Bitrate is how much data per second an encoder may spend describing your audio — more kilobits, more fidelity, bigger file. The honest part most converter sites skip: past a certain point, more bitrate buys you nothing you can hear. This guide tells you where that point is for each use.
The practical ladder
64 kbps (mono): the podcast standard. For a single voice, this is genuinely transparent — a 1-hour episode in ~28 MB. Music at 64 kbps sounds visibly compressed; don't.
128 kbps: the historic 'MP3 quality'. Fine on phone speakers and for casual background listening; on good headphones, cymbals and reverb tails give it away.
192 kbps: the sweet spot. In blind tests, most listeners on most equipment cannot reliably tell 192 kbps MP3 from the CD. This is Sonvert's default for a reason.
256–320 kbps: insurance. For archiving from lossless sources, for treated rooms and trained ears. The difference from 192 is subtle to nonexistent — but storage is cheap, so archiving high is reasonable.
VBR (variable bitrate): the efficient choice. The encoder spends bits where the music is complex and saves them where it isn't — VBR quality 0 averages ~245 kbps and matches CBR 320 in quality at ~25% less size.
The one rule that matters
Output quality can never exceed input quality. A 128 kbps source exported at 320 kbps is a bigger file with identical sound — the detail the first encoder discarded is gone forever. Match or slightly exceed the source bitrate and you've kept everything there is to keep.