Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Signal Coloration? (Read 23073 times) previous topic - next topic
0 Members and 2 Guests are viewing this topic.

Signal Coloration?

Reply #25 –
Quote from: Atmasphere on
Arnold, I detect a strawman in your conversation regarding: tubes vs transistors --> triodes vs pentodes. These are two distinct issues;  one knocked down is not the same as the other (which is the definition of a strawman).


Consider that BJTs are a lot more like Pentodes than Triodes. There was a logical progression - Triodes which have tons of built-in local feedback, therefore end up with relatively low loop gain, and are therefore run with low or no loop feedback. Pentodes, which have far less local feedback (that's what the two extra grids do), therefore end up with more loop gain, and need to and are generally run with more loop feedback. Transistors, which have even less local feedback, have tons of gain, and are generally run with lots of loop feedback.


The power paradigm is itself a straw man. Designing amps as constant power sources is soo backwards. It results in amps that can only work well when the speaker and amp are dsigned together which almost never happens. The problems of high source impedance were understood even in the 1950s, which is one reason that by the 60s just about every serious audio amp dropped triodes and had as much loop feedback as the designers could make stable given that they had to live with transformers.

A tiny minority serving a miniscule marketplace mostly composed of people with far more money than understanding of audio.  There's lots of talk about SETs but their actual market share is almost zero. Only by making artificial distincitions like saying that only the high end matters can anybody construct situation where amps with high source impedances are even relevant.


Quote
The fact of the matter is that most transistor designs sound dreadful on such speakers but Nelson pulls it off.


The fact tof the matter is that you are speaking out of the back of your neck. The vast majority of horn-loaded systems in actual use are designed for use with voltage source type amps. So-called full-range speaker designe sound dreadful regardless.

Quote
It is my contention that feedback (or lack of it) plays a bigger role than whether the device is tube or solid state. I appreciate your input if thoughtful-


That's an unfounded assertion. Inverse feedback works. It got us to the moon, and relatively pollution free-cars to name just a few areas where it has been a great success. There are almost no audio power amplfiiers built without it. Many of the cases where people claim to have built feedback-free audio gear they just traded loop feedback for local feedback.  The big problem with inverse feedback is that it really helps to understand calculus and analytical geometry to exploit it, and that leaves out your typical basement amp fiddler.

Quote
I would like to have someone really take that article on the Atma-Sphere site apart, but you have yet to present the convincing argument, although you did score some points.


I'd first have to teach you electronics in order to really convince you.

Quote
I would like to know: what transistor amplifier (other than those designed from the ground up) are known to have THD less than 0.5% at full power open loop?


Well, actually just about all of them. If you build a SS amp output stage, it usually has reasoanbly low distortion (0.1-0.5%) until you push it near clipping. When you build the driver stage you actually have a choice whether you build a drive stage with less local feedback and lots of gain and then exploit the gain with loop feedback to further linearize the output stage, or whether you put lots of local feedback into the driver stage, and let the output stage go it alone. Just about everybody does the former because they have thought things out.


Quote
My experience with transistors says that you open the feedback loop at your own risk!


Now that shows an amazing lack of insight. Of course you remove the loop feedback at your own risk because in the first place you are also imcreasing the amps gain by whatever the feedback factor is. You can easily end up with 20 to 40 dB more gain, which is like turning a power amp into a mic preamp.  If you don't turn the signal generator down the amp is going to suddenly clip like there was no tomorrow. If there are any problems with input-output leakage, the amp may even immediately become unstable. If you address both of those issues, then you come to the fact that you designed the input stage of your power amp as an input stage for a power amp not as a mic preamp, and the compromises you made for more voltage handling at the expense of noise come around to bite you.

Quote
as the amplifier needs that input to servo correct DC levels as well as distortion.


Well, that too if you lack the insight it takes to address that issue. Hint- see my former comments about using local feedback in the driver. The output stage usually has great bias stability all by itself.

Quote
So I would be interested to see what designs have been operated in real world situations open loop. My guess is that there are none- rendering the comment a bit of a red herring. Please prove me wrong!


Most output stages are almost laways emitter followers or somthing like them. What's to design in an emitter follower other than to try to pick output transistors with linear gain across the range of currents that are involved?  Look at the power amps that aren't simple emitter followers such as those by if memory serves Bryston?  The aswner is that the the output stages still have relatively low voltage gain, still unity. They still use lots of loop feedback. They do not have excpetionally good perforamnce compared to amps built more traditionally.

Quote
Regardless, the assertion that distortion signatures play a huge role in signal coloration seems to stand.


Only as a straw man that lives only until the first good listening test.


Quote
In case there is any issue about this, enhancements of odd-ordered  harmonics in the realm of 100ths of a percent are audible as a brightness or hardening in the sound


I invoke TOS 8.  I've done my homework and there is no reason to believe that the above statement can be demonstrated in a DBT because its been tried many times and it fails.


Signal Coloration?

Reply #26 –
Don't you just love it when someone shows up at HA touting his fringe technology, then has to actually defend it? 

Signal Coloration?

Reply #27 –
Quote from: pdq on
Don't you just love it when someone shows up at HA touting his fringe technology, then has to actually defend it? 


Feedback phobia is not so much fringe technology as anti-technology.

Signal Coloration?

Reply #28 –
I am happy to listen to one amplifier at a time, myself.  If a hundred of them strung together produce an audible coloration why should I care?  All that matters to me is that the particular one I am listening through should produce no audible coloration.  Even the amplifier in my iPod appears to meet that criterion!
Ed Seedhouse
VA7SDH

Signal Coloration?

Reply #29 –
Quote from: Ed Seedhouse on
I am happy to listen to one amplifier at a time, myself.  If a hundred of them strung together produce an audible coloration why should I care?


Of course the number of repeitions that need to be tested depends on the application, but I can only explain the need for 2 repetitions of power amps.

In the case of telephone repeaters, we found that 58 is the maximum number of real-world repetions, plus maybe a dozen or two  more for land lines on each end.

Quote
All that matters to me is that the particular one I am listening through should produce no audible coloration.  Even the amplifier in my iPod appears to meet that criterion!


Actually, an iPod headphone amp is one of the easier engineering "challenges" in the world.  Power amps actually take a little work to get right, still. It's ironic that a huge industry has been built on top of the fallacy that something is dreadfully wrong with iPods, other than of course that they're a little short on gain and power for people with less-sensitive earphone preferences.

Signal Coloration?

Reply #30 –
"Don't you just love it when someone shows up at HA touting his fringe technology, then has to actually defend it? biggrin.gif"

pdq, I like it too. So far in four years of talking about this, this is the only forum where I have seen any refutation that is worth taking a look at.

Arnold, so far yours has not been a bona fide refutation, although I think you are probably capable of it (please prove me right in the future)... BTW, here is what a strawman is:  http://www.nizkor.org/features/fallacies/straw-man.html
Your comment about the Atma-Sphere article does not meet that definition. However your comments about BJTs **does**.

So far, only theory has been flaunted regarding removal of feedback from an existing transistor amplifier. I've been in this business since 1974, in that time experience has shown again and again that theory and real world often have big gaps in them. Enough to drive my Bronco through in some cases!

Andy, I am very familiar with the Ayre, an amplifier I respect a lot. Here is an amp that was designed from the ground up as zero feedback, and **also** conforms to the Voltage Paradigm. I'm sure you also know that its also respected as one of the better sounding transistor amps you can buy. I often point to it as an example of where we might look to see advancements in the art. I'm not a fan of any of the Wavac designs, IMO they and most SETs are designed with specific intent to color the sound (through their distortion characteristics). I did make a point of explaining that in an earlier post and will not repeat myself. The larger SETs seem to me particularly problematic.

Arnold, a previous comment got by me without comment:
"Good tubed amps and good solid state amps sound the same in level-matched bias-controlled tests. No bona fide exceptions are known to exist."
If this were really true, we could all just go out and buy a cheap (but good) used unit from 20 years ago (or more) and make it work since there would be no sonic benefit from newer designs. IOW we could all just build a good, old amp design from 20-25 years ago and it should perform as well as the best out there today. That is a logical conclusion we can take from this comment, do you believe that? I would appreciate it if you would provide a reasoned explanation rather than the sort apocryphal statements and snipes you've been making (which do not forward your cause).

So far, signal coloration being caused more by distortion than frequency response variation still stands. If you've not done a study of the human ear, it might help to know that frequency response and frequency linearity as a rule of human hearing is a bit down on the list of those rules compared to how we perceive loudness and spatial imaging. It might help to think of your ancestors who had to survive a hunting tiger in the forest- it is not important to here the full bandwidth of a growl, nearly so much as how loud it is and where its coming from!

Signal Coloration?

Reply #31 –
Quote from: Atmasphere on
Arnold, a previous comment got by me without comment:
"Good tubed amps and good solid state amps sound the same in level-matched bias-controlled tests. No bona fide exceptions are known to exist."
If this were really true, we could all just go out and buy a cheap (but good) used unit from 20 years ago (or more) and make it work since there would be no sonic benefit from newer designs. IOW we could all just build a good, old amp design from 20-25 years ago and it should perform as well as the best out there today. That is a logical conclusion we can take from this comment...

I think it is accepted canon here, and amongst the rationalists, that Arnold's assertion is correct.  We beg for a proper test proving this wrong.

PS - your comment on theory vs practice is misleading.  Theory vs Practice is a problem of the social sciences - sociology, economics.  Design vs Implementation is a more apt description of what you're talking about and it is only when we get our language strict and correct that some answers become uncloaked. 

Creature of habit.

Signal Coloration?

Reply #32 –
Quote from: Atmasphere on
Arnold, a previous comment got by me without comment:
"Good tubed amps and good solid state amps sound the same in level-matched bias-controlled tests. No bona fide exceptions are known to exist."
If this were really true, we could all just go out and buy a cheap (but good) used unit from 20 years ago (or more) and make it work since there would be no sonic benefit from newer designs. IOW we could all just build a good, old amp design from 20-25 years ago and it should perform as well as the best out there today.


This has been true since at least 1990.  It might well have been true in 1970 since several tubed amps designed in the 1940s were likely very close to audible transparency.  Amps that were specified and measured in audio magazine reviews at +- 1db from 20hz to 20khz and < 0.1% distortion were available, if my memory serves me, in the 1950s.

Ed Seedhouse
VA7SDH

Signal Coloration?

Reply #33 –
Quote from: Atmasphere on
So far, only theory has been flaunted regarding removal of feedback from an existing transistor amplifier. I've been in this business since 1974, in that time experience has shown again and again that theory and real world often have big gaps in them. Enough to drive my Bronco through in some cases!


Your statement about not being able to physically open the loop is a red herring.  Look at any spec sheet of any op-amp, and you'll see a large number of open-loop parameters specified: open-loop gain and phase vs. frequency and so on.  And yet it's not physically practical to open the loop because the op-amp will hit the rail due to its high DC gain.  None of these specified parameters depend on being able to physically open the loop in order to measure them.  They're computed from various types of closed-loop measurements.  Another example is the new low-distortion op-amps from National.  The distortion of these is too low to measure reliably even with the best AP equipment when the gain is set to typical values (high feedback).  So they have to reduce the feedback, and calculate what the distortion would have been for the target configuration.  The theory for doing all this has been known, and much more importantly, tested in the real world for decades.  The same considerations for determining open-loop behavior apply with feedback power amplifiers, since they are essentially "op-amps on steroids" - except for a small number of oddball designs.

Quote
Andy, I am very familiar with the Ayre, an amplifier I respect a lot.


I'm sure it's a very good amp.  My point in bringing it up was as an example of a class AB transistor amp having distortion lower than 0.5 percent without feedback.  This AES article by Bob Cordell (PDF file) has an example of an amplifier which, with a few resistor changes, can be configured to work with or without feedback (figure 13).  The two 8.2k resistors reduce the DC gain for the no-feedback configuration.  They get removed when feedback is applied.  I've seen schematics of earlier Ayre power amps, and topologically speaking, they're very similar to the open-loop portion of typical complementary feedback amplifiers.  In fact, it would be easy to modify them to use feedback with a very small number of changes, much like Bob's article that I linked above.  The distortion of such amps is dominated by that of the output stage (because it is class-AB), so nothing magic happens to the open-loop distortion as a result of such reconfiguration.  In fact, removal of the 8.2k resistors in the article referenced above will reduce the open-loop distortion by unloading the voltage amplifier stage.
 

Signal Coloration?

Reply #34 –
"I think it is accepted canon here, and amongst the rationalists, that Arnold's assertion is correct. We beg for a proper test proving this wrong.

PS - your comment on theory vs practice is misleading. Theory vs Practice is a problem of the social sciences - sociology, economics. Design vs Implementation is a more apt description of what you're talking about and it is only when we get our language strict and correct that some answers become uncloaked. wink.gif"

I do know what the 'accepted canon' here is, and it might be coincidence, but the use of the word 'canon' is in keeping with my viewpoint of what I am seeing here now. Its a religious viewpoint, not supported by reality- IOW, its made up. Its easily enough proven- the difference between a high quality <a href="http://www.v-cap.com/">V-Cap Teflon cap</a> and a Xicon film cap is easily discerned in many amplifiers. If you can change a cap, which does not affect gain or bandwidth, then the sound of that amp would be changed with regards to a 'double blind' test. One of the problems is double-blind tests BTW is the dumbing down of connectivity- and the fact that changing on-the-fly is not the way to do the audition- the track has to be repeated. The second thing is- you are not going to hear a lot of difference between competent transistor amps. IMO, unless a SS guy develops a breakthrough, there is no point in running such a test- its moot. The real differences, which seem to me obvious, are the differences between tube and solid state. However, as we all know, tubes are not going to double power as you cut the impedance in half, so you would have to choose a speaker that displays what the tubes and transistors can do on the same footing- likely something with a flat impedance curve and at least 8 ohms. Then we can eliminate EQ, which is part of the dumbing down (in case you are not aware of this, there are no EQs that can do their job without also adding a distortion signature). 

I appreciate your PS too and agree with it. But- that comment flies in the face of the topic of this thread.

The most important thing I have been maintaining here is that the Rules of Human Hearing are number one, from which all thing audio must emanate. It cannot be any other way else the electronics are going to sound like electronics, which is 99% of what exists in the world today (no question that that makes me a fringe element- so go ahead and castigate me for wanting things to be better- you have to try harder than that to set me off). IOW we as an industry have placed bench measurements ahead of how our ears behave, **and stuck to that even when greater knowledge of human hearing has been available for decades". Gentlemen, this is an example of the Emperor's New Clothes.

So my raison d'etre for being here is this: if we can understand how the ear works, then maybe what should be done is design equipment that follows those rules and even takes advantage of them. **NOT** the other way around, which is to create bench specs that have little meaning to the human ear. The proof of this: take any 'good' set of specs, and tell me how that amp is going to sound. You can't do it, and audiophiles have known that for decades. Yet every manufacturer says they have the 'best', just as y'all here are saying you have the 'best' way. Since 'best' has a very specific meaning, most everybody has to be lying. Faced with this for the last 40 years, that is why audiophiles so frequently have to audition components in their own home. If the specs were instead based on the Rules of Human Hearing, they would then reflect our listening experience.

Its a shame that I have to point out- that this is just common sense.

Signal Coloration?

Reply #36 –
Quote from: Atmasphere on
Its easily enough proven- the difference between a high quality <a href="http://www.v-cap.com/">V-Cap Teflon cap</a> and a Xicon film cap is easily discerned in many amplifiers. If you can change a cap, which does not affect gain or bandwidth, then the sound of that amp would be changed with regards to a 'double blind' test.

Are you claiming that the change of this capacitor made NO measurable difference in IM, THD, noise or any other electrically measurable parameter, yet was heard in an ABX test?

And why did you put quotes around "double blind"? Was that to distinguish it from what we at HA would refer to as double blind?

Signal Coloration?

Reply #37 –
I always love the audible capacitor brand argument.

I don't know much about circuit design, but I have google, and these few circuits I just saw online had six to twenty capacitors in each. But audiophiles rarely say where that audible capacitor would be in the amplifier circuit. Kind of: "V-cap teflon makes sound velvety while Xicon film adds air and scene and soundstage", doesn't matter if they are in the signal path, or in the power supply, or where.
Ceterum censeo, there should be an "%is_stop_after_current%".

Signal Coloration?

Reply #38 –
If you want some real entertainment, check out this capacitor thread.  There's all kinds of gems in there, including "Jupiter Beeswax Capacitors", people chiming in that it's the best thread they've ever read, and so on.  Of course it's a sticky thread because it's soooo valuable  LOL!

Signal Coloration?

Reply #39 –
Quote from: Atmasphere on
I appreciate your PS too and agree with it. But- that comment flies in the face of the topic of this thread.

Then you misunderstand my PS.
It appears to me you try to colour the argument as one of "theories are one thing - but it takes a skilled practitioner (yourself) to make it work."
I call BS on this apparent belief.
It is a field of compromises and a market where corners get cut - but that is a far cry from a theoretical guessing game where artistic practitioners hone their craft to the point of making discoveries heretofore unbeknown to electrical engineers.

I won't even get into your attempts to dismiss the established rigors of DBTing, for you are not waving your hands at some dogmatic belief by some fringe cult (here at HA) but rather attempting to write-off established scientific procedure.

Quote from: Atmasphere on
I do know what the 'accepted canon' here is, and it might be coincidence, but the use of the word 'canon' is in keeping with my viewpoint of what I am seeing here now.  Its a religious viewpoint, not supported by reality- IOW, its made up. Its easily enough proven- the difference between a high quality <a href="http://www.v-cap.com/">V-Cap Teflon cap</a> and a Xicon film cap is easily discerned in many amplifiers.

emphasis mine
Choice that you should seize upon my colourful choice of words and attempt, once again, to dismiss with much hand waving.  If it is so easy - then do it or show where it has been done.  I have never seen a successful DBT of the audibility of capacitors which is unpredictable by their electrical characteristics.  Be a man - step up - be the first.

edit:  made nice
Creature of habit.

Signal Coloration?

Reply #40 –
Quote from: Atmasphere on
Its easily enough proven- the difference between a high quality <a href="http://www.v-cap.com/">V-Cap Teflon cap</a> and a Xicon film cap is easily discerned in many amplifiers. If you can change a cap, which does not affect gain or bandwidth, then the sound of that amp would be changed with regards to a 'double blind' test.


If it's so easy to prove an *audible* difference, then prove it.

Otherwise, you're just another audiophile blowhard.


Quote
One of the problems is double-blind tests BTW is the dumbing down of connectivity- and the fact that changing on-the-fly is not the way to do the audition- the track has to be repeated.


Then go ahead and repeat it. That's not forbidden in DBTs, btw.  And what is this 'dumbing down of connectivity' of which you handwave?

Signal Coloration?

Reply #41 –
Quote from: Atmasphere on
Arnold, so far yours has not been a bona fide refutation,


Depends what level of evidence you need for a refuation to be bonafide.


Quote
I've been in this business since 1974,


New-comer then. I've been in this business since 1959.

Quote
in that time experience has shown again and again that theory and real world often have big gaps in them.


If you keep up with technology, then you know that those gaps have a way of getting filled-in.


Quote
Arnold, a previous comment got by me without comment:
"Good tubed amps and good solid state amps sound the same in level-matched bias-controlled tests. No bona fide exceptions are known to exist."

If this were really true, we could all just go out and buy a cheap (but good) used unit from 20 years ago (or more) and make it work since there would be no sonic benefit from newer designs.


Been there, done that.

BTW here's the last 20-year old amp that I worked on:

http://cgi.audiogon.com/cgi-bin/auc.pl?amp...mp;3&4&


Quote
That is a logical conclusion we can take from this comment, do you believe that?


What you can read from my lips is that "all good amps sound the same in any reasonable comparison", but that as you increase the size of your view of performance, technology continues to march on in favorable ways.

Quote
I would appreciate it if you would provide a reasoned explanation rather than the sort apocryphal statements and snipes you've been making (which do not forward your cause).


The explanation is simple - show me a situation where two good amps sound different. Don't bluster, don't make veiled insults, don't drop names, don't try to give me a history lesson. Show me!  That's HA TOS 8 and and you appear to be willfully in violation of it.

Quote
So far, signal coloration being caused more by distortion than frequency response variation still stands.


Oh, I forgot to metion unqualified, unquantified, unsupported assertions. Please avoid those, too.


Quote
it might help to know that frequency response and frequency linearity as a rule of human hearing is a bit down on the list of those rules


So you think that *any* frequency response difference is less important than *any* difference in nonlinear distortion? So far, that's a statement that you would seem to support.



Quote
compared to how we perceive loudness and spatial imaging.


Well, loudness relates to how we hear differences in frequency response. I think you may have just contradicted yourself.

Quote
It might help to think of your ancestors who had to survive a hunting tiger in the forest- it is not important to here the full bandwidth of a growl, nearly so much as how loud it is and where its coming from!


Bandwidth and frequency response are really two different things. By conflating them, we can see where you are coming from...

Signal Coloration?

Reply #42 –
I invoke TOS 8.  I've done my homework and there is no reason to believe that the above statement can be demonstrated in a DBT because its been tried many times and it fails.
[/quote]

Occam's Razor suggests that you resist for reasons having nothing to do with audio. So Arnold, get someone else to push your questions- someone who will be reasoned instead of apocryphal. I don't mind a reasoned debate, but I am not going to play a verbal abuse game with you. Obviously the DBT had a flaw in its setup or its assumptions.

FWIW, I suggested a method earlier so you could at least prove the odd-ordered harmonic issue to yourself. I got no takers... TOS 8 yourself.

I will clear one thing up though. Odd harmonic content is how the ear determines how loud a sound is- **not** Fletcher-Munson. Arnold was confused.

Regarding capacitors: one can easily measure changes in bandwidth and distortion based solely on the replacement of a capacitor. This should be no surprise when one looks at the specs of the dielectrics. Is it a double standard of some sort when those measurements are eschewed by a measurement community such as this??

The Emperor is still naked...



Signal Coloration?

Reply #43 –
Quote from: Atmasphere on
Regarding capacitors: one can easily measure changes in bandwidth and distortion based solely on the replacement of a capacitor. This should be no surprise when one looks at the specs of the dielectrics. Is it a double standard of some sort when those measurements are eschewed by a measurement community such as this??

The Emperor is still naked...

Wow - we're mixing and matching to suit our needs, aren't we?
I think my relevant quote on capacitors was...wait for it...
Quote from: Soap on
I have never seen a successful DBT of the audibility of capacitors which is unpredictable by their electrical characteristics.

(added emphasis)



I'm curious, though, about one thing.
Quote from: Atmasphere on
If you can change a cap, which does not affect gain or bandwidth, then the sound of that amp would be changed with regards to a 'double blind' test.

These magic caps, are they changing bandwidth or not?

Sheesh, I don't know why I'm asking you.  You seem uncertain yourself.



Creature of habit.

Signal Coloration?

Reply #44 –
Quote from: Atmasphere on
Quote

I invoke TOS 8.  I've done my homework and there is no reason to believe that the above statement can be demonstrated in a DBT because its been tried many times and it fails.


Occam's Razor suggests that you resist for reasons having nothing to do with audio.



You seem to be highly confused. I'm not resisting TOS 8, you are. It's time for you to show us what you've got, besides baseless assertions.  Show us your DBT results or cease and desist.

Signal Coloration?

Reply #45 –
Quote from: Atmasphere on
Regarding capacitors: one can easily measure changes in bandwidth and distortion based solely on the replacement of a capacitor.


Well yes, if I don't use capacitors with the same basic characteristics, such as capacitance, ESR and ESL, then I might change how some circuits work.

There have been standard guidelines for applying different kind of capacitors that have been around for more than half a century.

Many of these standards were old when  Marsh and Jung were young, and well before they wrote their so-called "landmark" science fact/science fiction piece for The Audio Amateur, and then Audio.

Quote
This should be no surprise when one looks at the specs of the dielectrics.


That all depends on which dielectrics and which application. The idea that every dielectric has to make an audible difference in every application is fantasy backed up with absent technical testing and egregiously bad subjective testing.

The idea that you're going to make an audible difference in the majority of pieces of commercial equipment that are BTW egineered to contemporary standards is nonsense.  AFAIK there have never been even technical tests (which are far more sensitive than listening tests) to support many of these weird science ideas about capacitor dielectrics.

Quote
Is it a double standard of some sort when those measurements are eschewed by a measurement community such as this??


You are again confused. HA is not a "measurement community" nearly as much as it is a "properly done listening test" community.  It appears to me that I'm dealing with someone who is a stranger to both...