I know a little about tube guitar amps, and my understanding is that they are desired due to their “defects” compared to more accurate solid state transistors. I would assume headphone amps would want pristine amplification— why would you want tubes, then?

The slight touch of harmonic distortion from a tube amp can sound nice, even though it might not be totally transparent. Tube amps are also a lot of fun to build and tinker with because of the huge diversity in tube models, the operating point at which the tubes are run, the gain stage topology, power supply designs, etc. Personally I find that different tube designs almost always sound a little different from each other while well-designed transistor amps sound indistinguishable or almost indistinguishable from one another.

Pete Millett, who runs a DIY Tube site, made a good presentation on this: http://www.pmillett.com/etf_sod.htm

> Tube amps are also a lot of fun to build and tinker with because of the huge diversity in tube models, the operating point at which the tubes are run, the gain stage topology, power supply designs, etc.

You could say all of that about FETs and regular transistors too.

Yeah that is true, but with tube amps you can tinker around with the circuit and actually hear a meaningful difference, which is more difficult with solid state amplification IME

Those "defects" is just non linearity in the circuit. At low gain factors (much lower than a guitar amp) it creates prominent 2nd, 3rd, and 4th harmonic distortion (depending on the circuit) which corresponds to the musical intervals of an octave, octave + fifth, and 2nd octave. That sound is typically described as "warmth" and can be quite pleasing. It also increases the RMS of the signal without (greatly) affecting dynamics, which makes it perceptually louder.

Some people like the ultra hifi systems with crazy THD+N metrics, with super flat speakers in the deadest room they can find to get the most "pristine" sound they can. Other people like listening through systems that subtly color their favorite recordings to match their taste. Different strokes for different folks.

Tube amps (as do transistor amps) can vary quite a bit in how much they change the sound, e.g. part of the appeal of tube guitar amps is that you can drive them out of the area they work "most precise" into distortion, and that distortion sounds interesting. And some people like the sound changes tube amps cause for listening.

There is an story that when the Beetle's went to record at the BBC there was an argument with the BBC sound recording Engineer that no way was he allowing distortion to go out on the BBC :-)

I am no EE, but transistors also have problems giving precise accurate gains as well.

That depends on how much energy you want to waste.

https://www.electronics-tutorials.ws/amplifier/amplifier-cla...

The age old question: besides the obvious aesthetic attractiveness, is there really any kind of objective advantages to tube-based audio amps as opposed to solid-state semiconductor based ones?

Technically speaking there is zero advantage in using tubes instead of solid state when the amp is driven below distortion, that is for listening to music at normal level, and anyone claiming they can hear the difference between a decent 100 bucks class D amp and the most expensive tube amp in the galaxy is lying (probably also to themselves if they bought the latter). However, when tubes are driven to distortion, the waveform is clipped in a much gentler way compared to what is produced by solid state parts, so in a Hi Fi amp those high power peaks would indeed sound a lot cleaner, and in a guitar amp where distortion is looked for by design, using tubes would would produce a much nicer clipped sound, which is even more audible when playing chords. In HiFi amps the problem can be easily solved by designing (or buying) for more power than what is needed, so that the headroom reserve will be used to cope with those peaks without getting into clipping, while using tubes for guitar amps apparently still makes sense. Technology however has progressed a lot since the old days, most parts have become a lot cheaper, and even without resorting to digital manipulation todays solid state pure analog guitar amps can sound great without using tubes.

One important point though: when making guitar amps with newer technologies such as class D, we must redesign everything so that the amp -I mean the power module- will never clip; no matter what we throw at the amplifier main input, the power module must never be driven into distortion because class D distortion is plain awful. That requires a lot of redesign: no more clipping tubes, coils driven into saturation, purposely weak designed power supplies etc, but simply speaking we distort before and keep the level put in order not to distort after. This makes things even more interesting because one builds the guitar sound much before it gets to the final stages, the only further contribution to the sound being the speaker and the cabinet, not the amp module.

In audio most objectivists believe that currently known measurements (frequency response, power handling abilities, THD, PSSR, other non linear distortions, noise floor, etc.) can 100% conclusively define subjective sound. But then on the flip side there are very reputable people (eg. Paul McGowan from PS audio) agree that we may not have figured out all measurements necessary to nail down subjective audio quality. It is upto the ears of the listeners and their exposure to super high end audio to decide whether that is true.

I fall in the camp that thinks we haven't fully figured out what conclusively decides subjective sound quality. It makes things much more interesting and open to more research to settle debates on solid state vs tube, sigma-delta vs r2r dacs, full range driver vs other speaker configurations, cables/interconnects/power-cords matter or not and things like that).

Some interesting links (I don't endorse what these guys are saying but find them VERY interesting):

About DACs: http://www.streamplayer.io/v1/sharing?v=8Mn5PrnZV-k

About whether audio power cables make a difference (Ask Paul from PS audio): http://www.streamplayer.io/v1/sharing?v=8QuToO9JUfw

High end equipment makers have a vested interest in promoting pseudo science because it duped people into spending $5k for a DAC (e.g., PS Audio). Audio frequency range electrical engineering is very well understood and rigorously defined. The only extent to which there is an ‘x’ factor not covered by the legit engineering is that human perception does play a role in, well, human perception. So if you tell someone that this piece of equipment cost more than a car, people will genuinely believe it does sound better. Not too dissimilar from the very real placebo effect in medicine.

Anecdotally, I was recently listening to music in some audiophile house, streaming from ~15K$ eq. It wasn't even close to anything I've heard before, other than live music (in something like a jazz club, not a rock concert). Profound quality. I'm sure it's overpriced but man - if I had the money I would replicate his setup without a thought.

> Profound quality.

I've listened to a friend's $10k setup. It was more like a live concert than listening to a regular stereo. Yes, it had a tube amplifier. If you remember anything about the audiophile's setup, what was the setup?

This measurability nonsense is a red herring.

Think about it, the question can easily be answered by blind A/B testing using real listeners with varying aptitude for discerning any differences and the qualities.

Over the years people have tried and failed often to show much difference, when tests were done with proper controls and with scientific rigor.

To be fair I haven't read up on the subject recently so if there's anything current that shows otherwise I'd be glad to read a link to a paper from a credible source.

Yes. Approximately no-one believes that we have definitive measurements of subjective sound, any more than we have definitive measurements of how good a painting is.

The key thing is the equipment owner's self-suggestion, vs. whatever adequately controlled blind testing would reveal.

I've done a small amount of audio restoration work, in which you're often trying to judge (e.g.) how much of a noise-removal process to apply. How much of that high-frequency content is just noise, or is there actually some of the original signal still lurking in there? Apply too much noise-removal and you'll take away some of that signal.

In doing this you frequently bump up against the question, "Can I actually hear a difference?" My rather sobering experience when I actually put it to the test was that I was way beyond the point of being unable to identify different treatments of the same audio, any more than could be explained by random chance, and yet I was still telling myself I could hear a difference.

Which is why some professional audio plugins offer a “blind”-mode where all interesting but visually distracting meters, spectra etc are switched off. This is also the appeal of analog gear, where suggestive metering is uncommon.

False.

Rigorous double blind a/b/c/d tests using real listeners has been performed, and the research has been published in peer reviewed journals since the 80's.

What has happened is that most of the audiophile community saw research as a threat to their revenue streams and actively FUD around it. And the audio engineering community has simply moved on.

Here is a talk that could be very informative. https://www.youtube.com/watch?v=zrpUDuUtxPM

His book is worth a read as well.

What's one thing I said that was false?

I said studies have been done that failed to show any difference, that's absolutely true.

I said I was open to hearing about recent papers published showing otherwise, yet you've provided zero links to a peer reviewed paper. You offer a Youtube video? If it's a video about research can we just skip to a direct link to the science? Citation please?

I said it's not necessary to know if current testing equipment can detect differences, to know if people can tell the difference. True.

Where are the rigorous double blind studies, with full disclosure, conducted with no conflicts of interest in funding or affiliation?

And of course I assume we want to stay on topic and look at research that isolate amplifiers for some tests, not conflate them with moving parts such as speakers.

For personal listening maybe. PA systems for concert halls are sold in blind demos. Several vendors competing for the contract hang their systems behind an acoustically transparent curtain and the buyer decides which one sounds best.

That About DACs video is fascinating. It gets to the issue of what is it audiophiles might be hearing that we just can't seem to measure. (An explanation of why "the specs are so much better" but "it still sound wrong.")

Seconded. Great video, very informative.

Roughly speaking, tube amplifier goes in distortion adding pair harmonics when the solid state add odds ones. Odds harmonics aren't nice to the ear (see by yourself by listening to a Sawtooth wave) while even ones are much more flattering (see by yourself by listening to a triangle wave). This distortion relates to the entire theory that's behind the creation of chords and also to psychoacoustics.

Both square and triangle waves contain only odd harmonics, with the high harmonics of the triangle wave rolling off faster. The sawtooth wave contains both odd and even harmonics. You can't have a wave with pure even harmonics because the fundamental is counted as an odd harmonic. If you half-wave rectify a sine wave you get only even harmonics above the fundamental, but to my ears this just sounds like two separate tones: the sine fundamental and a separate higher pitched odd-harmonic wave.

Objectively speaking, if the subjective enjoyment of your music is generally increased, then that's a net positive. So it at least depends on the target audience.

If the target audience is over 35 than chances are their wallets are fat but their ears are already shot. That won't stop them from blowing a lot of money on stuff that doesn't make a difference at all (either measurably or subjectively) but if it makes someone feel better they should be free to waste their money as they see fit.

Ha, yeah there's a lot of snake oil in the business.

However, there is definitely a difference in sound, however subtle to some, between most tube and solid state amplifiers when it comes to recording music, as musicians tend to try to draw out certain sounds that are a consequence of the design of the amplifier, whether by overdriving the input or exploiting analog flaws, etc.

Even when comparing just tube or solid state amps, there can be a lot of variance between them.

This doesn't really apply to consumer-grade amplifiers where the goal is perfect sound reproduction. A good tube amp and a good solid-state amp shouldn't sound different to an untrained ear. But amplifiers used in production often try to differentiate themselves in some way sonically.

What if I'm balancing an increase enjoyment from feeling I've got better sound against a decrease in enjoyment from feeling I've being taken for a fool by snake oil salesmen?

Refer to my response to Jacques below.

My rule is: If it's real, then it can be measured, and if it can be measured, then it can be reproduced in solid state circuitry, perhaps using digital techniques if necessary. With that said, whether anybody is actually doing so satisfactorily is an open question. Actual test data for things like musical instrument amplifiers is remarkably sparse. And I won't begrudge anybody the pleasure of owning or building historic circuits.

You will be very interested in the Carver challenge [0]. In 1985, Bob Carver challenged the Stereophile magazine by claiming that he could perfectly reproduce the sound of any amplifier they chose by tweaking one of his (solid-state) amplifiers. They took him up with a high-end tube amplifier, and he won. Note that he did not use measurements but a brute-force approach of feeding the difference in output signal of both amplifiers to a speaker, and listening for (and manually correcting) any residual output.

https://www.stereophile.com/content/carver-challenge

Not really. Tube amps use quite a lot of power and their sound can be captured very easily by some of the more expensive modelling amplifiers.

It's a little bit like the tonewood debate in guitars, no one has been able to objectively prove it in electric guitars but some people swear by it.

I love tube amps. Would like to start building my own. Could you please share your favorite resources if any? Thanks.

Did you want to make hifi amps, or guitar amps? The electronics knowledge and construction techniques are nearly 100% transferrable, but the design lore diverges, and the online DIY communities can be very diverse. I've built a dozen or so guitar amps, preamps, etc.

I don't personally build hifi stuff, but I love John Brodskie's tubecad.com for the analysis of classic circuit topologies, and tons of original ideas. He offers nice looking PCB's for preamps and power supplies.

For me it was useful to determine what I wanted to build first, and find the related online sites and communities.

HiFi. Short anecdote: I once went to a tube amp repair class. After about 5 minutes I realized this was only for guitar amps. I just pretended to be a guitar player and sat through. It was awesome though.

Anyway the TubeCAD site looks very interesting. Thanks.

Pete Millet's site is a pretty good resource, he's posted a lot of different designs ranging from preamps to power amps to headphone amps. His "Starving Student" design was my first foray into building a tube amp about a decade ago, though that one is really a tube/transistor hybrid and it's designed for headphones but I used it as both a speaker and headphone amp.

http://www.pmillett.com/