Compress Audio

Bitrate is the default mode. You choose how many kilobits per second the sound may use, from 32 to 192, and the size follows from the length of the recording. It is the better mode when you care about how it sounds and want every file in a batch to have the same quality, such as a series of podcast episodes or an album.

Target size works the other way round. You type the size the file must fit under, in MB or KB, and the compressor picks the bitrate that gets there. Use it when a limit is printed somewhere: an email provider, a school or job portal, a voicemail system or a messaging app. Type a value a little below the limit, since some services count megabytes slightly differently.

The arithmetic is simple enough to do in your head. One minute of sound at 128 kbps is 7,680 kilobits, or a little under 1 MB. Half the bitrate gives half the size, and twice the length gives twice the size. That also shows the limits of target size: a long recording forced into a small target ends up at a very low bitrate, and no setting can hide that. In that case, trim the recording or switch to mono first.

Speech is narrow. Most of what makes a voice understandable sits in a limited range of frequencies, and a single speaker needs only one channel. That is why lectures, interviews, meeting recordings and voice notes compress so well: mono at 64 kbps, or even 32 kbps for dictation, keeps every word clear.

Music is wider and more demanding. Cymbals, strings, reverb and stereo width all take data, and low bitrates show up as a swishy, underwater sound on those parts. Keep music in stereo and at 96 kbps or higher. If a music file must become very small, Opus holds up better than MP3 at the same bitrate; its specification, RFC 6716, was written for both speech and music over the internet.

The checklist below covers the settings that make the most difference, roughly in order of how much they help.

A recorded lecture or meeting to share. These are long, so the bitrate matters most. MP3 or M4A at 64 kbps in mono keeps the speaker clear and turns an hour-long WAV into a file that fits comfortably in a cloud folder or a course platform. If the recording was made far from the speaker, run it through the Audio Converter page with Normalize volume afterwards.

A podcast episode for a hosting service. Many podcast hosts publish limits and preferred formats in their help pages; check them first. MP3 is the most common requirement. A voice-only show at 64 to 96 kbps mono sounds clean; a show with music beds can stay in stereo at 128 kbps.

A voice note or interview to send by email. Use Target size with a number just under the provider's attachment limit. Short clips will barely be touched, long ones get a lower bitrate automatically.

Songs for a phone with little storage. Convert a lossless library to M4A or Opus at 128 kbps in stereo. You keep the originals elsewhere, and the phone gets copies that sound very close at a fraction of the size.

Many files at once. Up to 20 files can go in one batch, each up to 95 MB, and all of them get the same settings. In Target size mode every file is fitted to the same limit on its own; the limit does not apply to the batch as a whole.

Output format offers MP3, M4A, Opus, OGG or the same format as the source. MP3 is the default because it plays everywhere, from old car stereos to web forms that accept audio. M4A holds AAC audio, plays on Apple devices and nearly all phones, and sounds better than MP3 at low bitrates. Opus is the most efficient of the four, ideal for voice messages and web use, but some older players and editors do not open it. OGG usually means Vorbis audio and is common in games and open-source software.

Keeping the same format is useful when software expects a particular type, for example a podcast host that asks for MP3 or a game engine that wants OGG. Note that a lossless source such as WAV or FLAC cannot stay lossless and still get much smaller; picking Same as source for those formats limits what compression can do. For a plain change of format without aiming at a size, Audio Converter offers more targets, including the lossless ones.

Every lossy compression removes something, and doing it twice removes more. Always start from the best copy you have: the original WAV from a recorder, the file as it came from your phone, or the master of a song. Compressing an MP3 that was already compressed for a messaging app gives a worse result than compressing the original to the same size.

Listen before you send. Play the busiest part of the compressed file, such as a chorus or a moment where several people talk at once, on the kind of device the listener will use. If it sounds thin or watery, raise the bitrate one step or switch to Opus or M4A.

Be careful with the sample rate on music. Dropping a song to 22.05 or 16 kHz removes the high frequencies that give cymbals and voices their air, and no bitrate can bring them back. For speech, the same change is barely noticeable and saves real space. The default, Same as source, is the safe choice whenever you are unsure.

Recordings that live inside a video are a separate case. To make a video lighter, use the Compress Video tool; to keep only its sound, the Video to Audio tool extracts the track, which is usually a small part of the original file.