I'm not uncomfortable.
To me, I'm not really bothered by the virtues of WHY, or WHY in context of this thread. I just state that there is 'tube sound' I hear it, for better or worse. IMO tube sound is both, better and worse. But no one owns the rule book or holds the position of righteousness.
Well, Chris!... I think maybe you must have burned your fingers just one time too many or something.
...just kidding. I believe the view expressed has a tad more validity than I might want to accept because "those" devices, those unspeakable abominations do not glow, you see.
As for my beloved tubes, "the tube sound"...hmm, it depends; a good tube sound, a bad tube sound, or a tube rig that is neutral?
In my view it seems to revolve in the mixed company of: 1.) The frequency variable characteristics of a transformer (if there is one)
2.) The variation of reflected load a speaker presents through the frequency domain, the mass of the associated moving elements (and the effect of respective back-EMFs).
and 3.) The valve: characteristic curves, how it is loaded, biased, the effects of capacities within it, the derived harmonic spectrum.
What the tube sound should be according to Aubrey:
There should not be Much of a tube sound, just music and some pretends-to-be-candle light. However, I must recall that experiment whose results were presented in The Wireless World (IIRC). When given a knob that adjusted harmonic distortion and asked to tweak it to taste, I believe it was around 2% and 3% that almost all participants settled into? I've wondered more than once if our extreme sensitivity to changes in harmonic spectrum distribution (identifying the timbre of a "brass" versus a "wooden" instrument, for example) might be pleased to have a tube restore some of the harmonics attenuated by recording elements of microphones. I'm not holding my breath for an answer, nor concluded that it is so, just curious.
Ultimately, the tube sound is in part what it actually is and in part what Your brain Interprets from what your ears hear. I believe it is sort of a mantra in the British audio communities, "If it sounds good to you, it Is good", I like that. I also like devices which have not devolved from the emanation of light from within!
Regards, Aubrey
...just kidding. I believe the view expressed has a tad more validity than I might want to accept because "those" devices, those unspeakable abominations do not glow, you see. As for my beloved tubes, "the tube sound"...hmm, it depends; a good tube sound, a bad tube sound, or a tube rig that is neutral?
In my view it seems to revolve in the mixed company of: 1.) The frequency variable characteristics of a transformer (if there is one)
2.) The variation of reflected load a speaker presents through the frequency domain, the mass of the associated moving elements (and the effect of respective back-EMFs).
and 3.) The valve: characteristic curves, how it is loaded, biased, the effects of capacities within it, the derived harmonic spectrum.
What the tube sound should be according to Aubrey:
There should not be Much of a tube sound, just music and some pretends-to-be-candle light. However, I must recall that experiment whose results were presented in The Wireless World (IIRC). When given a knob that adjusted harmonic distortion and asked to tweak it to taste, I believe it was around 2% and 3% that almost all participants settled into? I've wondered more than once if our extreme sensitivity to changes in harmonic spectrum distribution (identifying the timbre of a "brass" versus a "wooden" instrument, for example) might be pleased to have a tube restore some of the harmonics attenuated by recording elements of microphones. I'm not holding my breath for an answer, nor concluded that it is so, just curious.
Ultimately, the tube sound is in part what it actually is and in part what Your brain Interprets from what your ears hear. I believe it is sort of a mantra in the British audio communities, "If it sounds good to you, it Is good", I like that. I also like devices which have not devolved from the emanation of light from within!
Regards, Aubrey
I love tube amplifiers, and since most of the ones I make are OTLs there is plenty of opportunity for burning the fingers!Well, Chris!... I think maybe you must have burned your fingers just one time too many or something.
As for my beloved tubes, "the tube sound"...hmm, it depends; a good tube sound, a bad tube sound, or a tube rig that is neutral?
In my view it seems to revolve in the mixed company of: 1.) The frequency variable characteristics of a transformer (if there is one)
2.) The variation of reflected load a speaker presents through the frequency domain, the mass of the associated moving elements (and the effect of respective back-EMFs).
and 3.) The valve: characteristic curves, how it is loaded, biased, the effects of capacities within it, the derived harmonic spectrum.
What the tube sound should be according to Aubrey:
There should not be Much of a tube sound, just music and some pretends-to-be-candle light. However, I must recall that experiment whose results were presented in The Wireless World (IIRC). When given a knob that adjusted harmonic distortion and asked to tweak it to taste, I believe it was around 2% and 3% that almost all participants settled into? I've wondered more than once if our extreme sensitivity to changes in harmonic spectrum distribution (identifying the timbre of a "brass" versus a "wooden" instrument, for example) might be pleased to have a tube restore some of the harmonics attenuated by recording elements of microphones. I'm not holding my breath for an answer, nor concluded that it is so, just curious.
Ultimately, the tube sound is in part what it actually is and in part what Your brain Interprets from what your ears hear. I believe it is sort of a mantra in the British audio communities, "If it sounds good to you, it Is good", I like that. I also like devices which have not devolved from the emanation of light from within!
Regards, Aubrey
Your comment about having a tube amplifier "restore some of the harmonics attenuated by recording elements of microphones" is similar to another one made a couple of days ago by someone else on this thread, which I questioned and received no reply to. I'm not clear exactly what you are suggesting here. Are you suggesting that the microphone selectively filters out harmonics, leaving predominantly fundamentals? How would it "know" whether an 800Hz tone was a harmonic of a 400Hz fundamental, and therefore filter it out, or whether it was an 800Hz fundamental in its own right, and therefore not filter it out? I don't get what the mechanism you are proposing really is.
In any case, if the bottom line is that people like to hear more harmonics, whether originally present or not, then indeed there is nothing wrong with that. Provide a knob on the amplifier, like in the Wireless World experiment, and let them twiddle it to what they like best. This does rather accurately fit the description of an "effects box," I think.
Chris
Those are linear phenomena, not nonlinear or stochastic, at least to first order. But... you can't do these things noiselessly or perfectly, so in recovery, there's still some corruption (though easily made smaller than any audible threshold for EQ, not so much for Dolby). IOW, play them back with correction.
The Second Law is one you can't break.
I’m getting the impression that much of the failure to communicate and understand views effectively on this thread comes from the fact that we are talking about a number of different and distinct processes.
If we assume that the “subject” under consideration is the typical DIY Audio amplifier builder and that “acoustical” in this case refers to “that which is listened to” (from akoustos: "heard, audible") then we have these processes:
Process 1. Acoustical input > memory > acoustical output.
Here the subject is motivated to re-create a listening experience which is compared to the memory of the original listening experience. This has poor reliability because each stage is highly fallible, but it has good face validity. Face validity here is “subjectively viewed as covering the concept it purports to measure”. Listener experiences would be of the kind that subjectively states “to me it sounds real”.
Process 2. Acoustical input > acoustical output.
Here the subject listens to live instruments (in a DBT test if you like) and compares them with a recorded reproduction of the same instruments in the same space. Though this has eliminated the memory stage, this still has poor reliability but again good face validity.
Process 3. Electronic input > acoustical output.
Here the subject listens to a recorded reproduction which is compared with measurements of the electronic input to the amplification system. In this case we have dramatically improved reliability on the input end which permits a number of inferences to be made. But though we have improved reliability we now have poor face validity. For the subject under test this is not “subjectively viewed as covering the concept it purports to measure”.
In none of these processes do we have both high reliability and high face validity, which may explain why nobody has answered the original question posed – how do we get a mathematical model of what a listener perceives as “tube sound”. We have any number of inferences based on mathematical models of electronic inputs and outputs but these remain on the level of inferences.
If we assume that the “subject” under consideration is the typical DIY Audio amplifier builder and that “acoustical” in this case refers to “that which is listened to” (from akoustos: "heard, audible") then we have these processes:
Process 1. Acoustical input > memory > acoustical output.
Here the subject is motivated to re-create a listening experience which is compared to the memory of the original listening experience. This has poor reliability because each stage is highly fallible, but it has good face validity. Face validity here is “subjectively viewed as covering the concept it purports to measure”. Listener experiences would be of the kind that subjectively states “to me it sounds real”.
Process 2. Acoustical input > acoustical output.
Here the subject listens to live instruments (in a DBT test if you like) and compares them with a recorded reproduction of the same instruments in the same space. Though this has eliminated the memory stage, this still has poor reliability but again good face validity.
Process 3. Electronic input > acoustical output.
Here the subject listens to a recorded reproduction which is compared with measurements of the electronic input to the amplification system. In this case we have dramatically improved reliability on the input end which permits a number of inferences to be made. But though we have improved reliability we now have poor face validity. For the subject under test this is not “subjectively viewed as covering the concept it purports to measure”.
In none of these processes do we have both high reliability and high face validity, which may explain why nobody has answered the original question posed – how do we get a mathematical model of what a listener perceives as “tube sound”. We have any number of inferences based on mathematical models of electronic inputs and outputs but these remain on the level of inferences.
OTL...tubes..and the smell of seared flesh, What A LIFE!! I envy the, oh enlighted one! I hope I came across in good humor as intended. I did think you likely a tube hating siliconite, however, I hope you will please forgive.
Ok, so about the microphone thing. My thinking is that microphones have a linear response, where human ears have a log response, for starters. Mic elements have some mass. So let's say you can hear a harmonic of a guitar string fundamental note, the fundamental seems 8-ish times louder than the harmonic that you hear(well then, your hearing is better than mine!) so about 30 db down is where we find the harmonic when listening. The mass responds to the fundamental, but won't the same mass "mechanically" attenuate the harmonic? I don't pretend to know, but intuitively it is my guess that the mass would no longer be complicit to the urgings of a force many times smaller. So, what I suggested as if to "restore" would, in this case, be acoustic signal lost to diaphragm mass that never entered the electrical recording chain to begin with. I wonder if this might explain the preferred 2-3%D: the restoration of some harmonics that were original content of the instrument.
? - Aubrey
This is another one of those lovely theoretical debates where we can hold onto whatever position we like without actually proving anything in real life.
All amps are different and the notion that we can achieve wire with gain is laughably deluded.
The fact that some people think that we can remove the human interpretative element from a complex imperfect system is similarly laughable.
The fact that we are having another one of these pointless "theological" debates is the most laughable thing of all.
I say, go listen to your amps, and enjoy them for what they are - warts and all.
Shoog
All amps are different and the notion that we can achieve wire with gain is laughably deluded.
The fact that some people think that we can remove the human interpretative element from a complex imperfect system is similarly laughable.
The fact that we are having another one of these pointless "theological" debates is the most laughable thing of all.
I say, go listen to your amps, and enjoy them for what they are - warts and all.
Shoog
I hear you. It does all seem at times as if we are, in the words of UK jazz drummer Phil Seamen "trying to row the Queen Mary through an ocean of Mars bars"...... But exercising the brain is still healthier than foot dangling....
Obviously I deviated from focus on the original question, my pologies for that.
I don't think a useful model could be made without limiting methods, like the observation of a distribution curve and what "most" people would agree on the "tube sound" to be. everyone has ears with different hearing characteristics and brains with different interpretations of what was heard.
Without limits of some kind, the model could turn into the audio equivalent of the new and improved, faultless lie-detector (and I don't think such a thing could exist unless we stopped evolving).
- Aubrey
I don't think a useful model could be made without limiting methods, like the observation of a distribution curve and what "most" people would agree on the "tube sound" to be. everyone has ears with different hearing characteristics and brains with different interpretations of what was heard.
Without limits of some kind, the model could turn into the audio equivalent of the new and improved, faultless lie-detector (and I don't think such a thing could exist unless we stopped evolving).
- Aubrey
I think the debates only become theological because some people are intent upon blurring the lines between what is measurable and quantifiable on the one hand, and what is subjective opinion and preference on the other.
You say "the notion that we can achieve wire with gain is laughably deluded." In the sense that there is no such thing as a "perfect" amplifier, you are, in a rather trivial sense, correct. But what really matters is whether the amplifier can be made accurate enough that for all practical purposes it behaves as a "wire with gain." Measurements can be made with great enough accuracy that this can be an answerable question. If, for example, measurements show that the output, driving into the loudspeaker load, is simply a pure constant multiple of the input voltage to within an error of X nanovolts, and if the threshold of human hearing would correspond to a signal level of 1000X nanovolts, then it is safe to assert that the amplifier is behaving, for all practical purposes, as a pure "wire with gain," in the sense that it would not be humanly possibly to hear the difference.
It seems to me that you are trying to hide behind the "nothing in this world is absolute, so therefore we can never actually say anything" kind of thinking. This is just a lazy excuse for not confronting the difference between decidable questions and undecidable ones.
Another example concerns what we hear. If you say "I prefer listening to this particular amplifier, because it makes me feel more in touch with the music," then that is fine, and no one can argue against it. But if you say "I can hear a clear difference between amplifier A and amplifier B," then that becomes an objectively testable assertion. It can be tested by means of double-blind comparisons. If these confirm that you can reliably distinguish A from B, then your assertion is validated. If on the other hand you fail to be able to discriminate between A and B in the tests, then your assertion is proven false.
The topic of this thread concerns whether there is a distinctive "tube sound." Interpretations of what this actually means differ as to whether a distinctive difference that is merely the result of "trivial" and easily quantifiable effects like output impedance and frequency response count as "tube sound" or not. But it is a perfectly reasonable question to ask whether a solid state amplifier with a few resistors and capacitors added in order to match the output impedance and frequency repsonse of a tube amplifier will be distinguishable or not. And this is a perfectly decidable question, that can be settled by means of double-blind testing.
Chris
I hope I came across in good humor as intended. I did think you likely a tube hating siliconite, however, I hope you will please forgive.
? - Aubrey
That is exactly the spirit I took it in 🙂. I like tube amplifiers very much. Not because I think they sound any better, but because I like the warm glow of the tubes, and the feelings of nostalgia.
Chris
I disagree with the narrow definition of a double blind test. I would seek a different type of test. If a person had equal access to a pair of amps A & B for an unlimited amount of time, which amp would he gravitate to. i think that is a far more informative real world test which can tell us more about what people prefer and why.
I also think that most of the test batteries applied to a real world amplifier are currently far from adequate to capture all that is going on within the listening experience, and in this case I would rather have the human component firmly within the subjective loop of deciding what real people actually prefer.
I think we need real world tests for real world amps.
Shoog
See my article in Linear Audio vol 2 for how to set up a test like this, and the results of one done properly. Spoiler: ears-only, without peeking, the preferences still come out randomly for two boxes of gain with similar frequency response, noise, impedance, and all those other boring engineering parameters.
You could try it yourself and see if you get a contrary result.
It's also the same person who spent years doing sensory analysis professionally and published articles about it. 😀
- Status
- Not open for further replies.