In other words, I think this circuit is a good starting point and I would like to see where you end up with it. Remember, the perfect is the enemy of the good. Perfectionism kills most Hi-Fi projects. First build a good amplifier, then perfect it.
      Headphones and more headphones. It seems as if about half the e-mail I receive pertains to tube headphone amplifiers and the other half to phono stages. This leads one to the conclusion that what the world really wants is a tube phono stage capable of driving headphones. Well, maybe, the world was a bit too broad, as I, for one, want something else: what was in fact pictured at the Headwise website, a portable tube headphone amplifier. You see, I can only listen to headphones driven by solid-state equipment for about one hour; driven by tube equipment, two to three hours. A portable tube headphone amplifier would allow me to listen while walking about town or even just moving from room to room in my house without having to lug a power amplifier.
     But surely "portable tube headphone amplifier" contradicts itself: portable and tube just do not go together. Large power transformers, heavy chokes, fat high-voltage capacitors, thick high wattage resistors, heavy gauge steel chassis, low gauge filament hookup wires, tall tube sockets, and hot, bulbous vacuum tubes -- all work against compactness, lightness, and ease of movement. True enough. However while it is true that solid-state easily wins the portable battle, portable tube equipment has existed in the form of hearing aid amplifiers, small picnic radios, Geiger counters, and microphone preamplifiers. So ultimately, the issue is one of degree: how portable do you want it? Small enough to fit in your pocket or small enough to fit in a briefcase? Battery or wall powered? The punch line is that battery powered, nuvistor or Subminiature based, pocket headphone amplifier is possible. Possible, but possibly not desirable... A short overview is needed.

    Headphones and tubes seem a natural pairing. First of all, headphones have a much higher impedance than loudspeakers, which moves them closer to the tube's end of the playing field. Additionally, headphones do not require a more than few mW to roar. Yet complications arise.
      While most headphone have an impedance that is higher than a loudspeaker's 4-8 ohms, this impedance wildly varies from one model of headphone to another: from 16 to 2000 ohms. This variability is truly burdensome, as optimizing a power amplifier usually requires optimizing for one load impedance. For example, even when designing a direct-coupled, solid-state amplifier (almost always heavy feedback designs), the load's impedance must still be factored into the mix. For example, solid-state amplifiers for auto use are optimized for working into a 4 ohm load, as most car speakers have 4 ohms impedances. Thus, solid-state amplifier for auto use are designed to deliver a high output current and a low output voltage. If car speakers came with 16 ohms impedances instead of 4 ohms, then the design of car amplifiers would be inversed: high output voltage and a low output current. Now the ratio between 4 and 16 ohms is only 1 to 4, but the ratio between 16 to 2000 is 1 to 125. All of which leads to a large complication: determining how much current and voltage should a good headphone amplifier swing to meet a 10 mW output specification. The math is simple enough:

   Ipeak  = Ö(2Wrms / RLoad)
   Vpeak = Ö(Wrms x 2 x RLoad).

At the 2000 ohm extreme, peaks of 3.16 mA and 6.3 volts are needed to make 10 mW of output power; and at the 16 ohm extreme, peaks of 35.35 mA and 0.566 volts, for the 10 mW of output power. While the 6.3 volts output specification is easy made by any tube, 35 mA is quite a hefty current draw for most miniature 9-pin tubes such as a 6DJ8, 6FQ7, 12AX7, 12AU7, 12BH7, or even a 5687.