Are tubes more musical?

[Chr1]

Personally I think that it is all about the (evil) added distortion/harmonics... It ain't hifi, but it sounds good. Here's the quandary.
My 2c

 

[pablolie]

...
Want a tube because they feel it sounds different from a solid state amp: is going to colour the sound in some pleasing way.

It *is* going to be less true to the recorded original compared to a good solid state amp, as far as accuracy according to several measurable parameters. Whether or not it is pleasing or not is up to the person using the amplifier. If someone tells me they prefer the sound of a tube amp, all the power to them. If someone says tube amps are universally technically superior, I will not engage in further discussion. :-D

But then they go looking for speakers that are “ as tube amp friendly as possible.” Which means among other things higher sensitivity and a benign impedance for a tube amp to drive.

I don't think it is contradictory. Some combos of tube amps and speakers sound actually pretty sweet (and I mean that literally) [1]. Again, whether that's a strength or a weakness is up to the buyer to establish. But I'll start smirking if someone tells me the combination is magically revealing and technically superior.

But this would seem to mitigate some of the very ways in which a tube amp could change the sound! For instance if you’ve got a very easy to drive speaker, doesn’t that suggest that you’ve done away with some of the impedance variations that would’ve differentiated the tube from a solid amp in the first place?

"Easy to drive" means high sensitivity, i.e. more output for less power. It says nothing about the quality or linearity or accuracy of the output. Which means you may end with an overlay of mutual weakness between the amp and speakers. Or a magically pleasing combo for the type of music you personally like. In general, the matching of an amp and a speaker becomes more unpredictable the lower the damping factor of the amplifier and the higher the sensitivity of the speaker.

I personally prefer to listen to recordings as true to the original as possible. I could not care less how sensitive a speaker is, since these days you have solid state amp designs that drive/control any speaker to sound like its designer intended it to sound. And if they didn't, at least IMO they should publish a list of amps their speaker works well with as they tested (and ideally measured) the sound of the combo [1].

But I found the combination glorious. Versus a solid state amp, the clearly seem to have trouble keeping a grip on the lower frequencies of the speakers. But the result was a very “ ripe” there was actually quite appealing to me, and gave me the sensation the bass was going lower and they were performing slightly like larger speakers.

Generally, it is in the lower frequencies where mismatches in DF and sensitivity and linearity become quite hearable. But there is potential impact all over the frequency spectrum. I fact the latter is quite likely. Whether some prefer it or not does not really matter. The term "HiFi" demands accuracy. I have zero issue with someone preferring "ArtFi", a coin I have termed with a great friend of mine who prefers it and is actually extremely proud of.

[1] Incidentally that makes me wonder why no one as design an integrated, active speaker based on tubes - the designer-engineer ought to be the best to match things optimally. If anyone were to try to establish the superiority of tubes, that'd be the way to do it, I guess.

 

[pkane]

According to Paul, he had a module that could do the DvsF thing, but it was not part of the release.

It's not part of the released version of DISTORT mostly because I am reluctant to make something, less than user-friendly, public.

The feature is in my internal build, so I can configure it to simulate common distortion vs frequency shapes to create the desired transfer function, and then apply it to any music track. I'll need the 1kHz distortion plot where I can read the harmonic levels, and a 20-20k THD vs. frequency sweep. Here's an example of an REW measurement sweep sent through a DISTORT transfer function, simulating a tube amplifier and showing variation of THD by frequency:

 

[pablolie]

I'm not sure tube hi-fi amps would sound particularly nice when forced to operate well outside anything approaching linearity. The kinds of distortion that sound pleasing when applied to an electric guitar signal through very bandlimited speakers are going to sound very messy and nasty with a complex music signal through a full-bandwidth audio system.

I agree "standardized" testing protocols may not always tell the complete story to potential consumers. As in, measuring distortion at maximum output may be meaningless to some/many users, for example. Try that with a Purifi's 400W+ and let us know what your ears, neighbors and melting speakers think about it... One needs to read through measurements and be aware of what matters to their particular use case... and what doesn't.

 

[atmasphere]

But this would seem to mitigate some of the very ways in which a tube amp could change the sound! For instance if you’ve got a very easy to drive speaker, doesn’t that suggest that you’ve done away with some of the impedance variations that would’ve differentiated the tube from a solid amp in the first place?

No- they are taking advantage of what the tube amps offer, which aren't good at low impedances or crazy phase angles. Its sort of like buying a fine British sports car from the 1950s and then putting bad oil in the engine and chessy tires. IOW you give it the best conditions to show its stuff.

It's not part of the released version of DISTORT mostly because I am reluctant to make something, less than user-friendly, public.

The feature is in my internal build, so I can configure it to simulate common distortion vs frequency shapes to create the desired transfer function, and then apply it to any music track. I'll need the 1kHz distortion plot where I can read the harmonic levels, and a 20-20k THD vs. frequency sweep. Here's an example of an REW measurement sweep sent through a DISTORT transfer function, simulating a tube amplifier and showing variation of THD by frequency:

Sweet!!
Can you take this distortion spectrum and then model a 6dB/octave rise at 1KHz?

 

[SIY]

No one keeps a tube amp just because it sounds as good as a solid state one (unless they also need heating).

Ahem.

 

[pkane]

Sweet!!
Can you take this distortion spectrum and then model a 6dB/octave rise at 1KHz?

Can you suggest an actual amplifier with the measured distortion vs frequency? I'd prefer to simulate something that actually exists

 

[SIY]

Can you suggest an actual amplifier with the measured distortion vs frequency? I'd prefer to simulate something that actually exists

And that is claimed to be "bright."

 

[Multicore]

I'm not sure tube hi-fi amps would sound particularly nice when forced to operate well outside anything approaching linearity. The kinds of distortion that

I was talking about compression, not about clipping. And the kind of compression that results from power supply sag is not unlike the kinds used in mastering to make our precious recordings sound oh so nice.

 

[pablolie]

And that is claimed to be "bright."

Well, I can understand why something linear sounds "bright" to someone used to an inverse bathtub response. I put zero value on such personal perceptions. Show us the measured frequency response.

 

[DonH56]

"Are tubes more musical?"​

Only when you tap them with a spoon.

 

[atmasphere]

Can you suggest an actual amplifier with the measured distortion vs frequency? I'd prefer to simulate something that actually exists

I get that. When I repaired a pair of early 2000s Adcom GFA656s about 7-8 years ago, when I went online I found a measurement that showed a rise in distortion at 1KHz. It stuck in head because I also remember being impressed that the measurement was published. Now searching on that amplifier I can't find it. I've been gaslighted over this and I have a business to run; right now I think the spec I saw was when the amp was bridged, but the current amp I have to conclude is better than ones from 24 years ago.

It is my contention that this property is audible even when THD is low. SIY is trying to corner me into saying a particular amp and indeed the Adcom came off as bright in our system. Some might say that is due to the new boards not being broken in, but my surmise is the amp employs enough feedback that break-in isn't a thing. I've heard lots of systems that were bright over the years because I've been doing audio shows since 1989. But I've not kept track of which amps were in those rooms as I discounted them out of hand. I do remember a Symphonic amp the size of a steamer trunk that came off that way. I doubt its been measured though.

 

[Jimbob54]

PS (and with some trepidation) - I have been very favorably impressed by the ongoing dialog between @SIY and @atmasphere and I hope that it will not escalate into a (ahem) urinating contest. So many passionate mavens in this business are incapable of intelligent and focused debate on issues, but - from my perspective as a semi-informed amateur - these two have historically proven to be far, far better in that regard than the mean (or the median, for that matter).

I second that and perhaps those esteemed gents would perhaps like to dial back the fervour and tone of their debate a bit from recent posts. Much obliged.

 

[atmasphere]

Well, I can understand why something linear sounds "bright" to someone used to an inverse bathtub response. I put zero value on such personal perceptions. Show us the measured frequency response.

If this were the case I would expect that too but many tube amps have bandwidth higher than most class D amps. So its a pretty good bet that bandwidth isn't why.

 

[pablolie]

If this were the case I would expect that too but many tube amps have bandwidth higher than most class D amps. So its a pretty good bet that bandwidth isn't why.

I dunno, input the word "tube" in the electronics review index in here... kinda hard to discern stuff when the SINAD is that low.

 

[Blumlein 88]

I get that. When I repaired a pair of early 2000s Adcom GFA656s about 7-8 years ago, when I went online I found a measurement that showed a rise in distortion at 1KHz. It stuck in head because I also remember being impressed that the measurement was published. Now searching on that amplifier I can't find it. I've been gaslighted over this and I have a business to run; right now I think the spec I saw was when the amp was bridged, but the current amp I have to conclude is better than ones from 24 years ago.

It is my contention that this property is audible even when THD is low. SIY is trying to corner me into saying a particular amp and indeed the Adcom came off as bright in our system. Some might say that is due to the new boards not being broken in, but my surmise is the amp employs enough feedback that break-in isn't a thing. I've heard lots of systems that were bright over the years because I've been doing audio shows since 1989. But I've not kept track of which amps were in those rooms as I discounted them out of hand. I do remember a Symphonic amp the size of a steamer trunk that came off that way. I doubt its been measured though.

Well SIY is trying to get you to see if you can hear this brightness. Kind of hard to develop something reasonable without a target. What % at 1 khz rising by 6 db per octave would you want to simulate. The Adcom 565 which is listed as .004% at 1 khz at rated power both bridged and not bridged? Stereophile measurements of that amp in 1991 show .003 % at 1 khz at rated power so the amp likely is the same basic amp.

 

[Overseas]

As long as musicality is not defined, then YES, tubes are the most musical.

 

[atmasphere]

I dunno, input the word "tube" in the electronics review index in here... kinda hard to discern stuff when the SINAD is that low.

Citation needed.

Well SIY is trying to get you to see if you can hear this brightness. Kind of hard to develop something reasonable without a target. What % at 1 khz rising by 6 db per octave would you want to simulate. The Adcom 565 which is listed as .004% at 1 khz at rated power both bridged and not bridged? Stereophile measurements of that amp in 1991 show .003 %.

0.004 but look at what happens at higher frequencies when bridged- up by a factor of 5 at 10KHz.

'Brightness' is a thing that has kept businesses like mine going as such businesses/amplifiers provide an alternative without loss of bandwidth. Its a bit pesky when something sounds bright but you put it on the bench and FR is flat. So something else is afoot. I explained that earlier.

Whether I can hear it or not is a different thing since my hearing is so degraded from only 7-8 years ago. But if it gets too bright I still don't like it so I think I can. Obviously we're all trying to sort this out- If I'm wrong and distortion rise isn't what causes the brightness that keeps tube amp manufacturers going, it will be a data point and so other causes will have to be investigated.

SIY does not need to resort to personal attacks to move this along.

 

[DLS79]

Obviously we're all trying to sort this out- If I'm wrong and distortion rise isn't what causes the brightness that keeps tube amp manufacturers going, it will be a data point and so other causes will have to be investigated.

I'd assume they are still around, because some people (like with Nixie clocks), like the way they look. I assume some people think they like or like the way they sound.

 

[DanielT]

In that case I'm putting the Tiger comment up to a communication error. As I recall it was a painful amp.

cut 1: Algunas Bestias
cut 2: Sandino

Its been my theory that above 10KHz a distortion rise with frequency won't be heard for obvious reasons. So this has my interest piqued.

Citation needed; this appears a highly speculative and a rather subjective observation. FWIW, SET users (Single Ended Triode amps typically are zero feedback) commonly describe SETs are being more dynamic rather than less. Its my contention this is caused by how the amplifier distorts and how that in turn skews the ear's mechanism for sensing sound pressure (IOW, its an illusion). Actual, real dynamic qualities reside in the recording and are not actually a function of amplifiers whether tube or solid state.

'Rolled off response' makes no sense as many tube amps have wide bandwidth; that was a big deal with Stu Hegeman, the designer of early HK tube equipment including the Citation series. The Citation 2 of this thread has bandwidth that often goes higher than the published specifications, which can be regarded as nominal (I just put my Citation 2 on the bench and it easily made full power to 70KHz). The Citation 2 BTW employs enough feedback that it is able to act as a Voltage source as long as the load impedance does not get too low so FR variation is not a variable. Tube headphone amps tend to have much wider response and headphones tend to be a very easy load.

I commented earlier on the subjective quality of 'dynamics'.

A simple way to answer than of course is to obtain a good tube amplifier and find out. IME a side by side demo easily shows the difference to anyone in the room as its usually not subtle. The speakers used for such a demo should be of the kind that both the control amplifier and the tube amp both find to be an easy load so FR variation can be ruled out.

IME, these are the differences, which are easily measured:

1) more prodigious 2nd and 3rd harmonic than seen in most solid state amps; which the ear interprets as 'warmth'. I know some people think this is because of a hump in the bass or some such but if you put such amps on the bench you see there isn't anything like that unless the amp lacks feedback. If it lacks feedback the amp is meant to be used on a speaker that is designed with that in mind to prevent coloration.
2) A 'smoother more relaxed sound'; my theory being that the 2nd and 3rd harmonics are masking higher orders in the same way as seen in loudspeakers.
3) The reduced generation of higher orders at clipping, which allows for a graceful overload characteristic. Quite often people are overloading their tube amps and don't even know it. Guitar players in the past are well-known for using tube amplifiers for exactly the purpose of overdrive; that is how Marshall, Fender, Ampeg and so on made their names.

Once you really understand these facts it allows an access for designing and building a solid state amp that does the same thing should that be desired. You can even get the 3rd aspect of soft clipping if SITs are used in the output section. Its highly likely that a multi-ordered feedback loop will be needed and care in the circuit design should be taken to minimize distorting the feedback signal at the feedback node.

Actually, I only got hooked on a small thing in Audiophonics sales text. That is,Audiophonics wrote Sound tube instead, as is customary as far as I know, to say tube sound. I admit that was a rather silly detail to pick up.

I follow what you and other technical amplifier experts write the thread so thanks for your detailed answer. Say what you want, but this thing with tube amps and their being or not always takes off.