Which fade curve is best for podcast narration?
Logarithmic usually works better than linear or equal-power for narration, because it rises fast right at the start (√t), avoiding the "dropout" effect a too-slow entrance causes on speech. Equal-power is still the right choice for background music, where keeping constant power matters more than a fast entrance.
Why did my fade in and fade out end up shorter than I set?
Because the maximum duration of each fade is always half of the span it applies to, up to the absolute 60-second cap. On a 20-second clip, asking for a 15s fade in and a 15s fade out results in both being limited to 10s each, because together they cannot exceed half the total duration.
Can I apply a fade to just one segment in the middle of the episode, without touching the rest?
Yes, by setting a region. Everything outside it gets gain 0 (silenced) and the chosen curve applies normally inside it, so it is possible to insert a 30-second music bed in the middle of a 40-minute episode with a smooth fade at both ends, without cutting or re-exporting the rest of the file.