Reader Juan-José from Mexico. "Reading your review of the S.A.Lab Blackbird reminded me of what I heard at my home some years ago. I owned a Pass Labs XA30.5 and an F5 with a pair of Vapor Cirrus bookshelf speakers. I preferred the more lit-from-within quality of the F5 to the XA30.5 but there was no denying that the bigger Pass Labs amp moved and controlled the speakers better. However, one day a friend of mine brought his Art Audio Carissa 845 to my home, which I think you reviewed a long time ago and then found too slow with your Zu Audio speakers. I wasn't expecting much but imagine my surprise when the 16-watt tube amp really knocked out both of my Pass transistor amplifiers with classical music (I'm mainly a classical music lover and go to live concerts every weekend). Yes, the thing couldn't do heavy rock or percussion stuff as well as the transistor amps but the difference with classical music was staggering. Even with macro dynamics being inferior, the tube amp possessed a liquidity to the sound, a "legato" expression which delineated melodies in a way you could feel goose bumps with, just like in a concert hall. By contrast, the solid-state amps seemed totally mechanical and you could even say broken. But another interesting thing happened too. With the tube amp, when hearing a quartet for example, you could realize much more clearly the interplay between instruments; that the instruments were responding to one another. The coherence was much much greater with the tube amp. I don't know how to describe this or what the cause is but when going back to the Pass amps, you had the sensation that you stopped hearing music and began hearing sounds. And that had nothing to do with detail retrieval which was slightly higher with the transistor amps. I wonder what the magic of tubes is and why they manage to play music so well after all these years (it's ancient technology, as some would say)."

I thought Juan spot on. He even described the legato effect like I had. It's easy to hear, hard to pin to words. To be sure, I mostly hear that with direct-heated triodes. Once one transcends the unfairly popular 300B—unless it's a very different circuit like a Berning—and enters the more exotic 45, 10Y & Co. realm, this peculiar quality surfaces without unduly affecting tonal balance or excessive THD. Obviously Vincent didn't fit a pair of DHT to his standard d1-six. Hence one shouldn't expect that particular result. For contrast, my friend Dan briefly loaned me his Tobian Sound Systems T9 DAC. That black Swiss beast mates 16-bit NOS Philips 1541 silicon—R2R on a chip— to two 31 triodes. Voilà, Chopin in action. Think temporal ebb and flow with extended decays and softer water-colour transitions. Think the moon's effect on water. That's the effect of such esoteric triodes on music. If that be your desire, the d1-six-tube won't do it. Neither will my own valved converters. 6922, 6201, 12AU7 & Bros. don't bottle that flavour. Now remind yourself. Most classical music does without a regular percussion section. Modern music is nearly unthinkable without it. Doesn't it make perfect sense that the fluidic melody-over-rhythm emphasis which Juan talked of will charm string quartets far more than an E.S.T. groove based on very taut time keeping?

But Juan had more. "I’m glad you’re reviewing the new TotalDac d1-six since it’s a DAC I’ve been really interested in (not necessarily the six but any of their one-box DACs). I’d really like to know how it compares to your DAC arsenal. It’s especially interesting to me whether a well implemented Sabre chip can hold its own or even surpass a well-implemented discrete R2R DAC. I mean, all those naked Vishay resistors cost a great deal and technically—by audiophile wisdom at least—they should surpass a microchip DAC. True, implementation and the analog section account for more than the DAC type but even so, when both are well implemented, it’ll be interesting to see the differences. But I think Vincent should have sent you the normal six, not the six tube, because:
1. I’ve seen you tend to prefer transistors for the source.
2. From what I’ve read, the 12AX7 is not the best of the 12A_7 family tubes, having the worst distortion graphs. The consensus seems to be that the 12AY7 is a better sounding tube, with the 12AU7 in the middle.
3. And maybe most important to me, it lacks a headphone output, which, from what I’ve read on various forums, is really good. It’d have been nice to read a comparison with the Bakoon."

With my Fore Audio DAISy1 and Aqua Hifi LaScala MkII converters both fitted with tubes, I rather thought that Juan had misread my allegiances. Whether I prefer some colour(iz)ation at the source or not depends on what follows, primarily the amp/speaker interaction. As to headfi, the d1-six has the 6.3mm rear-panel jack option. Vincent simply didn't fit it to my loaner. As to the 12AX7, that'd been an initial mistake based on Vincent making a typo. He actually uses 12AU7. I didn't know whether that caused the low 1.4V output voltage (the 12AU7 is lower gain than a 12AX7); or whether Vincent ran his bottles as pure buffer without voltage gain. Whilst I had word in, I also asked what, exactly, the advantages were of paralleling multiple resistor matrixes over the d1-dual; how his digital volume control operates; and down to what attenuation value he considers it lossless.

"The 12AU7 are pure current buffers (no voltage gain) which lower the output impedance. Paralleling my resistor ladders lowers the noise and divides the output impedance by a further factor of six. As to the volume control, I've experimented with high-end pots, relays, Shallco + stepper motor, LDR and so forth. The best sound came from my digital volume embedded inside the FPGA and running at 69-bit resolution. It was the only way not to hear standard volume control parts. This doesn't mean that an active preamp couldn't improve the sound. For example, some customers use my volume with their Shindo preamp set to fixed output. It is not easy to say up to what level of attenuation mine operates without loss. This depends on the type of recording format and whether you consider noise, distortion, something else and to what other type of volume control you compare. Analog volume controls have their own limits, even relay-switched resistors. Putting it roughly, my digital volume control is the best I tried for limited attenuation. For very high attenuation, some analog variants can be better. Chances are that when you apply very high attenuation, the listening level becomes very low and now our hearing is far less sensitive." Here is an advantage to the d1-six-tube's low output. In direct-drive mode—DAC-->amp without preamp—you'll invoke less digital-domain attenuation than if Vincent's converter put out from 2-6V as many modern machines do. To complete that picture, we of course still must consider an amp's gain, speaker sensitivity and how loud/quietly you like it (which also is a function of room size and listening distance). Regardless, Vincent considers his digital volume control superior to all analog alternatives; if you use his "within reason". Where reason ends and wishful thinking begins we don't exactly know.