Over the permissable range it doesn't make a huge difference with incandescent's as they're NTC devices - the hotter the fillament the higher its resistance so they effectively regulate themselves. When I was a kid I built a Wien bridge oscillator which actually exploits this in its feedback loop.

However if you put an incandescent bulb on a dimmer then you lose this effect which can be a little annoying for e.g. mood lighting when the grid voltage isn't stable or the electical installation dosn't meet vd requirements.

And you need to rate the semiconductors in the dimmer for the cold inrush current - as when very dim, the filament resistance never rises to the final value. For tungsten lamps figures  of 8:1 to 10:1 come to mind,  so a 60watt lamp (1/4 amp, ~ 1000 ohms when hot) has a cold resistance of about 120 ohms and needs at least a 2A triac.  The life shortening is not negligible though as high efficiency filaments run quite close to failure - 90% volts is almost double life expectancy and 120% is near instant failure.
As this graph illustrates.

From Vaughn and Hughes 'Lamps and Lighting'

Mike

One of my friends at uni went on to work in broadcasting.

He told me that the stage lighting is dimmed down to imperceivable lumens for this very reason – stage lights are very expensive, especially if one fails during filming and the director notices. Keeping the lamp warm is terrible efficiency but means that during filming the failure rate is almost zero.

This is also why bulbs are rated where they are. If you look at that chart you might wonder why people don't deliberatly drop the voltage to get a longer bulb life. Well it would cost you more in energy than the cost of the bulb. Contrary to popular opinion it was never a conspiracy by bulb manufacturers to sell more bulbs - it was a very sensible energy efficiency measure.

One of my friends at uni went on to work in broadcasting.

He told me that the stage lighting is dimmed down to imperceivable lumens for this very reason – stage lights are very expensive, especially if one fails during filming and the director notices. Keeping the lamp warm is terrible efficiency but means that during filming the failure rate is almost zero.

This is also why bulbs are rated where they are. If you look at that chart you might wonder why people don't deliberatly drop the voltage to get a longer bulb life. Well it would cost you more in energy than the cost of the bulb. Contrary to popular opinion it was never a conspiracy by bulb manufacturers to sell more bulbs - it was a very sensible energy efficiency measure.

And photoflood lights were run at the white hot end of the curve and had a full power life of tens of hours.

There is a whole science of this sort of lighting slowly being forgotten and going the way of the gas mantle (which oddly you can still buy new).

Mike