For what it's worth.
Build your replacement boards on perf board so you have lots of mounting options.
There are only a few parts. Do a point to point hook up drawing and then draw a schematic.
The Ham way to get HV diodes is to stack in series and has been used since Germanium. Conventional wisdom for many years was (and for some still is) to use voltage equalization across each diode. Something like 470K and .01 uf. Diodes got CHEAP so sacrificing a few is no big deal. Resistors & caps price has increased, A lot of discussions were had on the need for voltage equalization and I quit bothering with equalization probably 30 years ago. I usually add an extra diode or 2 to the string in case of Murphy and have yet to blow a string on high voltage supplies for HF linear amps and what have you by diode junction PRV failure. Done some really dumb things and let the magic smoke out.
1N4007 are 1KV @ 1A and cheap as dirt. If you need a few PM me with an address and a few will show up in a plain brown envelope. I have a few thousand of them. If you are so inclined make a small donation to AK.
The board looks like the Selmax module may be 2 strings or singles pointing at or away from each other. A DMM on ohm or diode is a good way to check if it is not blown. You can check against a known good diode. Think 0.6V drop per silicon junction if you have a diode scale.
A complete photo of the front and list of where the wires go would help. Are black & white that go to what I would guess is the transformer primary 120 VAC? None of the wiring looks like HV insulation so you may well be under 600 VDC
You may well have a voltage doubler supply.
Is there any marking on the round thing that looks like a torrid with orange leads held by screw nut on top. The board seems a bit too old to have a powdered iron or ceramic torrid.
I expect if you hacked into the little Selmax module you would find several diodes in series potted. Theory purports potting helps equalize temperature so fewer diodes are needed, I think just to hide what's in there.
Never had electro static speakers so might miss something.
My take on the caps is: use what's cheapest. Probably a ceramic disc. It's just a almost no draw power supply unless I am missing the boat.
On to voltage testing. The supply will be just as fussy about the input impedance of your test meter as the meter will be if the voltage is over 600 to 1KV = SMOKE!
If you have a DMM, even a modern cheapie, I expect the input impedance is around 10 meg. Make a string of 1 meg 5% or just measure a handful and select 10 that measure the same / closest with your DMM.. Hook them in series and hook your meter across the bottom - ground side one and you have a 1:10 voltage divider to protect your meter. You want the resistor value you hook across about 1/10 the input impedance of your meter. Not perfect but will give you a good indication of what you have. If the voltage is low enough move the higher probe up one or 2 resistors and the calc for the voltage.
Remember, you don't need to look for HV. Make a mistake and it will find you! The old saw of keeping one hand in your pocket when probing is still a good idea The supply appears to be low current and caps almost nothing so danger is not much, but HV always deserves / requires respect.