choosing a treble-boost cap value

[stormbringer]

Hi-pass shelving filter theory is covered in Merlin Blencowe's 2nd-edition 'Designing Tube Preamps' book - Ch 4 P 95-96.

The filter covered in that section is made up of a resistive attenuator (R1 series resistor and R2 shunt resistor), with a bypass cap across R1. So in my situation (which is like the Marshall 470k R || 470p C configuration right before the gain pot), R2 would be replaced by the gain pot, and the output of the attenuator would be from the wiper of the pot. With the pot set to 50%, there are two series resistances, only one of which is bypassed.

[tubeswell]

The filter covered in that section is made up of a resistive attenuator (R1 series resistor and R2 shunt resistor), with a bypass cap across R1. So in my situation (which is like the Marshall 470k R || 470p C configuration right before the gain pot), R2 would be replaced by the gain pot, and the output of the attenuator would be from the wiper of the pot. With the pot set to 50%, there are two series resistances, only one of which is bypassed.

That's nearly the same type of circuit, except the pot is a variable R2, which has the effect of setting up a HF rolloff in conjunction with the input capacitance of the following stage (see Fig 4.22). So you still basically have a hi-pass shelving filter, and the pot as R2 provides variable amounts of HF roll-off (as well as attenuation) depending on the pot rotation.

[pdf64]

I don't think it's worth bypassing the 100k feeding the pot, can't envisage that the difference would be perceptible, a 1.5dB shelf at best.

[Chuck H]

+1
I simulated the circuit on Duncan TSC and it's definitely a no gratification circuit. The OP could try switching the cap in and out and see.