Best balanced tube preamp for NHT XD System?

I'm hoping to receive some guidance on the best tube preamp for a NHT XD 2.1 (hopefully soon to be 2.2) setup. I'm presently using a Parasound JC2 and am mostly pleased. The Parasound is fed exclusively by a Benchmark DAC1. The sound is wonderful, but slightly on the "clinical" (as opposed to "warm") end of the sound spectrum. I'm hoping that the tubes can solve the problem. I've read a single (highly favorable) review of someone who matched the Cary SLP 03 with an XD setup, but the reviews for this preamp are scarce and somewhat mixed. I have a less-than-ideal Sunfire Classic Vacuum Tube preamp that needs to be refurbished and would prefer to go another route. Finally, and as a reminder, I would also prefer XLR balanced in and out.
I used a Rogue 99 Magnum and a BAT 3iX with the Xd. I think the BAT worked very well, but ultimately there wasn't enough warmth in the upper midrange (particularly massed violins) for my taste.

I've heard from others who had very good results by replacing the Xd processor with a DEQX unit, but you'll also need two stereo amps for this. However, this would give you a lot of flexibility in "tuning" the sound of the speakers. See the replies to my review at audioasylum. Also, you might want to wait until you have the second Xdw as apparently that can affect the sound a lot.
I believe the XdA has an input impedance of 10K ohms (on the low side), so you might give the preamp's worst case output impedance some consideration.
As Bob says, the input impedance of the XdA is 10k Ohms. The BAT 3iX I mentioned specifies a minimum input impedance spec for a mated amp as 10k Ohms. My Bryston has the same input impedance (10k Ohms/phase for balanced input), and I know it sounds too lean with the BAT. Which makes me think that I was *not* getting optimal results with the BAT+Xd combination, and I have to take back the recommendation. Probably best to stick with an SS pre-amp with low output impedance for the Xd.
Daverz, there are balanced tube preamps out there where 10K ohms in an amplifier is no concern at all, just FWIW.
BAT seems to be saying that it shouldn't be a problem in their FAQ:

"Is the requirement for minimum input resistance dictated by the output resistance of your preamps?

No. However, this is a popular misconception. It is common to apply some old rules-of -thumb to things like preamplifier to power amplifier interfaces and state that there should be some magic ratio between the output resistance of the preamp and the input resistance of its load. People commonly mention numbers in the 10:1 or 20:1 area, some as high as 100:1. Unfortunately, there is no truth to such claims. As many of us know, one can perfectly transmit a signal in a system where the load impedance is equal or even lower than the source impedance (witness any cable TV system). The unusual design of Balanced Audio Technology's preamplifiers allows their gain stages to drive loads with resistance much below what their specified output resistance ratings would imply. For example, the VK-50SE preamplifier can supply a load with 65mA peak current - an amount well beyond that of the great majority of preamplifiers on the market today. However, it is also important to understand the role of the output coupling capacitor, present at the output of BAT preamplifiers. It is the size of this capacitor that will dictate the minimum power amplifier input impedance, not the output resistance of the preamp."

However, VTL gives a 200:1 figure on their website:

"If different makes of pre- and power amplifier are being considered it may become necessary to compare the output impedance of the preamplifier and the input impedance of the power amplifier. Generally for full frequency response it is advisable to match component impedances with a look-up factor of at least 200.

This is not to say that the components would not work together. The audio industry does have open architecture and a common interface between components. However what can happen with disparate impedances is that there could be substantial frequency response roll-off between the two components, particularly in the low frequency bass region. This is only potentially a problem when considering a tube preamplifier to drive a solid state amplifier, as many tube preamplifiers have very high output impedances, (designed to drive high input impedance tube power amplifiers) and most solid state amplifiers have low input impedances. Sometimes a high output impedance preamplifier can be very sensitive to the capacitance of the interconnect cable, which tends to roll off the high frequencies also. "

The BAT/Bryston combo seems much leaner than it should be. I'd like to try some SS amps with higher input impedances.
Daverz, the general rule of thumb is 10:1, but you see these higher numbers due to the fact that a lot of tube preamps have a coupling cap at the output that will limit their low frequency response/impact. Its another way of saying that their output impedance goes up as you go into the bass frequencies.

IOW 10:1 is OK as long as the actual output impedance is the same at all frequencies, which is not the case with a lot of tube preamps. Either the coupling cap has to be made a lot larger (which reduces transparency) to deal with 10K at your amp at 20Hz, or done away with altogether (direct-coupled). With direct-coupling you have no issues with bass impact at all :)
Teaman, what did you end up doing? What was your concern with the Sunfire tube pre-amp. I've read good things about them, and was thinking about trying that.

So far, I've used an Adcom GFP-750, a Primare PRE30, a B&K PT-3II, and briefly a Sonic Frontiers Line 2.

All different, and good, but none really brilliant. the Adcom, Primare and SF all sounded a bit lean and disambiguated, where the B&K sounded more coherent and "rich" but with less detail and separation of instruments - also a bit darker. (SF Line 2 was the brightest, interestingly).

I am still looking for some combination that sounds more coherent and musical, but still has good detail and articulation.