This voltage multiplier doubles voltage Charging the first capacitor and at another point causing it to discharge, charging the next capacitor. This is added in stages so that the starting voltage is higher, which results in a doubling.
We have a few point of loss in this design, one is the diodes, as you mentioned, but also the rate at which the capacitors discharge. If C1 discharges faster than c2 can charge, and with standard parts this could happen, up to 20% faster, and C2 is always conducting to the next stage, then you could have a situation where C1 is not fully charging C2 in comparison to draw. Since some of this is caused by how long the cycle of discharge, caused by the switching frequency of the IC, there is some fudge range, and 9v does not sound too bad, but not the best and "should" not make a lot of different in volume, but some, as this design is already "quite"
In a design I have been working on, using the same basic supply http://www.diystompboxes.com/smfforum/index.php?topic=94778.0
you can see my voltages starting at 9v, at third stage with load I saw 25v.
Looking up your speaker, it looks to be a very inefficient speaker as it is rated for 116db at 150 watts @ .5 meters. Do you have a speaker you can test with? another guitar amp you can just unhook the speaker for a second and test?
The Murder one is putting out less than .4 watts. (65mw max for that tube and your not going to get half that If I remember correctly)
This is one of those Just Loud enough amps, I love them, great for practice and for recording.
I am on travel till Friday, and I just got my transformers in, just need a voltage drop resistor for the heaters and I will have mine completed. I will measure my voltages and let you know how it comes out. I have an SPL meter and will measure with my efficient Red Coat speaker and my speaker from a Mic stand monitor to see how much difference I get.