Private browser-based audio compression

Audio Bitrate Changer

Change supported audio to a controlled 64–320 kbps delivery copy in your browser. The tool measures the source average bitrate, expands each speech, podcast, sharing, or higher-quality preset into real settings, estimates payload size, warns about up-bitrating and generation loss, and verifies the produced bitrate.

Local processing: this file, its name, tags, artwork, and audio content are not uploaded.
Status Choose audio file

Output

InputOutput
File size--
Container--
Codec--
Average bitrate--
Sample rate--
Channels--
Duration--
Metadata in output--
Processing time--

What changing bitrate actually does

Bitrate sets an encoded-data budget per second for lossy output. A lower target often makes a smaller file because the encoder has fewer bits to represent the same duration. The encoder decides which detail can be approximated or discarded, so size and audible trade-offs change together.

This page creates MP3, M4A/AAC, or OGG/Vorbis because the reviewed build exposes bitrate control for those outputs. It does not present WAV or FLAC as bitrate targets. After encoding, ffprobe measures the result instead of treating the selected number as the actual average.

Recommended bitrate ranges for speech, podcasts and music

These are starting points, not objective quality scores. Content, codec, listening conditions, and the source encode all matter. Presets remain editable and show their format, channel count, and bitrate so the label never hides the actual command.

UseStarting pointChannelsWhat to check
Speech copy64 kbpsMonoConsonants and background noise
Podcast96 kbpsMonoVoice clarity and music beds
Music sharing192 kbpsStereoHigh frequencies and stereo detail
Higher-quality delivery256 kbpsStereoWhether the source justifies the size

Why up-bitrating does not improve a low-bitrate source

If the measured source average is about 64 kbps, writing 256 kbps output cannot recover frequencies or spatial detail removed earlier. The new encoder only receives decoded samples from the existing source. It may preserve that decoded version with fewer additional losses, but it is not a restoration process.

The estimate marks a target above the source as possibly larger. This is useful when a platform requires a particular format, but it should not be described as an audio enhancement.

CBR, VBR and actual average bitrate

A selected bitrate is an encoder target. Container headers, tags, padding, and rate-control behavior mean the final file size and measured average can differ. Variable bitrate allocates more data to complex passages and less to simple ones; constant-rate workflows aim for steadier allocation.

The payload estimate is duration × target kbps × 1000 ÷ 8. It deliberately excludes an exact overhead claim. The result card separately shows estimated bytes, actual file bytes, and actual average bitrate.

Re-encoding and generation loss

MP3, AAC, and Vorbis are lossy. Decoding and encoding a lossy source again can add artifacts even at the same target bitrate. Keep the best available source and make delivery versions from it when possible instead of repeatedly converting a previous delivery copy.

A/B preview helps you decide whether the result is suitable. The two players pause each other, and the page attempts to align playback position when switching, but a listening comparison remains subjective and is not reported as a quality score.

File-size examples and when to use the compressor

At 128 kbps, ten minutes of encoded audio payload is about 9.6 MB before container overhead. At 64 kbps the same duration is about 4.8 MB; at 320 kbps it is about 24 MB. Actual results vary, which is why the output is measured.

Use the main compressor when you want a quality preset or a target-MB workflow. Use this changer when a platform, archive convention, or production handoff specifies a bitrate and you need the source comparison to remain visible.

FAQ

Does 320 kbps always sound better?

No. The source, codec, and listening conditions matter. A 320 kbps re-encode cannot restore information already lost.

How is estimated size calculated?

Audio payload bytes equal duration in seconds multiplied by target kbps and 1000, divided by 8. Container and metadata overhead are separate.

Why is the actual bitrate different from my target?

Encoder rate control, stream complexity, headers, and tags can make the measured average differ from the selected target.

What is generation loss?

It is additional degradation caused by decoding and encoding an already lossy source again.

Should speech be mono?

Often, when spatial information is not needed. Listen first and keep stereo when it carries meaningful content.

Can the output become larger?

Yes. A target above the source average or additional container data can increase file size.