OPUS to AMR Converter

Converting Opus to AMR: What You Are Trading Away

Opus is a modern general-purpose codec — the IETF standardised it as RFC 6716 in September 2012, and it scales from 6 kbit/s narrowband speech all the way to 510 kbit/s stereo music. AMR is the opposite: a 3GPP speech codec from the late 1990s built to squeeze a phone call down a GSM channel, mono only, with a fixed 8 kHz (narrowband) or 16 kHz (wideband) sample rate. Converting Opus to AMR is a lossy-to-lossy pass through a speech codec, so plan on the result sounding like a phone call. Do it when a telephony platform, an IVR, a voicemail system, or a test harness demands an .amr file — not to "improve" a recording.

How to Convert OPUS to AMR

1. Upload Your OPUS File: Drag and drop the .opus file onto the page or click "Add Files" to browse. Several files can be queued and converted with the same settings.
2. Pick the Audio Codec: In the Audio Codec list choose AMR Narrow Band (8 kHz, the GSM/voicemail classic) or AMR Wide Band (16 kHz, HD-voice systems). Nothing is pre-selected, and this choice also fills the Constant Bitrate list and locks Audio Sample Rate to 8000 Hz or 16000 Hz.
3. Set the Quality Preset: Under File Compression, use the Quality Preset dropdown — Highest through Lowest — to pick where in the codec's mode ladder you land. Audio Channel offers Mono only, because AMR has no stereo mode.
4. Convert and Download: Click Convert and save the .amr file. No sign-up, no watermark.

Walk-through: Codec Choice, Then Quality Preset

The Audio Codec radio list on this page has exactly two entries, and neither of them starts selected. That matters more than it looks:

• Leave the codec alone and you get AMR-NB. The server falls back to narrowband for an .amr target when nothing is chosen, which is the safe default for legacy telephony gear.
• Choosing a codec is what populates the Constant Bitrate list. Before you pick Narrow Band or Wide Band, the Constant Bitrate dropdown renders with nothing in it — every rate entry is gated on a codec being selected.
• Use Quality Preset, not Constant Bitrate. The Constant Bitrate entries currently collapse: whichever AMR rate you choose, the value that reaches the encoder is clamped to the codec's ceiling, so an AMR-NB file lands on 12.2 kbit/s and an AMR-WB file on 23.85 kbit/s regardless. Quality Preset does spread across the codec's range — roughly 5 to 12 kbit/s for narrowband and 6 to 24 kbit/s for wideband — and the encoder then snaps that to the nearest legal mode.
• Trim before you convert if you only need a clip. The Trim group takes a Start time and a Duration in seconds or HH:MM:SS.sss, so you upload and encode less of a long recording.

Opus vs AMR-NB vs AMR-WB

AMR Mode Ladders and File Size

Both ladders are fixed by the codec — you cannot invent a rate in between. Sizes below are the payload arithmetic for one minute of continuous speech, before the one-byte header AMR adds to each 20 ms frame (about 3 KB per minute).

Common Errors and How to Fix Them

• "The Constant Bitrate dropdown is empty" — No codec is selected yet. Pick AMR Narrow Band or AMR Wide Band in the Audio Codec list and the matching rate ladder appears.
• "Every bitrate I choose produces the same file size" — That is the Constant Bitrate collapse described above. Switch to Quality Preset, which actually moves the encoded mode.
• "The output sounds muffled and thin" — Working as designed. AMR-NB keeps roughly 300–3,400 Hz of a single mono channel; anything Opus captured above or below that is discarded, and no bitrate brings it back.
• "My playback app rejects the file" — An .amr holding wideband audio starts with the #!AMR-WB magic instead of #!AMR, and players written only for narrowband will refuse it. Re-convert with AMR Narrow Band.
• "The stereo mix collapsed to one channel" — AMR has no stereo mode at all. The Audio Channel dropdown offers Mono only for that reason.
• "I wanted a smaller file, not a worse one" — For speech that still has to sound like speech, Opus to MP3 at a low bitrate is a far better trade than AMR.

When Opus to AMR Is the Wrong Move

If nothing downstream is actually demanding an .amr file, this conversion only costs you. AMR was designed for a 1990s mobile voice channel: mono, one narrow slice of the spectrum, no music. Ringtones, podcasts, song clips, and anything with instruments come out sounding broken. Use it for telephony platforms, IVR prompts, voicemail imports, MMS pipelines, and test fixtures for telecom software — and pick a general-purpose target otherwise. Going the other way later is possible but does not restore anything: AMR to MP3 re-encodes whatever AMR kept, and the discarded frequencies stay discarded.

Frequently Asked Questions

Why does my Opus voice note sound like a phone call after converting to AMR?

Because AMR is a telephone codec, and the conversion throws away everything outside the phone band. AMR-NB samples at 8 kHz and encodes roughly 300–3,400 Hz in a single mono channel; Opus can carry the full audible range in stereo. Folding one into the other discards the highs, the lows, and the second channel. That muffled result is the codec working exactly as specified, not a fault in the conversion.

Should I pick AMR Narrow Band or AMR Wide Band?

Match whatever the receiving system accepts. Narrow Band is the classic 8 kHz format that GSM and 3G voicemail, MMS voice, and most legacy IVR expect, and it is the safest choice when you are unsure. Wide Band runs at 16 kHz with a 50–7,000 Hz band — noticeably fuller — and is what HD-voice systems use, but older equipment rejects it outright because the file even starts with a different magic number.

Why does the Constant Bitrate dropdown have nothing in it?

Every entry in that list belongs to a specific codec's mode ladder, and the list only fills once you have selected AMR Narrow Band or AMR Wide Band in the Audio Codec group. Nothing is pre-selected there, so on a freshly loaded page the Constant Bitrate control looks empty. Pick the codec first.

Does raising the AMR bitrate actually improve the recording?

Only within the codec's narrow ceiling, and only if the control reaches the encoder. Moving from 4.75 to 12.2 kbit/s on AMR-NB does sharpen speech inside the telephone band, but no AMR mode reaches above ~3,400 Hz for narrowband or ~7,000 Hz for wideband. On this page the Quality Preset dropdown is the control that reliably moves the encoded mode; the Constant Bitrate list currently clamps to the codec's top mode whatever you choose.

What happens if I hand the encoder a bitrate AMR does not support?

It refuses to invent one. We checked the two encoders this converter uses — the OpenCORE AMR-NB encoder and the VisualOn AMR-WB encoder — and both log "bitrate not supported: use one of …" and fall back to the nearest legal mode in their ladder. That is why an out-of-range request lands on 12.2 kbit/s for narrowband and 23.85 kbit/s for wideband rather than failing the job.

How big is a one-minute AMR file?

Small, which was the entire point of the format. At AMR-NB's top 12.2 kbit/s mode the payload works out to about 92 KB per minute, plus roughly 3 KB per minute of per-frame headers; the lowest 4.75 kbit/s mode is under 40 KB per minute. AMR-WB at its 23.85 kbit/s ceiling runs about 179 KB per minute of payload. Compare that with a 48 kHz stereo WAV at roughly 11.5 MB per minute.

Can I convert music or a ringtone to AMR?

You can, but you should not. AMR's encoder models the human vocal tract, so instruments, cymbals, and reverb come out smeared and metallic — the codec has no way to represent them. For a small music file, use Opus to MP3 instead; only use AMR when the destination system genuinely requires the format.

Why is the output always mono, and can I keep the sample rate at 48 kHz?

No to both, and neither is a setting we hide. AMR is specified as a single-channel codec, so the Audio Channel dropdown exposes Mono alone. Its sample rate is fixed by the codec too: 8000 Hz for Narrow Band and 16000 Hz for Wide Band, which is why choosing a codec locks the Audio Sample Rate dropdown to exactly one value. An Opus stream always decodes at 48 kHz, so the conversion resamples down to whichever AMR rate applies.

How are my files handled during conversion?

Your file is uploaded over an encrypted connection, encoded on our servers, and deleted automatically a few hours after conversion — no sign-up, no watermark, never shared or made public. The practical limit is upload size and time rather than anything about the conversion itself. If you only need part of a long recording, trim it first with the Audio Cutter so there is less to send.