What bitrate should an internet radio station stream at?

Bitrate is the number of bits per second your stream spends describing the audio. Everything else follows from it: how good the station sounds, how much bandwidth it burns, how many listeners a connection can hold, and whether somebody on a train can listen without the player stalling.

It is also the setting people agonise over longest and change least often, usually on the wrong evidence — a comparison made on studio monitors in a quiet room, for an audience listening on a phone speaker in a kitchen.

What the number means

At 128 kbps, one second of audio is 128,000 bits, or 16 kilobytes. A three-minute song is about 2.8 MB on the wire. Double the bitrate and you double both.

Higher is not linearly better. Encoders spend their first bits on the parts of the sound humans notice most and their last bits on detail almost nobody hears, so the gap between 64 and 128 is enormous and the gap between 192 and 320 is, on most material and most equipment, inaudible. This is the whole reason a bitrate decision is a judgement rather than a maximisation.

What listeners actually notice

Roughly, on typical pop and rock material through consumer equipment:

The honest test is not whether *you* can hear the difference in the studio. It is whether your audience can hear it on the device they use, in the room they use it in, while doing the thing they do while listening. For most stations, that answer is no somewhere between 128 and 192.

What it costs

Bandwidth is bitrate multiplied by listeners multiplied by time, and it grows faster than people expect.

One listener at 128 kbps for one hour uses about 57 MB. At 192 kbps, about 86 MB. So a station averaging 30 concurrent listeners around the clock moves roughly 1.2 TB a month at 128 kbps and 1.9 TB at 192.

That difference is the actual bitrate decision. It is not "does 192 sound better" — sometimes, slightly — it is "is that improvement worth 50% more delivery, and what else could that budget buy". A station that spends its allowance on 192 kbps and then hits its limit two-thirds of the way through a month has made its audience's experience worse, not better, because unavailable is worse than 128.

Work out your own numbers with the bandwidth calculator, and the hours side with the listening hours calculator.

Format changes the answer

Bitrate is only comparable within a format. AAC at 96 kbps is roughly the quality of MP3 at 128, so a station that switches format can hold quality and cut a third of its bandwidth — or hold bandwidth and gain quality.

The catch is compatibility, and it is a real one for radio specifically: MP3 plays in every browser, car head unit, smart speaker, aggregator directory and cheap internet radio on the market. AAC plays in most of them. "Most" is a support burden when the ones it does not play in belong to your most loyal listeners. This is covered properly in AAC vs MP3 for radio streaming.

Sample rate and channels

Two settings sit next to bitrate and neither is worth much of your attention.

Sample rate should be 44100 Hz. It is what your music files already are, it is what every player expects, and resampling to 48000 for a stream adds a conversion for no benefit.

Channels should be stereo for music and can be mono for speech. Mono at 64 kbps genuinely sounds better than stereo at 64 kbps, because all the bits go to one channel instead of two — that is a real trade for a talk station on a tight budget. For music, stay in stereo.

Multiple streams

Running a high-quality and a low-quality stream side by side is the classic answer to "some of my listeners are on mobile data".

It is less compelling than it used to be. Mobile data is cheaper and faster than when the practice started, players do not switch between your streams automatically, and every extra stream is another URL to publish, another directory listing to maintain and another thing to be wrong. Most stations are better served by one well-chosen stream than two mediocre ones.

The exception is genuinely constrained audiences — a station with significant listenership on slow or expensive connections — where a 48 or 64 kbps stream is not a nicety but the difference between listening and not.

What we do

RadioHost Studio caps bitrate per plan: 128 kbps on a trial and up to 192 kbps on a station subscription. The cap is enforced at connection rather than by silently re-encoding, so a source configured above it is refused and you find out immediately rather than wondering why the stream sounds different from the file.

The cap exists for the reason described above: hours multiplied by bitrate is the real cost of delivery, and a plan that advertised unlimited both would be pricing on the assumption that you never use it.

Frequently asked questions

Is 128 kbps good enough for a music radio station?

For most stations, yes. It is the most common bitrate in online radio, artefacts are difficult to identify in normal listening conditions, and it costs roughly a third less bandwidth than 192. Move up if your audience is demonstrably listening on good equipment and your listening-hours budget has room.

Does a higher bitrate use more of my listening-hours allowance?

Listening hours count time, not bytes, so the hours themselves do not change. What changes is the bandwidth behind them — which is why plans that sell hours also cap bitrate: the two multiply.

Can I change bitrate on a live station?

Yes, but every connected listener is disconnected when the stream restarts, and some players do not reconnect on their own. Change it at a quiet hour, not during your best-attended show.

Should I upload my music at high bitrate even if I stream at 128?

Yes. Upload the best files you have. Encoding a 128 kbps stream from a 128 kbps source stacks two lossy encodes and sounds noticeably worse than one; the library should always be higher quality than the stream.