Tubes for a pre stage... matched NEW or unmatched identical NOS?

@tomcy6 - Bob does some great work. He used to do Tube Only repair and customization and has his own patented circuit designs. He lives a mile away and has done some repairs for me even after he was building the new Pres. It's a nice sight seeing 6-8 of them on the bench in progress.

On my Audio Research SP6B, the individual tubes can be biased, and it definitively improves sound quality. 
 

+1 re Brent Jessee, excellent guy, my matched Raytheon VT231’s and CBS/Hytron JAN 6S7N’s are wonderful, and he was a knowledgeable pleasure to work with in tube selection. 

Most preamp circuits are run class A due to low signal levels. You want them matched for gain and screened for low noise, again because the signals are low at this stage.

It's always important to match triodes 1 and 2 in a tube that is run in stereo. While triodes that are off by 10% may not hurt anything you will hear the imbalance. If it were just a matter of one triode having higher gain than the other you could adjust the balance control to fix it. But there's more to it. Each triode has it's own frequency response curve and selecting tubes with curves that are as closely matched as possible ensures that the left and right channels reproduce frequencies at nearly the same amplitude. Without this you will hear smearing of the image from side to side and a wandering in the image when a vocalist or instrument moves up and down the scale. So in order to keep a vocalist centered as she sings and a trumpeter locked in one place, the tubes need to be matched. I use an old TV-7 tube tester to see if my triodes are matched but it really only tells me that they are matched at 60Hz because that's the only frequency that it tests at. If you use a curve tracer or an Amplitrex tester you can see how they are matched from 20Hz - 20kHz. Getting as good a match as is practical is never a bad idea.

The Sphinx likely uses tubes for the phono section, so low noise is paramount.

For maximum RIAA fidelity and channel balance, the g1/g1 and g2/g2 sections need to be well matched, assuming the same topology for each channel.

The big fly in the matching ointment is that tubes seldom match across the entire operating range and matched tubes are matched in a matcher, NOT your amplifier and thus operate under totally different conditions. Tubes matched on one matcher may have very different numbers on another matcher.

Tubes matched today aren't matched tomorrow as they age at different rates.

See ieLogical Rolling