Changing runtime while also affecting pitch is easy. Just adjust the playback rate. No information is added or lost. The image analogy holds in that case.
Its changing the run time without affecting pitch which is trickier, but made simple with the advanced tools now available.
I was about to say that the slight pitch change produced by changing the run time of one file to sync with the second is probably not significant enough to matter, and regardless of how it's done I expect that is the case. But you've got me thing through this again, and I now realize that actually we do want to change pitch as well as timing, and I may have been selecting the incorrect mode in Samplitude's stretch algorithm. I'll have to check that.. I may have been stretching to sync timing between two files, and unintentionally throwing the pitch off slightly, by instructing the algorithm NOT to allow any change of pitch change with the stretch. Instead, we want the pitch changed as well, to bring it the two slightly different pitches back into union. I'll walk through it below, but if this is correct, it ironically proves to me that the typical amount of time stretch required isn't significant enough to cause serious pitch problems, since I didn't notice any that I may have unintentionally introduced!
Scenario:
Two different recorders with unsync'd clocks record the same sound (with a pitch and duration), sampling at a slightly different rate. Upon playback directly from the same machines on which they were recorded, using the same recording clocks for playback as were used for recording, the sound from both will remain in sync once initially sync'd. I used to play back this way regularly, having trained myself to hear the slight delay or lack of it once the two were effectively in perfect sync, the more difficult task of properly identifying which was leading and which was lagging, and perfection of a very rapid "play/pause/play" jab.
The same two files, transferred off the original machines on which they were recorded, are then played back with timing provided by a third clock. The rate of the third differs slightly from both of the recording clocks. Both play back at a slightly different pitch and duration than the original. One is adjusted to match the other, and that accurate relative sync between the two is good enough. Absolute sync to the exact timing and pitch of the original (by adjusting the rate of the third clock) isn't perfectly maintained and isn't critical, only the relative timing of the two. Playback rate and pitch going to vary ever so slightly with each different playback device based on the difference between it's own clock (the new 3rd) and the clock of the original recorder (the 1st), the file of which was never changed.