Let’s begin with the fact that capacitors are among the most non-linear parts of any amplifier -- so any real improvement is worthwhile. I don’t know what the existing caps are so none of us have any basis to comment on the upgrade value specifically.
Let’s assume they are not the best. Would be better if you said. Do you even know this to be true?
You discuss two types of capacitors: 1) coupling caps which block DC and allow two stages to to be linked, pass the music signal, but allow two different DC points to exist. These are directly in the signal path and a better cap means a better signal. 2) you discuss what is called a bypass capacitor on the power supply (B+ to ground). Broadly speaking we can think of capacitors (electrolytic primarily) that are efficient, with large capacity for their size, and cost, but are very nonlinear, especially as frequency rises. Alternatively there are various classes of capacitors that are the revers - large size, small value, but very linear and good sounding. Good design combines the two allowing one to provide capacitance and the others to provie linearity and efficiency at high frequencies. The best of these are film type capacitors (Mylar, or better yet poypropolene). Ceramic are also very good but have only tiny values or low voltages for the monolithic multilayer types (like << 50V).
You can pretty much know what will be the best sounding from one value: dialectic absorption. A Poly type is typically something like 100X to 1000X better than a "lytic.
As to whether they matter: yes. They make the power supply, well, supply, at high frequencies!
So get polyprop caps of >> whatever the applied voltage is (more is better, but gets $$ and big) and put them in. Poly caps are NOT polarized, but there is often a band that indicates the end to be grounded.
Justme.
