> bet money that for the same technology both wattage and max voltage will be lower for a smaller part.
Of course... but the cost of materials and the general shrinkage of gear means many newer parts are NOT the same old painted carbon technology. I've handled resistors smaller than a traditional 1W part yet rated 3W. What happens is they run 3 times hotter. That would burn the paint off, so they use ceramic glaze. Thermal expansion would be bad, so they diddle the leg-metal alloy and the body construction so it stays together at red-heat.
If you read the very fine print: the voltage-limit may or may not be lower. Carbon-composition had a real limit before you got arcs between the individual particles of carbon. And the general direction of modern electronics means "nobody" cares about voltage rating: everything is 3.3V or lower. You found specs suggesting 50V/mm, but do they actually make trouble past there or is it just all they are tested for?
And the peak transient power (critical for some B+ RC filters) is a function of both temperature limit and total amount of stuff (mass times specific heat). Even with higher temp limit, there is no replacement for "big".
For the general run of audio (not tubes, not power amps), I see no problem with any resistor big enough to pick up with your fingers. Tube preamp plate resistors should be calculated. Preferably on FULL supply voltage (so they don't burn the first time a tube shorts). Tube amp B+ dropping/filtering resistors should be grossly oversize, generally a large 5W part. Power amps have many small resistors and a few big ones.
And in DIY, it is better to super-size than to miniaturize. Your labor is worth more than the resistor. It is also wise to shop specialist suppliers: tube guitar-amp shops stock "full size" resistors because that's what their customers want and need.