What audio bitrate should you use?
Every audio converter asks you to pick a number — 128, 192, 320 — usually with no explanation of what it does. It decides two things at once: how large the file is, and how close it sounds to the original. Choosing badly wastes either disk space or sound quality, and it is genuinely hard to tell which from the number alone.
The short answer
Music you care about: 320 kbps.
Music in general, podcasts with production, most things: 192 kbps.
Speech — lectures, interviews, meetings, voice notes: 128 kbps.
Anything you will edit later: do not use a lossy format at all.
If you stop reading here you will make the right call almost every time. The rest explains why, and covers the one mistake that catches nearly everyone.
What a bitrate actually measures
Bitrate is how many bits of data are used to describe one second of audio. 192 kbps means roughly 192,000 bits — about 24 kilobytes — per second. Multiply by duration and you have the file size, near enough:
A four-minute song at 192 kbps: 4 × 60 × 24 KB ≈ 5.8 MB
The same song at 320 kbps: 4 × 60 × 40 KB ≈ 9.6 MB
A one-hour lecture at 128 kbps: 60 × 60 × 16 KB ≈ 57 MB
This is why bitrate is the only setting that matters for file size. Length and bitrate determine it entirely; the content makes almost no difference.
Why more bits sound better
MP3 and similar formats work by throwing information away — deliberately, and cleverly. They model what human hearing can actually detect and discard the rest: a quiet sound occurring just after a loud one at a similar frequency is inaudible to us, so it need not be stored.
A higher bitrate simply means the encoder has to discard less. At 320 kbps it can keep nearly everything that matters. At 96 kbps it is forced into aggressive choices, and you start to hear them — usually as a faint swirling or watery quality on cymbals, applause and reverb tails. Those are the hardest sounds to compress because they are close to noise, and noise is exactly what these formats are worst at.
That is also why the difference is easiest to hear on detailed acoustic music and hardest on speech. A voice is a narrow, predictable signal. There is not much to throw away in the first place.
Where the ceiling is
Repeated listening tests over the past two decades have landed in roughly the same place: at 256 kbps and above, most listeners cannot reliably tell a well-encoded MP3 from its lossless source, even with good headphones and careful comparison. Trained listeners on revealing equipment sometimes can, on particular recordings.
So 320 kbps is not audiophile territory — it is the top of the format, and it is genuinely transparent for practical purposes. There is nothing above it to reach for, because MP3 does not go higher.
The practical consequence: if you have the storage, 320 kbps for music costs you nothing but space. If you do not, 192 kbps is a very small compromise.
The mistake almost everyone makes
You cannot exceed the quality of the source. If a file was encoded at 128 kbps, converting it at 320 kbps does not recover anything. The detail was discarded when the original was made and no longer exists anywhere. You get a file two and a half times larger containing exactly the same sound — slightly worse, in fact, because it has now been compressed twice.
This comes up constantly with video soundtracks. The audio in an MP4 is often 128 or 160 kbps to begin with, soextracting it at 320 kbps just inflates the file. Match the source or go slightly above it, not far above.
It applies to every lossy-to-lossy conversion:M4A to MP3,OGG to MP3, and so on. If you know the source bitrate, match it. If you do not, 192 kbps is a safe default that rarely wastes much.
When the source is lossless
Converting from FLAC orWAV is the one case where a high bitrate is fully justified. The encoder is working from the complete original signal rather than from something already degraded, which is exactly the situation these formats were designed for. 320 kbps here produces about as good an MP3 as can exist.
It is also the only conversion you can repeat freely. A lossless original stays perfect no matter how many files you make from it, so keep it and treat the MP3 as a disposable copy for a particular device.
Formats other than MP3
Bitrate numbers are not comparable across formats. AAC (what anM4A usually contains) and Vorbis or Opus (inside anOGG) are more efficient than MP3, so they sound equally good at a lower number. A 128 kbps AAC is roughly comparable to a 160–192 kbps MP3.
This is why an MP3 converted from an M4A is often larger than the file it came from while sounding no better. That is not a fault in the conversion; it is the price of a format that plays on everything.
WAV sits outside this entirely. It is uncompressed, so it has no meaningful quality setting — its bitrate is fixed by sample rate and channel count at about 1,411 kbps for CD-quality stereo. That is why the quality control disappears when you choose WAV as an output.
A rule of thumb worth keeping
Ask what the file is for. Archiving or editing: stay lossless. Listening on good equipment: 320 kbps. Listening on a phone, in a car, in a gym: 192 kbps is indistinguishable in those environments. Speech of any kind: 128 kbps, and the file will be small enough to email.
The one thing not worth doing is choosing 320 kbps reflexively for everything. On a two-hour interview it produces a 288 MB file that sounds precisely like the 115 MB one.