If "Tube Sound" Is a Myth, Why Tubes?

[Julf]

How is it obscure?

Or maybe you didn't see the previous minor discussion about the sound of live instruments (e.g. brass).

I related my perception of live instrument and vocal sounds to what I look for and appreciate in reproduced sound, and how I find reproduced sound of those sound sources often depart from the "real thing." Curious what you are confused about.

Just confused about what your actual point is - apart from relying on fallible auditory memory.

 

[MattHooper]

Just confused about what your actual point is - apart from relying on fallible auditory memory.

Again, not sure how you can *actually* be that confused.

Yes, auditory memory can be fallible. All our memory, and senses, can be fallible. And yet they are reliable *enough* to allow us to mostly successfully navigate reality. Which means: how far do you want to push the point of "fallibility?"

Do you think your auditory memory is so fallible that you can't recognize your family's voices from one day to the next on the phone? Do you think I would not be able to recognize the sound of a saxophone vs a xylophone because I heard the saxophone yesterday and the xylophone today?
You can't "see" all the instruments when playing your recorded music at home - e.g. a jazz, rock band etc - and yet you generally accurately identify the trumpet, the sax, the drums, keyboards etc in a mix, without suggesting any need for blind testing right?

So...yes...auditory memory is fallible, but far from 'completely unreliable' so long as the sonic characteristics in question are large enough. It's in the ever more subtle aspects that auditory memory decreases.

Do you *really* think it's implausible that I am perceiving a *real* difference in the sound of a LIVE TROMBONE right in front of me vs that on a recording played in my little Spendor s3/5 speakers? You don't think there are some truly identifiable characteristics to distinguish those reliably...like most trombones I hear through my Spendors are indistinguishable from the real thing in front of me? Like, I'd need a blind test to know that a live drumset being played in front of me sounds different from an old recording through tiny Spendor speakers?

What, exactly, is your point?

 

[Robin L]

Just confused about what your actual point is - apart from relying on fallible auditory memory.

As do we all?

Subjective impressions of the sound of live, unamplified instruments are useful if one is a recording engineer [though the term "recordist" is closer to the mark as one doesn't need to be an engineer to be good at making recordings]. Getting a handle of how one gets close to the sound of a given instrument reqires a lot of subjective evaluation and comparison, that's baked into the equation.

 

[Julf]

As do we all?

For what? remembering how we believe we heard something, yes. For actually comparing two devices? Only if we don't care if the comparison makes any objective sense.

It's useful if one is a recording engineer [though the term "recordist" is closer to the mark as one doesn't need to be an engineer to be good at making recordings]. Getting a handle of how one gets close to the sound of a given instrument req

Indeed. Recording engineers are engineers in the anglo-saxon meaning of "someone who operates equipment" (see "locomotive engineer"), as opposed to the academic "someone with a higher degree in applied physics".[/QUOTE]

 

[Blumlein 88]

Yes but Amir measured this Amp (not just same model but this actual amp) with his AP and it was only down 0.1 dB @ 10 Khz and .5 dB @ 20 Khz so the drop starting @ 4 KHz seems like it must be something else when hooked up to a speaker?

The speaker is not a resistor so you'll have reactance which can cause the response changes which are not quite simple. So both Amir's and your results are possible.

 

[Julf]

Yes, auditory memory can be fallible. All our memory, and senses, can be fallible. And yet they are reliable *enough* to allow us to mostly successfully navigate reality. Which means: how far do you want to push the point of "fallibility?"

I tend to push it to "fine for everyday situations, nowhere enough to support any claims that contradict established science".

Do you think your auditory memory is so fallible that you can't recognize your family's voices from one day to the next on the phone?

Of course not. But that doesn't take much. I can recognize the voices of my family in a very noisy environment over a very distorted, chopped and bandwidth-limited connection. That pretty much shows that our brain is actually very good at *not* detecting small distortions.

Do you *really* think it's implausible that I am perceiving a *real* difference in the sound of a LIVE TROMBONE right in front of me vs that on a recording played in my little Spendor s3/5 speakers?

Again, of course not. I do think it is implausible that you would perceive a difference of the same sound reproduced though the same Spendor s3/5's through two different reasonably-measuring amps.

 

[MattHooper]

I tend to push it to "fine for everyday situations, nowhere enough to support any claims that contradict established science".

Excellent! That's exactly the heuristic I myself endorsed earlier in this thread (and other threads)!

In virtually all cases, a scientific approach is the gold standard for raising confidence in a belief. But the fact that rigorously controlled tests are the MOST reliable doesn't entail all other forms of subjective observation and reason are therefore completely unrelieable, or unfruitful, or unreasonable. If that were actually true, we humans couldn't have made it this far, and couldn't get through the day successfully.

Of course not. But that doesn't take much. I can recognize the voices of my family in a very noisy environment over a very distorted, chopped and bandwidth-limited connection. That pretty much shows that our brain is actually very good at *not* detecting small distortions.

I think there's some red herring in there. (It's pretty amazing just how low-fi you can make a familiar voice and still recognize it. Did you grow up with dial phones? The tinny-little sounds coming from those phones didn't stop most people from recognizing familiar voices). But the point was that when there are objective differences between acoustic sounds in the known-to-be-audible range, and if they are pretty distinct or large enough, our auditory memory is quite good at cataloguing those differences reliably. The question, then, is if one is referencing a sound between "A" and "B" where the sonic differences would be both objectively measurably different AND in the audible range AND gross enough to be memorable.

Don't you agree?

Again, of course not. I do think it is implausible that you would perceive a difference of the same sound reproduced though the same Spendor s3/5's through two different reasonably-measuring amps.

Oh, I thought you were being skeptical of my impressions of live vs reproduced instruments? Maybe not then.

If instead you are talking strictly now about differences between amps:

I agree that it's implausible to hear a difference between two amps that measure "reasonably" so long as "reasonably" means "measuring closely enough for the difference to be inaudible." Because after all, there are ways for an amp to exhibit audible distortion.
The question is whether the tube amp in question alters the sound to an audible degree.

As I have said: my perception is my tube amps alter the sound vs the SS amp. But here we are in to the types of audible differences that can get in to the disputable realm, hence I'm quite open to the possibility it's due to sighted bias. But the objective and audible differences between a live trombone playing in front of me in the street vs that on a jazz recording through my little spendors (or frankly through any of my speakers) ought not to be very controversial.

Cheers.

 

[levimax]

The speaker is not a resistor so you'll have reactance which can cause the response changes which are not quite simple. So both Amir's and your results are possible.

OK while the ST-70 vs Neurochrome FR differences made sense to me it didn't make sense that adding 1 Ohm in series with the Neurochrome amp had very little effect on the FR. I went back and double checked and it turned out I was using a .25 Ohm resistor. I ran the test again with ~0.9 Ohm resistor and it acted as I expected. By adding the "simulated internal resistance" to the SS amp it almost exactly duplicated the FR of the tube amp. I believe that the vast majority of "tube sound" is just FR differences due to different internal resistances reacting with the speakers. Attached is results from REW. I am still going to try listening tests (both with and without the resistor) but right now after adding the resistor to the SS amp even sighted I can't tell them apart.

 

[rkbates]

OK while the ST-70 vs Neurochrome FR differences made sense to me it didn't make sense that adding 1 Ohm in series with the Neurochrome amp had very little effect on the FR. I went back and double checked and it turned out I was using a .25 Ohm resistor. I ran the test again with ~0.9 Ohm resistor and it acted as I expected. By adding the "simulated internal resistance" to the SS amp it almost exactly duplicated the FR of the tube amp. I believe that the vast majority of "tube sound" is just FR differences due to different internal resistances reacting with the speakers. Attached is results from REW. I am still going to try listening tests (both with and without the resistor) but right now after adding the resistor to the SS amp even sighted I can't tell them apart.

Very interesting experiment - keen to see how it all ends up

 

[TimF]

Marketing: I see a spdif (optical) cable advertised with very shiny gleaming gold connector tips. Silly that I imagine.

 

[levimax]

Sounds fascinating levimax. I admire your effort - it's more than I'd be tempted to put in.

Looking forward to the test results.

In my subjective comparisons, last night I preferred the solid state amp for Rachmaninoff's Symphonic Dances - tube amp sweeter and more lush, but a sense of laid back presentation/dynamics. Solid state amp seemed to wake the musicians up, more solidity, energy of string section etc seemed better.

Hi Matt: OK here is my shot at being subjective.... I have been listening to my "true mono" amp test set up (One speaker in the center). While down mixed stereo ranges from OK to weird to bad compared to two speaker stereo some of my old "pre-stereo" mono LP's from the 1950's sound fantastic on one speaker.... in fact I prefer them on one speaker to "two channel" mono. For simple arrangements originally mixed for mono, like Julie London singing accompanied by a guitar, I think the illusion of the musicians being in the room is much more believable on one speaker than on two speakers. While my system has a solid center image there is always some "brain power" spent creating the illusion. With one speaker in the center there is just a singer and guitar player in the center of the stage with no brain power being spent on creating a phantom center illusion. I find it much more realistic and easy to listen to.

 

[Blumlein 88]

OK while the ST-70 vs Neurochrome FR differences made sense to me it didn't make sense that adding 1 Ohm in series with the Neurochrome amp had very little effect on the FR. I went back and double checked and it turned out I was using a .25 Ohm resistor. I ran the test again with ~0.9 Ohm resistor and it acted as I expected. By adding the "simulated internal resistance" to the SS amp it almost exactly duplicated the FR of the tube amp. I believe that the vast majority of "tube sound" is just FR differences due to different internal resistances reacting with the speakers. Attached is results from REW. I am still going to try listening tests (both with and without the resistor) but right now after adding the resistor to the SS amp even sighted I can't tell them apart.

85% of hifi is frequency response.
In Bob Carver production ss amps with the transfer mod of a tube amp one of the things is a high output impedance.

And yes, level matching is job one. Level matching must be well matched not just close or they sound different.

 

[MakeMineVinyl]

OK while the ST-70 vs Neurochrome FR differences made sense to me it didn't make sense that adding 1 Ohm in series with the Neurochrome amp had very little effect on the FR. I went back and double checked and it turned out I was using a .25 Ohm resistor. I ran the test again with ~0.9 Ohm resistor and it acted as I expected. By adding the "simulated internal resistance" to the SS amp it almost exactly duplicated the FR of the tube amp. I believe that the vast majority of "tube sound" is just FR differences due to different internal resistances reacting with the speakers. Attached is results from REW. I am still going to try listening tests (both with and without the resistor) but right now after adding the resistor to the SS amp even sighted I can't tell them apart.

Tube amps behave in a more complex fashion than just being a less perfect voltage source than a solid state amp which in turn produces frequency response errors. They have a different distortion characteristic, have less crossover distortion from running richer into the class "A" window, and generally use much less negative feedback - not to mention that they have an output transformer between the output stage and the load which introduces its own complex soup of non-linearities. Of course the differences are audible, and in my case are desirable since I use tube amps (and active crossovers) on high efficiency speakers and a class "A" single ended triode on the HF horn.

 

[Julf]

It's pretty amazing just how low-fi you can make a familiar voice and still recognize it. Did you grow up with dial phones? The tinny-little sounds coming from those phones didn't stop most people from recognizing familiar voices.

Oh yes. As well as being a ham radio operator, and having used lots of very bandwidth-limited aviation radios and other R/T gear.

But the point was that when there are objective differences between acoustic sounds in the known-to-be-audible range, and if they are pretty distinct or large enough, our auditory memory is quite good at cataloguing those differences reliably.

Absolutely. The human brain is extremely good at extracting and learning patterns, especially related to survival or engagement with other humans. The problem, in audio, is that what we memorize is not the actual sound, but our pattern-matched perception of it.

The question, then, is if one is referencing a sound between "A" and "B" where the sonic differences would be both objectively measurably different AND in the audible range AND gross enough to be memorable.

Don't you agree?

I do. I think we are mostly in agreement - it is always the edge cases and definitions that cause issues.

Oh, I thought you were being skeptical of my impressions of live vs reproduced instruments? Maybe not then.

No, just skeptical about how it applies to comparing amps.

Cheers!

 

[levimax]

Tube amps behave in a more complex fashion than just being a less perfect voltage source than a solid state amp which in turn produces frequency response errors. They have a different distortion characteristic, have less crossover distortion from running richer into the class "A" window, and generally use much less negative feedback - not to mention that they have an output transformer between the output stage and the load which introduces its own complex soup of non-linearities. Of course the differences are audible, and in my case are desirable since I use tube amps (and active crossovers) on high efficiency speakers and a class "A" single ended triode on the HF horn.

Are you sure the differences you hear are anything more than FR differences? Good output transformers are quite linear in the audible range, crossover distortion is a "solved problem" in any well designed amp, lower feedback means higher output impedance with more FR error. More distortion is interesting.... I have been playing around with distortion generators on both speakers and headphones on both tones and music and unless it is very high (like 0.5% on tones and 2% on music) it is impossible for me to hear. I also never heard distortion, even 2nd order distortion, that sounded "better" than no distortion. FR errors on the other hand, especially "low Q" broad errors, are amazingly easy to hear. For some speakers and amp combinations I am sure higher output impedance alters the FR in a way that can be both more accurate (flat) and or more preferred but it introduces a lot of variables that are hard to predict and control.

I used to have a DIY tri-amp system with active crossovers MOSFET amp on bass, 2 triode mono block tube amps on mid, and class A on highs. It sounded great and when people saw it the psychoacoustic effect was powerful and I got a lot of compliments on the sound. It also heated up my room like crazy and required a lot of maintenance. After awhile for summers I switched out regular SS amps for the tubes and class A and in addition to being cooler it didn't really change the sound. I then started doing research and found this site and learned about room EQ and DSP and found I could accomplish a balanced and pleasant sound easier and cheaper and more repeatable with SS amps. I still play around with different amps for fun and am kind of disappointed that I no longer believe in "magic" amps. On the other hand I have shifted my focus on improving my speakers/ room which has yielded better results for me.

 

[MakeMineVinyl]

Are you sure the differences you hear are anything more than FR differences?

As an amp designer, I am sure. Although I've never seen a differing opinion altered by conflicting views on this forum. In the words of a well known philosophiser, "it is what it is".

 

[levimax]

As an amp designer, I am sure. Although I've never seen a differing opinion altered by conflicting views on this forum. In the words of a well known philosophiser, "it is what it is".

Yea life is too short to worry about what other people believe.... I am just trying to figure out what I believe. In any case it is a fun hobby... enjoy.

 

[Blumlein 88]

Yea life is too short to worry about what other people believe.... I am just trying to figure out what I believe. In any case it is a fun hobby... enjoy.

One thing I like is this forum can challenge your belief without being an in your face test of manhood. At least it does so more often than most. Look at evidence, give yourself a chance by testing blind if you will, and you might find your belief was off the mark. Which might lead to a change of belief.
You aren't obligated to change it, and you don't need others approval if it differs, but life is sometimes interesting when you learn new things, no? If you can change your attitude from your belief being so personal to being somewhat disengaged from your sense of self it can be enjoyable even.

 

[Frank Dernie]

One thing I like is this forum can challenge your belief without being an in your face test of manhood. At least it does so more often than most. Look at evidence, give yourself a chance by testing blind if you will, and you might find your belief was off the mark. Which might lead to a change of belief.
You aren't obligated to change it, and you don't need others approval if it differs, but life is sometimes interesting when you learn new things, no? If you can change your attitude from your belief being so personal to being somewhat disengaged from your sense of self it can be enjoyable even.

This is very true.
It wasn't this forum that made me sceptical of "common wisdom" in the last few decades of hifi magazines, but being an engineer.
I looked at what was being written and it seemed daft to me.
I had had success in my career by being sceptical of "common knowledge" and working things out from first principles and, luckily, been encouraged rather than smacked down at University so I decided to apply the same approach to my hobby once I had retired and had some spare time for the first time in 40 years.
When I was first buying some hifi, in the late 1960s all the magazines were pretty technically focused but I had noticed the drift towards subjective reviewing, which seemed fair enough until applying a bit of first principle to what was being promulgated. It did not bear scrutiny.
The first seemingly ridiculous claims were, of course, cables. It seemed self evident that as long as they make good contact and are adequate in impedance there is no mechanism whereby a cable could have any influence at audio frequencies. Since I had been working long hours for good pay I had ended up with some expensive cables supplied with preamps, amps and CD players. I decided to do blind tests myself and found what I expected, no difference except thiose with in line filters.
Next I tested a few loudness levels. My amp has a volume control display in dB so it was fairly easy.
I found that if I reduced volume from my normal listening level by the time I had reduced it by 60dB I could only just barely hear it.
From this I deduced, perhaps wrongly but IMO logically, that since I could barely hear it on its own a distortion level that low would certainly be inaudible given that the music would be playing at normal levels at the same time.
I now consider the idea that distortion levels need to be better than -60dB, or 0.1% to be not credible.
Lets face it there are loads of hifi enthusiasts who think LPs sound best and they are nowhere near that good.

Anyway the usual suspect hifi forums were not generally, with a few member exceptions, receptive of pretty well any technically based comment so it was with great joy I joined this forum after Amir suggested it may suit me following some facts I wrote about record players which were unwelcome on another forum.
So here I am!

 

[Julf]

As an amp designer, I am sure.

Just curious, what is your sureness based on? I do wish a card saying "amp designer" would make me technically and factually infallible, but unfortunately it doesn't seem to work that way (just witness some hot-shot "amp designers" spewing absolute BS).