Sound Quality Vs. Measurements

Status
Not open for further replies.
Let's calculate on the napkin.

We have an inverting amp (you like parallel feedback, right?), with couple of additional resistors: R1 from source to input, and R2 from output to input.
Let's assume that the amp has infinite input resistance and zero output resistance.

That means, current through resistors is (Vo-Vs)/(R1+R2).
If output voltage is Vo, amplification factor of the amp Aa, that means input voltage Vin is Vo/Aa.

Voltage drop on R1 is Vs-Vin, i.e. current through resistors multiplied by R1 value.

Vs- (Vo/Aa) = R1* (Vo-Vs)(R1+R2)

Find from this formula ratio between Vo and Vs, and you will get resulting amplification factor equal to ratio between resistors with error reverse proportional to the open loop amplification factor. This reverse proportion is what DF96 meant.
 
nigel pearson said:
but I feel uneasy with it . To me a loop must have echos
Engineering intuition is a bit like moral conscience - when unconstrained by reliable reference points it can lead you astray. For engineering intuition the reference is given by sound theory and the accompanying mathematics. Get that right, then your intuition will be a reliable guide.

At audio frequencies an open-loop amplifier behaves like a filter, not a transmission line, so there are no echoes. As I tried to explain, re-entrant distortion arises from algebra not time delays. If you can do the algebra you will understand; if not you will just have to take my word for it.
 
Thanks for that . In a mechanical sense we see that as we work . Even if the maths are put to one side it is almost like playing a musical instrument how distortion increase and decreases with manipulation . I find tube amps get you in trouble quicker . Not surprising with output transformers and other problems . It forces good compromises perhaps ?

I can see how the maths works . I often use similar in other things I do although it is about the limit of how far I use maths . The liberation is when we bend the maths and the magic happens . Not contradicting it , just finding cancellations if possible . The Quad 405 almost was magic . I remember the storms it caused . Poor Peter Walker and him so much better versed than me . If I am right all that is wrong with it is penny pinching in the main . It is sad because it made people like me say the mathematical approach had let us down . So badly let us down as to make many of us think ones who use maths seriously and perversely lied . Like they had no shame . Not about what they could prove , but what they couldn't . I am only just now trusting maths with sound again . For years I just listened .

Peter Walker did his PHD late in life . If I remember he said it was much more difficult than he expected . How honest of him .

As much as I shouldn't I will keep loops in my belief system . When dealing with resonance in engines I perhaps would take a different view . There were chaotic failure of low revving very high power diesel engines . This was never successfully analyzed . However oil squirted in sequence with the fuel injection under the piston crown solved it . The problem was the engine running under identical conditions for ages would suddenly develop piston hot spots . Think carefully , a diesel engine is remarkably stable . It can not be transient effects . The brainwave was to think that it must be related to fuel injection so just squirt in sequence . It worked . The thing is until the power reached this level the effect was unknown . The piston was receiving good cooling . Squirting seemed the thing it needed . Perhaps not just cooling ? Something dynamic and not just thermal .

I remember how perfect digital was . Funny how it got better .

I found this refreshing . It is all stuff said before however it reconfirms my observations .

https://passlabs.com/articles/audio-distortion-and-feedback
 
Last edited:
Engineering intuition is a bit like moral conscience - when unconstrained by reliable reference points it can lead you astray. For engineering intuition the reference is given by sound theory and the accompanying mathematics. Get that right, then your intuition will be a reliable guide.

At audio frequencies an open-loop amplifier behaves like a filter, not a transmission line, so there are no echoes. As I tried to explain, re-entrant distortion arises from algebra not time delays. If you can do the algebra you will understand; if not you will just have to take my word for it.

That's beautifully put . No sarcasm , I was almost tearful . Algebraic ghosts I think I will call them .

Thanks about ECC 82 and 81 . I had found those things . I like the 81 .
 
Seriously non-linear feedback systems can exhibit chaotic behaviour. That may be what happened with your diesel engine. It is also surprisingly easy to create apparently simple mechanical systems which are chaotic, such as some compound pendulums.

The Pass paper you link to is notorious for denying that emitter/source/cathode degeneration is feedback (R4 in fig 9). Not to start that argument again here, but be aware that Nelson Pass is not mainstream however much his fans wish he were.
 
Local and global feedback are conceptually exactly the same, and obey exactly the same mathematics. In either case the open loop system is a combination of amplifier and filter. The filter bandwidth may be quite different, but that is not a fundamental change.

The particular circuit shown by Pass was a single stage, so in this case 'local' and 'global' feedback are over the same stage!

Anyway, we have already had this argument.
 
I'll throw my two cents worth in this huge pot. I have an overall theory about why two things can measure the same, and why they tend to sound different. My theory is that the whole mess of audio frequencies, and there are hundereds playing at the same time, will tend to cancel each other as well as amplify each other along the signal path. And depending on what variation of component you are using, the component (capacitor, inductor (speaker coils, and filters), wire composition, resistor types (metal film, carbon, etc.) will allow these interferences to interfere in different ways. I don't know if I am right, but that is what I think may be happening here.

When people measure things, they usually just measure one freq at a time, so the opportunity for harmonic frequency interference is not there. It is there however if there are the multiple hundereds of frequencies available during a musical piece.

For instance, why would one capacitor sound different than another. Well small vibrations between the plates can cause an audible change in sound, and if you have a cap that tends not to vibrate as much...Even if the caps are of the same capacitance, you will find audible differences. They of course can have the exact same measurable qualities, but will sound different.

I have a limited knowlege of electronics since I only have an associates degree in it, but if there is someone more familiar with frequency interference, especially in electronic components, please chime in if you haven't already!
 
When people measure things, they usually just measure one freq at a time...

Depends on what they're looking for. Multitone, MLS, and impulse response are all pretty common techniques these days. Single tone is useful, since it exposes many-if-not-most of the common practical distortions, but no-one claims it to be universal. There's also IM (two tone), which has been pretty routine since the 1950s.

To date, *no-one* has demonstrated an audible difference between audio components that measure the same, assuming by "measure," you're talking about doing a competent and reasonably thorough job of it.
 
To date, *no-one* has demonstrated an audible difference between audio components that measure the same, assuming by "measure," you're talking about doing a competent and reasonably thorough job of it.[/QUOTE]

I guess your right by saying "competent". The differences in components are apparent since they could be made from different materials, and still exhibit similar capacitance, but discharge at different rates, and have other physical properties that might also be measureable.

I guess if we live in the audiophile world, experimentation after measurement (limited to equipment availability), is what is the fun part of being into improving your audio setup to your specific tastes.
 
John, you should seriously consider starting to smoke.

There's nothing like a good cigar after a good lunch. Or in the evening, sitting on your porch (or standing on your window if you live in a condo).

Feel free to retaliate with good advice on drinking, because I don't drink. AND I have a strong preference towards Motorola/ON Semi transistors.

I don't gamble and I don't wench around, at my age chasing skirts is limited to downhill only. :))))))
 
Status
Not open for further replies.