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

[watchnerd]

Disclosure: my university education was in applied physics and signal processing, and I currently work in the software industry, so my knowledge of electrical engineering is not zero, but I'm no EE. So apologies in advance for any technical errors.

This thread is inspired in part by comments I've seen from @SIY , and others, that there is no such thing as "tube sound", per se, and that properly designed tube amps should be neutral.

Which then raises a number of questions, in my mind:

--If tube based amps can be made to avoid "tubey sound", why bother with tubes instead of solid state?

--If a tube based amp is designed to avoid "tubey sound", what impact does that have on tube rolling?

--If a tube based amp is designed to avoid "tubey sound", does it have any impact on the output impedance, current, or other factors that differentiate tubes from solid state?

--Does it matter at all if there are tube rectifiers in the circuit?

--Is there a reason to prefer transformers to boost converters?

--Between driver, power, and rectifier tubes, which are have the most impact on the sound, assuming a "non tubey" design goal?

--If tubes don't have a sound, what's the point of hybrid tube amps?

 

[hetzer]

Like other senses, our hearing gets compensated by other information that brain has. There are little muscles below our hair cells that sense the sound. The brain use this muscle to hear what it wants to hear more clearly. This is why we can talk to other people in noisy environment.
When we see a music coming out from tube amp, our brain emphasizes certain sounds in accordance with our bias. Someone who thinks that tube amp sounds 'warm' will focus on bass. Someone who thinks that tube amp has its own 'texture' or 'harshness' will focus on highs.
This also means that no matter you intend to make tubey sound or not, if there is a tube or analog VU meter on top of the amp, people will 'hear' tubey sound. The designer's choice will be the color of LED to make tube glow.

 

[watchnerd]

Like other senses, our hearing gets compensated by other information that brain has. There are little muscles below our hair cells that sense the sound. The brain use this muscle to hear what it wants to hear more clearly. This is why we can talk to other people in noisy environment.
When we see a music coming out from tube amp, our brain emphasizes certain sounds in accordance with our bias. Someone who thinks that tube amp sounds 'warm' will focus on bass. Someone who thinks that tube amp has its own 'texture' or 'harshness' will focus on highs.
This also means that no matter you intend to make tubey sound or not, if there is a tube or analog VU meter on top of the amp, people will 'hear' tubey sound. The designer's choice will be the color of LED to make tube glow.

In that case why not just put VUs and non-functioning tubes (just a heater element, no audio circuit) as a facade on a class D amp?

 

[PaulD]

Without bothering to answer all of the questions above individually, it is fair to say that a competent design can be achieved with various devices. So it should not matter what technology is used in the rectifier, output stage or input stage, but if the designer is not so expert then the characteristics of the devices can affect the performance, perhaps in a way that can be audible.

Some valve amplifiers sound like the preconceptions and some people like that, it can even be advantageous in some systems with other nonlinearities that are either compensated for or avoided or covered by the poor amplifier.

Similarly it is of little importance if the devices are BJTs or JFETs or MOSFETs IF the design is competent, then it will be transparent.

Never underestimate the power of marketing and cognitive biases with amplifier sales.

 

[watchnerd]

Without bothering to answer all of the questions above individually, it is fair to say that a competent design can be achieved with various devices. So it should not matter what technology is used in the rectifier, output stage or input stage, but if the designer is not so expert then the characteristics of the devices can affect the performance, perhaps in a way that can be audible.

So if the technology shouldn't matter, why bother using bigger, hotter, more expensive tubes, instead of ICs?

 

[Thomas savage]

https://audiosciencereview.com/forum/index.php?threads/technically-what-exactly-is-tubey-sound.8215/

 

[PaulD]

So if the technology shouldn't matter, why bother using bigger, hotter, more expensive tubes, instead of ICs?

Marketing

 

[DownUnderGazza]

Nostalgia, bragging rights, buying into audiophile community group-think...

I once came across a quote that basically said, the more you have personally emotionally and financially invested in a certain belief, the harder it is to accept you were wrong and thus relinquish that belief.

Basically if you have extolled to others the benefits of tubes, cables, cable-lifters, vinyl, and invested your own hard earned cash in the same, the harder it is to admit you were a fool and wasted your money.

 

[SIY]

--If tube based amps can be made to avoid "tubey sound", why bother with tubes instead of solid state?

--If a tube based amp is designed to avoid "tubey sound", what impact does that have on tube rolling?

--If a tube based amp is designed to avoid "tubey sound", does it have any impact on the output impedance, current, or other factors that differentiate tubes from solid state?

--Does it matter at all if there are tube rectifiers in the circuit?

--Is there a reason to prefer transformers to boost converters?

--Between driver, power, and rectifier tubes, which are have the most impact on the sound, assuming a "non tubey" design goal?

--If tubes don't have a sound, what's the point of hybrid tube amps?

In order:

1. Fun. It's enjoyable for some (like me) to use old technologies and see what performance can be scraped out of them. Think of someone restoring a 1959 Avanti. I also take advantage of the high available voltage swings for things like mike and phono preamps as well as electrostatic direct drive.

2. I think the notion of "tubey sound" is bullshit. Bad tube designs can be bright, dull, fuzzy, overly sharp, whatever. A bad amp is a bad amp, a good amp is a good amp, whatever active devices are being used. Tube rolling is something to keep technical ignoramuses busy, much like cable swapping or re-capping. You don't need to know anything, you just have to have a credit card.

3. The whole point of making a good tube amp is to get a low output impedance, sufficient current/voltage swing for the intended load, and anything else that defines a good amp, irrespective of active devices.

4. Yes. There may even be advantages for a guitar amp given the voltage sag under load. For audio reproduction, tube rectifiers are at best benign and at worst a severe performance degradation.

5. All of them, but most of all, output transformers are the key to getting rid of the amp's effect on sound.

6. See number 1.

 

[Sancus]

In that case why not just put VUs and non-functioning tubes (just a heater element, no audio circuit) as a facade on a class D amp?

They're called "hybrids" and yeah, when implemented well, that's pretty much how it ends up.

 

[Thomas savage]

Personally I bought some tube amps/pre's because I read about them being so much better than their nasty modern SS equivalent.

Turned out this was bullshit ,at least with the stuff I bought ( Audio Research) sounded quite thin and not at all 'tubey' . The amp bored me to death and the pre did nothing extra va a SS , So no expanded soundstage etc.

They looked kinda cool though.

 

[audimus]

There is one school of thought that an amp should just be transparent in that what comes out should be a perfect replica of what goes in as indicated by various measurements. The closer you get to this asymptotically, less it matters how it is achieved since the thesis is that you cannot really tell the difference between them with well-designed listening tests. This camp does not claim there is no “tube sound”. But holds that the “tube sound” comes from deviations from the perfect replica of sound and so not of interest (to them).

There is another school of thought that the goal isn’t a perfect replica but rather to achieve what is perceived as more pleasant or musical or less fatigue-inducing or emotionally satisfying for them individually. The measurements would only be relevant to the extent that they don’t indicate distortions or characteristics that would lead to less of those desired goals, like clipping, gross non-linearities, etc. Achieving perfect replicas is not the point at all.

Both are perfectly legitimate schools of thought even if they vehemently don’t agree with each other and go to extreme lengths to invalidate the other which is largely a futile and a “religious” argument (because of the differences in the axioms of each school, rather than the methodology).

In an ideal world, both would co-exist peacefully but are drawn into conflict when realities of the marketplace force manufacturers to claim that their equipment simultaneously satisfy both those schools of thought and therein lies the problem.

 

[SIY]

They're called "hybrids" and yeah, when implemented well, that's pretty much how it ends up.

Here's my hybrid phono stage ("The Equal Opportunity") IM at 1.3V out.

 

[Blumlein 88]

In that case why not just put VUs and non-functioning tubes (just a heater element, no audio circuit) as a facade on a class D amp?

I remember reading how as a grad student J_J had people come and listen to tube vs SS amps. The tube amp was able to light up the tubes, and looked functional, but was never used. It was SS both times. Yet the tube sound was heard until one of the instructors caught on.

 

[PaulD]

I remember reading how as a grad student J_J had people come and listen to tube vs SS amps. The tube amp was able to light up the tubes, and looked functional, but was never used. It was SS both times. Yet the tube sound was heard until one of the instructors caught on.

I've done similar things several times, once all I changed was the front LED from blue to yellow - most listeners (sighted and uncontrolled) said it sounded warmer...

 

[Blumlein 88]

My thinking 20-25 years ago was different from the mainstream in the audiophile world. It was often suggested if you needed power, but liked tubes, a tube preamp got you more than halfway there. I never found that to be true. I didn't find it to get you at all there in fact.

Tube amps did sound different. But I was using ESL's. Very high impedance in the lower 200 hz (like 15-40 ohms), and very low impedance in the upper octave and a half (like an ohm or two). A good push pull triode with a moderately high output impedance (maybe SIY says bad design) would elevate the low end enough to notice, and soften the treble enough to notice. I could measure that. It wasn't just bias, it really sounded different. Into my ESLs it is also possible the tube amp at least on peaks had some audible distortion that wasn't obvious enough to sound like distortion.

Now I thought more than that sounded different. Some of the sense of space, a 3D quality SS amps didn't have. I did my own tests at one point like what the Swedish AES did. Load an amp output with resistors, drop the level enough that input to output was unity gain, and feed that into another amp powering the speakers. At the time I did this I thought tube amps did things SS couldn't do in the space and 3D area. I was using a VTL and a Spectral. To my surprise, the VTL fed into the Spectral exhibited all the space and 3D qualities which meant whatever the triode tubes were doing was captured accurately by he Spectral. It was triodes being of superior fidelity, it was triodes having a pleasing coloration. This is not what I was expecting at all. Reversing positions the Spectral had all the coloration of a meter of interconnect. You could tell if it was just interconnect to the amp or going thru the Spectral. In time I tried this with a few other amps.

Now this wasn't blind listening on my part. I did do this for one of my audiophile buddies. I switched with him out of the room several times. He could within seconds correctly tell me if the VTL was in circuit or not. He was not able to reliably tell as to whether the Spectral was in circuit, and took longer to answer. Still single blind, and a couple minutes to switch. Plus I know the FR was different.

In any case, people who want tube sound or believe in it, could use low powered tube amps like preamps feeding them into clean SS amps to have their cake and eat it too. I've even had in mind making a preamp with the right character. Using 12AT7s in push pull circuits able to deliver a watt or so, transformer coupled and loaded by resistors on the output to be like a mini push-pull triode amp. You could even switch in a few elements to simulate loudspeaker loading. Probably should have done it years ago without saying what I was doing. Come up with some marketing story to sell.

 

[PaulD]

There is a solid amount of high-quality valve modelling in professional audio tools, from modelling amplifiers (hardware DSP) for guitarists, to modelling plugins etc for guitar amps and effects (LA2A compressors etc). The best ones essentially run something like SPICE in realtime and they have been refined over the last decade to be indistinguishable from the real thing (yes, I've been involved in blind testing). I can imagine this making it into hifi before too long so people can choose it easily if they wish.

 

[audimus]

I can imagine this making it into hifi before too long so people can choose it easily if they wish.

I agree and my vision goes one step further. People will not have to choose different amps based on taste, they will be able to configure the modeling inside the amp using an included app to make it sound “right” for them.

The app will be like the audio equivalents of the image editing tools of today. Start with current in the middle and they can tap in one of many directions to add more color so to speak or less bright or warmer etc until they settle on what sounds best for them. There will even be multiple profiles for each member of the family and for each type of content.

Mark my words, I prophecied it first.

The people who like their clean perfect sound replication amps of today will be relegated to the back pages of history and niche markets just like the vintage tube amps crowd today.

It will be the ultimate triumph of subjectivism enabled by objective scientific method.

 

[Wombat]

My thinking 20-25 years ago was different from the mainstream in the audiophile world. It was often suggested if you needed power, but liked tubes, a tube preamp got you more than halfway there. I never found that to be true. I didn't find it to get you at all there in fact.

Tube amps did sound different. But I was using ESL's. Very high impedance in the lower 200 hz (like 15-40 ohms), and very low impedance in the upper octave and a half (like an ohm or two). A good push pull triode with a moderately high output impedance (maybe SIY says bad design) would elevate the low end enough to notice, and soften the treble enough to notice. I could measure that. It wasn't just bias, it really sounded different. Into my ESLs it is also possible the tube amp at least on peaks had some audible distortion that wasn't obvious enough to sound like distortion.

Now I thought more than that sounded different. Some of the sense of space, a 3D quality SS amps didn't have. I did my own tests at one point like what the Swedish AES did. Load an amp output with resistors, drop the level enough that input to output was unity gain, and feed that into another amp powering the speakers. At the time I did this I thought tube amps did things SS couldn't do in the space and 3D area. I was using a VTL and a Spectral. To my surprise, the VTL fed into the Spectral exhibited all the space and 3D qualities which meant whatever the triode tubes were doing was captured accurately by he Spectral. It was triodes being of superior fidelity, it was triodes having a pleasing coloration. This is not what I was expecting at all. Reversing positions the Spectral had all the coloration of a meter of interconnect. You could tell if it was just interconnect to the amp or going thru the Spectral. In time I tried this with a few other amps.

Now this wasn't blind listening on my part. I did do this for one of my audiophile buddies. I switched with him out of the room several times. He could within seconds correctly tell me if the VTL was in circuit or not. He was not able to reliably tell as to whether the Spectral was in circuit, and took longer to answer. Still single blind, and a couple minutes to switch. Plus I know the FR was different.

In any case, people who want tube sound or believe in it, could use low powered tube amps like preamps feeding them into clean SS amps to have their cake and eat it too. I've even had in mind making a preamp with the right character. Using 12AT7s in push pull circuits able to deliver a watt or so, transformer coupled and loaded by resistors on the output to be like a mini push-pull triode amp. You could even switch in a few elements to simulate loudspeaker loading. Probably should have done it years ago without saying what I was doing. Come up with some marketing story to sell.

I think the 'tube sound' is generally related to low cost tube power output stages and simple tube power supplies, in particular SETs.
This is truly exploited in traditional guitar amps. That is not to say that guitar amp designs can't be complex in a search for a "clean"(not hifi) tone or stable 'high gain distortion' generating amps. This is one area I got theoretically and practically involved in for some time.
I have never considered a tube amp for HiFi - to many practical downsides.

The Williamson amplifier gives an idea of what a high performing tube amplifier design involves. SS easily outperforms it.

http://www.google.com.au/url?sa=t&rct=j&q=&esrc=s&source=web&cd=13&ved=2ahUKEwiPlKOk6pXkAhWlILcAHcfFCL0QFjAMegQIAhAC&url=http://www.r-type.org/pdfs/dtnw-amp.pdf&usg=AOvVaw2blbCmxSla8Qd4krVyoAc6

 

[PaulD]

The people who like their clean perfect replica amps of today will be relegated to the back pages of history and niche markets just like the vintage tube amps crowd today.

Maybe! I can see a time when "perfection" in the electronics of audio is a commodity that we don't value it any more, so manufacturers design for "taste"... It will be a sad day for me, as I do value being able to accurately reproduce what was recorded, but I'm probably an unusual case, and an endangered species.