Initializing... drag & drop files here
Supports: OPUS
Opus is already one of the most efficient audio codecs in existence, so most .opus files are small to begin with — there is far less to squeeze out than with a WAV or a 320 kbps MP3. Compressing an Opus file means re-encoding it at a lower bitrate, which costs you a generation of quality because the source is already lossy. Use this when you have to meet a hard upload or attachment cap, or to shrink long voice recordings where some quality loss is acceptable. Files are uploaded over an encrypted connection, processed on our servers, and deleted automatically a few hours after conversion — no sign-up, no watermark, never shared.
.opus track or click to browse. Batch uploads are supported, so you can shrink a folder of voice notes in one pass.Opus chooses its internal mode (SILK for speech, CELT for music) automatically, so the right number depends mostly on what is in the file. Figures below follow Xiph's published recommendations; "size per hour" is the resulting file for one hour of audio.
| Target bitrate | Size per hour | Best for | Quality notes |
|---|---|---|---|
| 96–128 kbps | ~43–58 MB | Stereo music, podcasts with music beds | ~128 kbps VBR is "pretty much transparent" for music per Xiph; going higher is wasted |
| 64–96 kbps | ~29–43 MB | Streaming-grade music, mixed content | Xiph's streaming/radio range; clean on earbuds and phone speakers |
| 48–64 kbps | ~22–29 MB | Interviews, group voice, stereo speech | Comfortable for multi-speaker recordings |
| 24–32 kbps | ~11–14 MB | Single-voice mono — narration, voice notes | Xiph's clean-voice range; intelligible, not for music |
| 16–24 kbps | ~7–11 MB | Wideband VoIP-grade speech | "HD voice" territory; audible artefacts but fully understandable |
Below roughly 24 kbps, music falls apart but a single mono voice can still be understood — that is the floor where Opus is genuinely better than older codecs. If your file is already at or below 32 kbps (typical for WhatsApp, Telegram, or Discord voice notes), there is little headroom left; trim the recording or accept that re-encoding will be audible.
Because Opus is already extremely efficient, most .opus files are encoded near their content's transparency point — around 128 kbps for music and 24–32 kbps for voice notes. There is little redundant data left to remove, so dropping the bitrate further trades audible quality for only a modest size cut. If the file is already a low-bitrate voice note, the biggest win usually comes from trimming silence with the Audio Cutter, not from re-encoding.
Yes. Opus is lossy, so re-encoding it at a lower bitrate discards more audio data and stacks a second generation of compression artefacts on top of the first. For spoken word the loss is often inaudible down to about 24 kbps mono; for music the practical floor is higher, around 96 kbps, before cymbals and reverb tails start to smear. If you have the original lossless or higher-bitrate source, encode straight from that instead of from the existing Opus.
For a single mono voice — narration, a lecture, a voice memo — Xiph's clean range is 24–32 kbps, and wideband VoIP-grade speech holds up at 16–24 kbps. For stereo music, stay at 96–128 kbps, where Opus is effectively transparent; 64–96 kbps is fine for casual listening. Interviews and group recordings with several speakers sit comfortably at 48–64 kbps. Picking a music bitrate for a voice note just wastes space; picking a voice bitrate for music will sound thin.
Choose Specific file size and enter your target in MB; the encoder back-solves the bitrate from the clip's duration. This is handy for fitting a long voice recording under an attachment or message cap. Aim slightly under your target to leave room for container overhead, and remember that switching Audio Channel to Mono on a stereo voice file roughly halves the size before you even touch the bitrate.
Usually no — Opus is more efficient than MP3 at every bitrate, so converting to MP3 to "save space" will make the file larger for the same quality. Convert to MP3 only for compatibility with a player or device that cannot read .opus; in our testing a 5-minute mono Opus voice note re-encoded to MP3 at a matching quality came out noticeably bigger than the original. If compatibility is the real goal, use Opus to MP3 rather than this compressor.
Files are uploaded over an encrypted connection, processed on xconvert's servers, and deleted automatically a few hours after conversion. The audio is never shared, never made public, and is not used to train any model — no account or email is required. To shrink a mix of formats in one job, the broader Audio Compressor handles MP3, WAV, M4A, OGG, FLAC, and Opus together.