "pimp the product". They needed to create iconic writers that the public could accept as respected professionals. The biggest con job ever pulled on the public. Second only to the "last year's models are lacking in performance". Yet another big lie. Today those reports have devolved into praise and worship rags. All competing for inclusion into a product brochure.

-Chris
 
vuki said:
Funniest thing with DM sample files is that there is difference in green pen sample tracks, but not in the hires/cd sample tracks. That is, there could be difference even there but outside hearing range.
Green pen works! 😀


But that's the crux, isn't it? It's not whether there is or is not a difference, I mean, how far down would you want to go? The question is, is it audible to the majority of the listeners with some factor of confidence.

Jan Didden
 
janneman said:



But that's the crux, isn't it? It's not whether there is or is not a difference, I mean, how far down would you want to go? The question is, is it audible to the majority of the listeners with some factor of confidence.

Jan Didden

I think hi-end audio is not about majority of the listeners, because majority doesn't care. It is about those really interested in hi-quality audio reproduction. But then, there you have those who are interested in that from the point of achieving the most spectacular measuring data and those who are interested in achieving the most spectacular audio reproduction. Mandatory for the later is that they love the music and attend live musical venues regulary, so they could have some kind of reference. Otherwise they could get impression that the sound they achieve at home with their excellent measuring devices is better than live.
(the last sentence is for you Jan :rofl: :wave2: )
 
syn08 said:


I disagree, a frequency response change is not a distortion. But this was already beaten to death here and elsewhere.

<snip>

syn08, normally the difference between "linear" and "nonlinear" is closely related to the alteration of the spectral content of a signal.

So, if the original signal is altered, by for example a change in frequency response, but no new spectral content is created during this alteration, than it is called a "linear distortion".

Otherwise, if new spectral content is created during a process, than it is called "nonlinear distortion".

Isn´t that a common definition for these terms?
Maybe it would have been better to define one mechanism as "deviation" and the other as "distortion" ?!
 
Jakob2 said:


syn08, normally the difference between "linear" and "nonlinear" is closely related to the alteration of the spectral content of a signal.

So, if the original signal is altered, by for example a change in frequency response, but no new spectral content is created during this alteration, than it is called a "linear distortion".

Otherwise, if new spectral content is created during a process, than it is called "nonlinear distortion".

Isn´t that a common definition for these terms?
Maybe it would have been better to define one mechanism as "deviation" and the other as "distortion" ?!

I (and many other) can't accept that filtering is a form of "linear distortion". It is also difficult to quantify the amounts of "linear distortion".

It also introduces more error sources; its not only the cap type/material that can contribute, but also the cap absolute value. It seems to me that any variations due to the cap material will be masked by the matching error.
 
Anatech; i just feel the need to comment the "religious" side of your post. 🙂

I think it´s grossly misleading just to accuse subjectivists of this sort of behaviour.
Quite often socalled objectivists were not seeking for scientific results but for confirmation of their own "belief system" .

That led to (often) poorly designed listening tests with very predictable results.
As posted earlier, back in the 80s i was really shocked by the reaction of "objectivists" if their methodology was questioned.
Even if the questioner could provide an analysis based on mathematical/statistical reasoning- see for example the debate after Les Leventhals JAES-Article.

So imho it was quite often just a fight of two diverging belief systems that led to the unhappy situation, where nearly everythings seems to be possible.

Basically these are questions of psychoacoustics and engineers normally do not know enough about proper testing methodology to get knew and better results, but fail to detect their own shortcomings in this regard.
 
SY said:


It doesn't speak to the question of "can wine A and wine B be distinguished," but rather "how consistent are the judgments of relative quality between distinguishable wines." The latter is not good, which is why I have had many strong disagreements over the years with wine writers who do stupid things like use a 100 point scale.

In the case of boxes of gain where A cannot even be distinguished from B in a basic blind test, the idea of an audio writer having the similar hubris to assign point scores is beyond ludicrous, it borders on outright fraud.

I have just read a summary and have to dig into his article for further comments.

At a first glance i thought he would point to the fact that blind tests without sufficient controls were providing quite unstable/faulty outcomes.
 
syn08 said:


I (and many other) can't accept that filtering is a form of "linear distortion". It is also difficult to quantify the amounts of "linear distortion".

It also introduces more error sources; its not only the cap type/material that can contribute, but also the cap absolute value. It seems to me that any variations due to the cap material will be masked by the matching error.

I thought the concept of linearity/non-linearity is well defined.

Let a signal enter a black box, if what comes out is different in the time domain then the signal is distorded.
The time domain distortion is a linear one if the process in the black box is linear: Xin *k=>Xout*k and Xin +Yin=>Xout+Yout otherwise it is non linear.

JPV
 
JPV said:


I thought the concept of linearity/non-linearity is well defined.

Let a signal enter a black box, if what comes out is different in the time domain then the signal is distorded.
The time domain distortion is a linear one if the process in the black box is linear: Xin *k=>Xout*k and Xin +Yin=>Xout+Yout otherwise it is non linear.

JPV

So did I 🙂 However, some insists that if the black box is altering the frequency domain response, this qualifies as "linear distortion".
 
syn08 said:


So did I 🙂 However, some insists that if the black box is altering the frequency domain response, this qualifies as "linear distortion".

The more general definition relies on the relationship between input and output of a system.

If input and output are related in a linear manner but the signal is altered than it´s "linear distortion", if the relation between input and output is nonlinear than it´s "nonlinear distortion" 🙂
 
syn08 said:


So did I 🙂 However, some insists that if the black box is altering the frequency domain response, this qualifies as "linear distortion".

Am I missing something?

If for example the black box is a linear differential equation, it will change the input signal waveform so it is distorded but the spectral content is changed too( for example a low pass filter). It is anyway a linear system therefore linear distortion and modification of the spectral content of the input signal.

JPV
 
Jakob2 said:


The more general definition relies on the relationship between input and output of a system.

If input and output are related in a linear manner but the signal is altered than it´s "linear distortion", if the relation between input and output is nonlinear than it´s "nonlinear distortion" 🙂


JPV said:


Am I missing something?

If for example the black box is a linear differential equation, it will change the input signal waveform so it is distorded but the spectral content is changed too( for example a low pass filter). It is anyway a linear system therefore linear distortion and modification of the spectral content of the input signal.

JPV

Jacob: could you please define input and output are related in a linear manner and signal is altered.

Y=a*X+b defines a linear transformation over a normalized vector space. In particular, for the Fourier/complex representation of signals, it is claimed that if a and b are complex quantities then this is a "linear distortion". Only if a and b are real the signal would not be distorted in any way, shape or form. To me, this definition of distortion doesn't make any sense, a low pass filter with a=1/(1+T*jw) and b=0 is not distorting the signal.

Edit:
JPV: no, a linear differential equation does not distort anything. Just remeber that if f(t) is a differentiable function with Fourier transform F(w), then the Fourier transform of its derivative is given by jw*F(w). This can be used to transform differential equations into algebraic equations - and you just got a= jw as above.
 
bear said:
That is why the SE amplifier can sound so good.


As I mentioned earlier the work of D.E.L. Shorter and more recently Dr. Earl Geddes seems to hold a clue to how this works. They both showed that it is the spectra of distortion that matters much much more than the absolute amplitude.



_-_-bear


I thought Dr. Geddes work concentrated on perception thresholds of various kinds of distortion. Extrapolations are dangerous and not necessarily valid for very low levels of distortion. Just because 10% seconds sounds better than 2% sevenths (making it up) does not mean anything at the .001% level.

In my experience SE amps can have clearly audible coloration and still be prefered by listeners. Yes, audible even to me 🙂
 
Status
Not open for further replies.