From my experience, a light yellow color on plain carbon steel is about as high as I would go. That is above my tempering temperature for 1095, which is 325-335 F. Other steels take different colors at different temperature, and I havent done enough research to make any useful statements. Something to think about is that colors are a surface oxide, and stainless steels resist oxide growth past a certain point, so how hot is a stainless when it finally turns blue? Basically, what M Wadel described is the practical limit. Temperature has a MUCH greater effect than time, as an increase in tempering time from 1-4 hours only produces a loss of about 1 point of hardness. IIRC, this was a plain carbon steel, and alloy steels will be slower than that. Something thats been nagging me is how can heat treatment be "ruined" by say, sharpening on a sander, where the blade heats up then immediately cools down, when tempering times are generally 1 hour or more? Seems the heat treaters could save some time by bringing the piece to tempereture for a few seconds then immediately cooling it, assuming the piece took less than an hour to come up to temperature. I need to do some more digging on the matter.