maneco said: "And i always thought that metal films had less thermal noise...but that could be a myth of "high end audio"..."
Thermal noise is thermal noise---the universe over. There is no such thing as excess thermal noise---in volts r.m.s. at a given temperature it is square root of 4kTRB (k = Boltzmann's constant in joules per kelvin degree, T = temp in degrees kelvin, R = resistance in ohms, and B = noise bandwidth in Hz). Or equivalently as a current noise in parallel with the resistor it is sq rt 4kTB/R, in amps r.m.s.
When you put voltage across a resistor and hence current through there can be excess noise, often with a 1/f spectrum. Metal films can be quieter in this regard---a lot depends on the quality of the end terminations.
Metal films with ionic contaminants can get nasty when there is d.c. across them and in the presence of moisture. A manufacturer in Thailand had something amiss and sent a bunch of parts of which several failed in the field in an amplifier for Ford. Unfortunately they were in deep denial and kept explaining the problem away as having loaded the wrong resistivity elements in the machine that cut the spiral groove to set the value---they claimed that it produced localized hot spots and burnout. They lost the business in the end.