Why speed and pitch are linked by default
Playing audio faster or slower by simply changing its playback rate raises or lowers pitch as a direct side effect — this is physics, not a software limitation: a sound wave compressed in time literally has a higher frequency, and stretched in time has a lower one, which is exactly what pitch is. This is the classic “chipmunk effect” when speeding up, and the equally familiar “slow-motion drawl” when slowing down.
Time-stretching algorithms exist specifically to break this link — they change duration while independently preserving the original pitch, using techniques that analyze and reconstruct the audio's frequency content rather than just replaying it faster or slower. Modern time-stretching is good enough that moderate speed changes (within maybe 25-50% of original speed) are often hard to distinguish from the original pitch-wise.
When you actually want the pitch to change
Not every use case wants pitch preserved — a classic sound-design trick uses simple speed changes precisely because they alter pitch, and some music production techniques deliberately exploit the speed-pitch link for effect. If you want a natural-sounding result for speech or normal music playback, you want pitch preservation; if you're going for a specific creative effect, you might want the opposite.
Extreme speed changes still reveal artifacts
While modern time-stretching handles moderate speed changes cleanly, pushing to extremes — several times faster or slower than original — starts to reveal audible artifacts (a slight warble, robotic texture, or smearing) because the algorithm has to work with less real information for very fast speeds, or fabricate more for very slow ones. This isn't a flaw in any particular tool; it's an inherent limit of how much a time-stretching algorithm can reconstruct convincingly.
Practical uses beyond the obvious
Slowing down speech (podcasts, lectures, tutorials) to catch details on a first pass, or speeding it up to review content faster, are the most common practical uses — most people already do this natively in podcast and video apps without realizing it's the exact same speed-change technology being applied. Music practice is another common case: slowing down a song without changing its pitch lets musicians learn fast passages at a manageable tempo while still hearing the correct notes.