Vacuum tubes have been around for ages, and for better or worse, they have their advocates for use in amplifiers and preamps. However, tubes are simply inconvenient devices. Even a 12AX7 preamp tube is huge relative to a handful of transistors, tubes require weird voltages, and each and every one of them is a through-hole device that doesn’t lend itself to machine assembly.
This changed recently with the introduction a strange new tube from Japan. Noritake and Korg recently introduced a triode that uses the same packaging as VFD displays. The Korg Nutube is a vacuum tube that operates at lower voltages, is smaller than the usual preamp tubes, and still has the vacuum tube sound.
For his Hackaday Prize entry, [Kodera] is building a headphone amp with this new tube. Is a tube-based headphone amp particularly novel? No. But this is the first we’ve seen anyone playing around with this new, interesting piece of technology.
The requirements for this Nutube are simple enough, and the minimum anode voltage of this tube is just 8 V. [Kodera]’s circuit is running the tube at 12 V, and the only other circuitry in this preamp are a few coupling caps and an op-amp just before the power stage.
[Kodera] has crammed this circuit into a proper amplifier using a 2 x 15 W class-D chip from TI. It’s really a phenomenally simple circuit that’s also remarkably tiny. These kits are actually available on Tindie. Time will tell if the Nutube is picked up by some big-time manufacturers, but we’re happy to see someone is playing around with the latest advances in tube amp technology.
You can get a VST plugin for ‘the tube sound’.
no digital device or software can ever (will ever!?) recreate the genuine harmonics of a good ‘tube’ amp, and that includes the combined might of google’s AI and all of the other silicon valley outfits. Even after the singularity, the computer overlord will probably have to go out and buy some real hardware to recreate the sound. Get yourself a nice proper tube amp at 100W or so (1959SLP or JTM45 etc., none of that mosfet rubbish), play a decent guitar on it (maybe a nice US tele or something like that) through a nice big 4×4 cabinet, turn up the main volume till it starts to distort. All kinds of harmonics come through, I’ve never ever ever heard any kind of digital device (or even a distortion pedal) come even close to the sound. By the way, ‘tubes’ are also called ‘valves’ in some parts.
http://www.ecircuitcenter.com/Circuits/pushpull/image006.gif
This circuit compensates completely for all distortion in the push-pull stage, while remaining efficient. You can low-pass-filter a DAC to pulse-density-modulate perceptually continuous decibel levels which would make you instantly deaf. A CPU can synthesize a signal like that of several megahertz without even breaking a sweat, so simulating a tube amp circuit in realtime is easy and commonly done. I could not find the VST I originally used because there are *so many of them* now.
You are wrong and you are wasting your money. Spend it on concert tickets if you care about audio.
look, the only way to get the proper sound of a good valve amp (and I mean a proper dirty sound, but not overdriven beyond all recognition) is to buy a good valve amp (e.g. Marshall etc). Thats just the way it is. Its nothing to do with how clean the ‘clean’ sound is, or with simply getting odd harmonics as the waveform distorts due to signal saturation (anybody can make an overdriven circuit with a couple of cheap transistors). It is down to the non-linear harmonic behaviour you get when the valves are operating at high power. I’ve tried all the various digital effects systems, nothing gets anywhere near the real thing. For people who just want to play ‘flat’ chords, maybe some of the blip-blop rubbish that comes out of a CPU might be okay, but if you want to play proper riffs and licks that have some actual feeling and variety to them, you need a real amp. You also need a real guitar, made of real wood by a proper craftsman, with good pickups. In fact, the amp and the guitar are two essential parts of the overall instrument. You also need to stand near to the speakers to get a bit of feedback. One thing you don’t need is any of that ‘oxygen free copper’ nonsense, thats just a gimmick.
you severely underestimate computers and massively overestimate human senses, we are so far beyond what humans can sense or perceive in precision or reproduction today.
as a thought experiment;
if you had a piece of source audio and piped that through the tube amplifier you want and record the resulting waveform with enough precision then one could reproduce the sound so exactly that there would probably be more variance from hour to hour on the actual physical tube amp itself, a variance that one could characterize and emulate if you really wanted.
You are so wrong. First of all, I used a VST plugin that emulates tube amps in 2001 on Celeron 500MHz with 512MB of RAM. There was also iZotope Ozone plugin for Winamp. Presets were rubbish, but with some tweaking it achieved great results. Emulating tubes is actually not too hard – you need to add some harmonics and then a bit of soft overdrive that doesn’t cutoff the signal amplitude the way transistors and diodes do. I could probably do it on some dsPIC, and you wouldn’t know the difference. Just look around and do some double blind tests. Your ears are not golden and your brain is lying to you based on your expectations…
Hmm… Why do I get the impression that himmy is the one with the ‘oxygen free copper’?
Probably something to do with the obsessive-neurotic ‘nonsense’ he’s talking.
Valve amps are not ‘better’ than ‘solid-state’ amps. Fenders are not ‘better’ than Gibsons. They are simply different.
Which you choose depends mostly on what you are trying to achieve, and for most musicians, what you can afford.
As an aside, originals are not necessarily ‘better’ than copies either. I’ve seen both bad originals and brilliant copies.
Build that circuit, try it out and come back. Tell us how good it sounds. If you actually build that you’ll see it sounds like trash
“no digital device or software can ever (will ever!?) recreate the genuine harmonics of a good ‘tube’ amp,”
Bullshit. The tubes are implemented by way of application notes and documentation from tube companies, and those charts and instructors are easily interpreted into software. We have software that can model all manner of extremely complex systems, including analog circuits (which is all a tube is…) – for fucks sakes we can model atomic bombs. The behavior of tubes is easily within the capabilities of software to model. I bet anyone with knowledge of the relevant audio APIs in a programming language, and rudimentary knowledge of how to handle audio sample processing, could design software to model a tube given a book or two on their characteristics.
You’re probably one of those idiots who thinks that “clock drift” in “consumer” CD players “causes distortion.”
it needs to be rubbed wit unicorn tears and build with wood polished by virgins deep in the Amazonas to get the full experience, anything else is just a poor cheap imitation
@himmy-jendrix says: “no digital device or software can ever (will ever!?) recreate the genuine harmonics of a good ‘tube’ amp”
Nonsense. The Laws of Thermodynamics theoretically disproves your claim himmy-jendrix.
But never-mind that. Let’s think about how you would almost perfectly (any difference would be imperceptible to a Human) “recreate” the genuine harmonics of a good “tube amp”, as you say…
1. Get a “real” degree in Electrical Engineering with an emphasis in Systems, Coding & Information Theory, and Digital Signal Processing (DSP).
2. Study the subject of Human “Psychoacoustics” in-depth (it is a deep subject). This is critical to put bounds on the efforts described below.
3. Beat your head against the wall for a while learning how to code any sampled physical Tube-Amp so it can be “cloned”. Typically you would use a decent PC sound card (they’re quite good these days, tame your ground-loops please!). This will require “capturing” real Tube-Amplifiers properly. (Your study of Human Psychoacoustics above will likely give you knowledge about how to properly “Mic” a capture session).
I think some work has been done here with Guitar Amplifiers:
http://guitarix.org/
4. Build a DSP based physical device that can mimic ANY type of “Tube-Amp” sound captured in Item-3 above.
5. Charge Zillions of Dollars for what you built! If you don’t charge a LOT of money for it, even though what you made is a functionally Perfect clone of the distortion in a Tube Amp, No “Audiophiles” Will Buy It!
Whilst I disagree with the idea that this is somehow fundamentally achievable because of the laws of thermodynamics, I think you at least recognise that it is not an easy problem to solve and outline some of the key challenges. The whole question of why certain things sound ‘good’ to human ears/brains is interesting. Ultimately its a matter of taste and everybody will have their own likes and dislikes. It is similar in some ways to the question of why certain wines taste good.
@zank-frappa,
You said, “Whilst I disagree with the idea that this is somehow fundamentally achievable because of the laws of thermodynamics…”
The pivotal contributions of Shannon and Nyquist in the area of Coding and Information theory are tied directly to the basic laws of Classical Thermodynamics. Almost everything you use today in terms of electronic devices depends on this, and how it is applied in-practice (the realm of Engineering).
https://en.wikipedia.org/wiki/Entropy_in_thermodynamics_and_information_theory
You said, “Ultimately its a matter of taste and everybody will have their own likes and dislikes.”
That’s the problem – “Taste”. Human “Taste” has no place in-fact. Objectively, the fact is: any analog system can be duplicated as a sampled digital system so close to exact given the bounds of the very well established understanding of Human Psychoacoustics, that an “un-biased” Human will NOT be able to tell the difference provided the sampled system is a functional duplicate of the purely analog system in the time & frequency domain (as a result of Digital Signal Processing).
For example: You have a “Taste” for tube amplifiers. I can take a vinyl record and play it through a tube amplifier so you listen to it. Next I sample the same vinyl record’s audio signal digitally and reproduce it through a properly (emphasis Properly) designed Class-D digital to analog amplifer using the same speakers, and you SHOULD mathematically NOT be able to perceive ANY difference! If you do perceive a difference, then it is quite likely due to your psychological bias toward Tube amplifiers. Something that is NOT based in-fact. The two results are the same, but you prefer the Tube result, simply because it has the word “Tube” involved – Nothing more!
You can use few op-amps to achieve “tube” sound…
you can get a decent sound using op amps (i’ve even got some interesting sounds using BC109s), but it isn’t the same as a genuine tube sound. ECC83 every time.
No-one has mentioned the output transformer, and maybe driver transformer, in valve guitar amplifiers, a transformer is a non-linear device also. It may be possible to play a very large collection of guitar riffs etc and test signals through a ‘classic’ amplifier in order to work out the overall transfer characteristic at all amplitudes, at all frequencies, at all combinations of amplitudes and frequencies, at all sequences of amplitude and frequency combinations etc etc etc, and build a mathematical model in order to simulate the “classic valve amplifier” effect but no-one has done so yet – all ‘plug-ins’ are just nice, or not-nice, sounding approximations.
I have no problem with nice or not-nice sounding approximations – some of them may be ideal for some music after all. But if you want an authentic “classic valve amplifier” sound – get a classic valve amplifier or a modern reproduction. By the way, carefully designed valve and transformer power amplifiers can be almost perfectly non-distorting. It was called HiFi and no, there is no audible difference between a very good valve HiFi and very good solid state HiFi.
I have Bias FX and an Engl Savage Mk2. There’s something to be said for being “in the room” with a real tube amp, but if I recorded both and sent it to a friend, he’d be hard pressed to tell the difference.
A sound of guitar cabinet is very much affected by listening position. Impulse responses are recorded from a single position, and monitors don’t have the same, let’s say “diversity of tone” as a result.
That said, if you record a tube p and cab, and apply an IR from the same cab and mic position to a digital amp recording, you’d have a very hard time telling a difference.
Tubes aren’t magic.
You can use a few good transistors to achieve “tube” sound…
You can use a few wires suspended in a vacuum to achieve… Oh
uh, I thought this sounded familiar… http://hackaday.com/2013/11/09/vfd-display-becomes-an-amplifier/
Tubes amplify differently because they distort the waveform. A high quality solid state amp that isn’t overdriven will give little or no distortion and sound different. If people really want the sound of a tube distortion, just bake it in in the digital domain before passing it off to the D2A.
As far as I can tell part of the problem is that class A amps have low efficiency and so are heat limited in power, Class B amps distort the waveform around zero amplitude, and class D amps are a weird high-frequency PWM that makes some people nervous (and requires a band-cut filter if not going to a magnetic speaker immediately). Also someone said there’s issues with transistor amps requiring DC offset filter caps that are electrolytic because they need to be big but used in a non-polarized application. I don’t buy that last bit as I’m pretty sure pairing caps to depolarize does really work and well enough to not distort anything.
Its just like records. Most of the record fanatics refuse to believe that sound engineers have to squash the living hell out of the audio master to make it sound good on a record and it is nowhere near what it sound like in real life. There was an interesting articles on what they have to do, Ill have to find it again.
This may be the one:
https://hackaday.com/2013/12/31/retrotechtacular-wax-on-wax-off-how-records-are-made/
If real life is second rate classical musicians and a lazy conductor playing in an auditorium with poor acoustics, or a singer in a bar with an accompaniment of shouting and police sirens, or a lacklustre local group playing covers of widely differing artists – then give me record any day.
Incidentally I only buy second hand records (flea markets and charity shops) not cds but for a new purchase it would be a cd. I don’t think my buying choices are science driven.
At the tail end of the tube era there were Nuvistors, which were metal-cased and about the size of a to-5 transistor. I’m not sure size was the selling point, they were low noise devices especially good for receiving, and in the VHF and UHF range. Not really power devices, RCA did circulate a schematic for a low power transmitter. They even had Nuvistors that worked fine with 12volts on the plate, it was the time of hybrid radios with tubes still handling the radio end and transistors for the audio stages. Thy saw a lot of use in tv tuners, and definitely in at least one Tektronics oscilloscope.
So they could be used to make small audio preamps.
There were also subminiature tubes, they were bigger, seen in hearing aids and portable radios.
Michael
“Time will tell if the Nutube is picked up by some big-time manufacturers”
Korg isn’t a big time manufacturer? Give me a break.
My thoughts exactly… also one of the few of the big players who are still trying to innovate instead of simply repackaging the same old sounds and samples.
Noice!
I’ve been scratching my small tube itch by playing with a few of the Soviet 1970s era miniature tubesm which are the size of an AAA cell, and still available in lots of 10 for like $1 each on the ‘bay. I will have to try one of those tubes in this circuit.
Also, many common or garden-variety tubes will also run at reduced voltage (eg 12 to 24 v), which can add more interesting nonlinearities.
For a different take, Pete Millet has also designed a little headphone amp using the nutube.
Why make every person interested in your comment go search for what you’re talking about? Just add the link.
http://www.pmillett.com/nuhybrid.html
He’s also a US distributor for the nutubes:
http://www.nutube.us/
PAIA for a while had a tube preamp, it was in Radio Electronics (or maybe it was Electronics Now by then). I think it ran on low voltage, the dual triode was there for the “sound”, but there were solid state parts in the signal chain.
Michael
http://www.paia.com/manuals/docs/9305-tubehead-schematic.pdf
Cool to learn about this “tube”, but the guy designing the headphone amp seems clueless. E.g., I immediately noticed C9 & C10 connected to the opamps without any resistors to ground for input bias. If the circuit works like this, it’s only occasionally, by dumb luck. And those opamps (not the TL072 of the schematic) – U$10! Get real.
Huh, someone’s actually doing something with these (and looks to have sold several on tindie even!) I remember this being released maybe 2-3 years ago, wondered if it was just a proof of concept or if there would ever be products based on it.
I have purchased one of his kits, the first one which uses the korg nutube in a headphone amp configuration. Its a very nice kit with a simple hybrid tube + opamp stage. Worth the money considering the nutube alone is $50 and hard to source.
Yeah…kind of a novelty at that price. As I mentioned above, you can still get Soviet-era mil-spec tubes including very small pencil tubes for a buck or two each.
All things considered any use of a vacuum tube in this age is a novelty. I have a box of Raytheon 6418 subminiature vacuum tubes to keep my Oatley headamp kit running. They are fun to play with but suffer from a laud ringing caused by any vibrations which makes them useless in a pocketable design.
Vox (owned by Korg) recently released their MV50 series of amplifiers, which utilize the Nutube and a class-D power amp. They sound pretty decent, too.
http://www.voxamps.com/MV50
All you people salivating over “The Supremacy of Vacuum Tubes Over Transistors”, here’s an opportunity to put your money where your brains aren’t.
I have an original RCA Vacuum Tube Manual, in very good condition. Used only in a college course on vacuum-tube electronics design.
It can be yours for $495.00.
What year? Mine are copyright 1959 (taped together) and 1964 (in fair shape, quite usable).
does it have the 12AX7/ECC83?
Everything has the 12AX7, it’s one of the most common tubes ever produced. It was one of the very first “miniature” tubes smaller than the octal socket. I have a drawer full of them pulled from various surplus disassemblies.
Most of the good tube manuals and specs are available online for free, but thanks anyway.
The tube hate is almost as irrational as the extreme audiophile wankery.
Whether you choose to acknowledge it or not, recording music is still a balancing-act of compromises and artifice. Almost nobody is listening to “pure sound” eg two very good mics going through the minimum of electronics to be recorded to either a highly optimized analog system, or the most accurate digital system. AND the very best speaker systems you can buy are still running at greater than 1% distortion, not to mention phase discontinuities around the crossover points path length differences, edge diffraction, etc. And then there’s the room itself.
So… many compromises and choices, There is no ONE right, ideal sound system.
There are subtle differences to be heard between tube and semiconductor circuitry, and it is indisputable that some of the “bad” things (tube and transformer nonlinearities, analog tale saturation) actually sound “good” in that the characteristics are pleasing, and they are still sought-after sounds in the recording studio.
Tube circuits often contain fewer active devices than solid-state circuits, since tubes usually provide higher gain, and there are still beneficial interactions between tube output stages and speaker impedance that you don’t get with the best ultra-damped SS amps.
So the tube afficionados aren’t necessarily wrong… but the differences are usually so subtle (and not necessarily more accurate) that the average listener doesn’t care, especially if they’re just streaming music, and so the huuuge prices for esoteric tube gear seem (and are) outrageous.
(To me, $50 for this Korg/Noritake device is outrageous, but whatever.)
all of your examples here are fairly well characterised and thus easy to emulate, plenty of modern music uses digital emulation of old sounds and errors.
it might be practically easier to get your results physically but it isnt the only way, that is my primary beef with tube worshippers, they claim that nothing could ever reproduce the results of their beloved tubes, in the modern age athat is clearly ridiculous.
with good enough data even a cheap modest cpu could reproduce everything, all the harmonics, all the distortion, everything, you might need a special sound board to do it but at this point i actually doubt it, many modern sound cards go well beyond human hearing.
Karplus-Strong string synthesis plus some carefully selected nonlinearities and even-order harmonics can get very close to the sound of a gold-top Les Paul with DiMarzio pickups running through a Mesa boogie amp. But of course, it’s much easier, faster, and more “musical’ to simply grab and play a gold-top Les Paul with DiMarzio pickups through a Mesa boogie amp. You’ll dial in a pleasing sound a lot faster than tweaking a DSP.
Similarly, you can add a DSP front-end to a good quality class D amp, and screw with the damping and it could sound like an old McIntosh tube amp… or you could simply get your hands on an old McIntosh amp.
If you want the sound of a tube circuit, … a tube circuit is often a good choice.
Many people enjoy the sound of tube electronics – as processors in the studio or at home. Does this justify the hype and the sometimes stratospheric prices? Not to many of us. But to some, yes.
And it’s fun to play with tubes, too.
i never said tube amps shouldn’t be used, all i have been arguing is that it is clearly wrong to heighten them to some mythical status where nothing else could ever reproduce the sound or effects.
as long as that is understood it is immediately easier to actually discuss tube amps for what they are.
I don’t think anyone is disputing that some of the ‘bad’ things can sound ‘good’. What is in dispute is that such sounds can *only* be achieved by using tubes. I assert that *no* tube amp can be built that could not be characterized and digitally simulated with results indistinguishable to the human ear (double-blind ABX test). There is no magic in tubes, only physics.
“a triode that uses the same packaging as VFD displays”
If you want a triode in a VFD form-factor, just use a VFD – at least that’s a hack.
The VFD is a single device while the NuTube is several in one package, and the VFD is sensitive to vibration because of the long unsupported filaments which are necessary to not obstruct the displays.
I wonder if a VFD could be used as a crude microphone.
It probably could, and it would be a pretty cool hack.
Am going to try it. Probably going to burn out old tape player haha.
Mullard (or whoever owns the name now) tried this some years ago. They were advertising a 12AX7 based on some CRT or planar layout. I think some engineering samples went out and it ended quickly.
As for the tube haters, who cares? I enjoy designing with them and building amplifiers. Using your logic no one should bother driving vintage automobiles or living in old homes.
Isn’t the output transformer a large part of what creates the “tube sound”?
Not really. The transformer just matches the high-impedance high voltage/low current tube signal to the low impedance low voltage/high current needs of the speaker. Those audio transformers are designed to have much higher frequency response than the amplifiers they’re made for, so they aren’t really modifying the signal in a way anyone could ever hear.
yes, massively.
some more info:
A transformer is a non-linear device, with lots of factors influencing the sound. Non-linearity of the magnetic flux varying under load (saturation), bandpass filter characteristics, non-linear distortion leading to generating harmonics, etc etc.
Gonna try to split the difference here. Yes a transformer (and its interactions with the speaker) is part of the “tube sound” particularly when it’s in the output stage of an all-tube amp and is being driven into the non-linear part of its transfer curve. (eg guitar amps).
A well-designed transformer, especially an input or interstage transformer, is just about transparent when used as designed.
I could never be an audiophile.
Don’t get me wrong, I appreciate music and I appreciate good sound. But to appreciate it down to the minutia of detail where these kinds of specialized electronics start to actually matter… I would need time to just sit and listen. Not listen and drive, not listen while tinkering on some project, not listen while working… just listen. I would need an uncluttered room to eliminate all of the spurious reflections. I would need my daughter and all her noise to be elsewhere. Finally, if I had all that I could be an audiophile for about 1.5 minutes at a time, the longest my wife ever goes without talking!
That’s ok because after about 3 minutes I would go crazy from not spending my time actually doing something.
I think this is actually a religious argument, no, not even joking. I think it really boils down to worldview.
I generally have no patience for audiofoolery, but when I hear statements like “you severely underestimate computers and massively overestimate human senses” I detect a person who has rejected the possibility of any enchantment in the world. I’m going to remain agnostic in that, I simply don’t know, but I do recognize that while some are ready for the singularity others prefer cedar groves and believe they can feel inexplicable things deep in their spine.
To each their own and keep an open mind.
what do you mean by enchantment?
mystery in the world? i have plenty of that, at the moment gravitational astronomy is a bit more exciting than basic electronics though.
there are plenty of actual mysteries without having to invent my own.
totally agree with you about keeping an open mind. It is getting kind of religious.
I think there might actually be two different arguments/concepts that are getting mixed up in this thread: reproduction of prerecorded sound, and actual creation of the sound in the first place. When it comes to reproducing a guitar sound that has already been recorded, I wouldn’t particularly care if it was tubes or mosfets or whatever, so long as there was NO distortion (ok, maybe some people want distortion during playback, thats up to them). In this case, digital to analogue is fine, I don’t believe humans can’t hear the difference if the amps are high fidelity enough and the sample rate is high enough.
When it comes to actually creating the sound in the first place, that is a completely different matter. Now it may be possible to get some really really nice dirty guitar sounds using a CPU and some signal synthesis, but it probably won’t sound the same as a decent valve amp. It may even be preferable to some people, but it won’t be the same. The same goes for the guitar itself, including the materials, craftsmanship and pickups etc. If you play a guitar made of plastic it will sound different to one made of hardwood. I haven’t heard of any kind of digital system that can make a plastic guitar sound like a wooden one, certainly not in real time. A non-linear electical circuit operating at high power is hard to model exactly, and the little details can make all the difference. Add in feedback and things get mathematically very tricky very quickly, its not even linear-time-invariant. Modelling these things is harder than you might think, harder than just reading a couple of books on Fourier transforms and Butterworth filters. This is why autotune software, despite being very expensive, still sounds utterly crap (to me anyway, some people might like it and each to their own).
I can pick up any number of conventional triodes for a couple of quid each on ebay.
These unobtanium flat triodes are north of forty quid.
Now I know audiophiles love to spend money, and will go for the novelty factor, but I can burn out ten Mullards or Mazdas for every one of these things.. I don’t see any real advantage.
Its not as if the new kid on the block is even all that much smaller, in fact as far as I can tell, they occupy more board space, not less, being short and flat, rather than tall and round.
They do look the part though.
If they had been produced in 1959, then they would have made a mint.. in 2017, I’m not convinced.
They remind me of the Sinclair flat screen TV. https://en.wikipedia.org/wiki/TV80 a novel idea, but not a huge success.
WARNING: This is not a troll, but just a thought…. When it comes to the Stradivarius violin, there is (to my knowledge) no duplicate, and scientists have been trying for a long time to analyze and reproduce the sounds — To me, this lends some credence to the Tube Amp vs Digital argument. For me, my ear is not that good, however, when I was in the audio engineering part of my life, I would design enclosures and I would coat all the inside surfaces with contact cement, it was like a little rubber mat, this would dampen noise, and people would claim a cleaner warmer sound. I so believe there is some truth to the analog side of things.
you are misunderstanding the issue here, i will guarantee with 100% certainty that you can reproduce the sound of a Stradivarius, you might not be able to build a physical violin that does but that is another matter.
sample it with a couple of billion samples a second and a fidelity far beyond what any human ear will ever be able to hear and play it over non colouring monitors and there you are, the sound will be closer to the original than the original is to itself a couple of degrees apart.
you might need ultra high quality equipment to do it but it certainly isnt impossible.
“vacuum tube sound” = harmonic distortion.
RCA Nuvistor triodes:
http://www.ebay.com/itm/5-Pcs-NOS-Philips-JAN-8056-Nuvistor-Audio-Radio-Vacuum-Tubes-NIB-/251324032755
So you have a “valve amp”, known for its nice harmonics, and you strap a D-class output stage to it, known for their horrible distortion. Way to piss off both audiophiles and scientists.
A product like this makes sense for instrument amplifiers, where a certain sound is desired, rather than a good reproduction.
I would consider using this for a guitar amp, but last time I checked, these modules seemed rather expensive and hard to get by.