Video to AMR Converter

How to Convert Video to AMR Online

1. Upload Your Video File: Drag and drop or click "+ Add Files" to select MP4, MOV, MKV, AVI, WebM, FLV, 3GP, MTS, WMV, or any of the 36+ supported video formats. Batch upload is supported.
2. Pick Constant Bitrate (or Quality Preset): AMR-NB only encodes at eight fixed rates — 4.75, 5.15, 5.90, 6.70, 7.40, 7.95, 10.2, and 12.2 kbit/s. The default 12.2 kbit/s matches GSM-EFR toll quality; pick 7.40 kbit/s for a noticeably smaller voicemail-grade file, or 4.75 kbit/s for the absolute smallest size at the cost of audible artifacts.
3. Set Audio Sample Rate, Channel, and Trim (Optional): AMR-NB is locked to an 8 kHz sample rate and mono channel by spec — the dropdown surfaces these so you can confirm, but other values get resampled/downmixed automatically. Use the Trim controls (start time + duration in HH:MM:SS.ms) to extract only the spoken segment instead of the entire video soundtrack.
4. Convert and Download: Click "Convert". Files process on our servers and the audio track is extracted, resampled to 8 kHz mono, and encoded with the FFmpeg libopencore_amrnb codec — no sign-up, no watermark, no install.

Why Convert Video to AMR?

AMR (Adaptive Multi-Rate) is a narrowband speech codec adopted as the 3GPP standard in the late 1990s and still baked into GSM, UMTS, and most carrier voicemail systems. It throws away everything outside the 200-3400 Hz human-voice band, so a one-minute spoken recording at 12.2 kbit/s lands around 90 KB — roughly 10x smaller than the same minute as a 128 kbit/s MP3. Pulling AMR audio out of a video file is useful whenever the destination is voice-only and bandwidth- or storage-constrained.

• Voicemail and IVR upload — Carrier voicemail systems, Asterisk/FreeSWITCH PBXes, and Twilio's <Play> verb all accept AMR (or AMR-WB) directly without a transcode step. Recording a greeting on your phone camera and converting to AMR keeps the upload payload tiny.
• MMS and legacy 3GPP messaging — MMS body audio is constrained to AMR by most carriers; extracting AMR from a 3GP or MP4 capture avoids the carrier-side downsampling that degrades MP3/AAC.
• Field interviews and transcription source — Journalists and researchers transcribing hours of interview footage can drop the file to AMR for storage. Whisper, AssemblyAI, and most speech-to-text APIs accept AMR or auto-convert it on ingest.
• Embedded and IoT speech playback — ESP32-style microcontrollers, automotive head units, and older Android devices have hardware AMR decoders but limited storage; a 1-hour lecture fits under 60 MB.
• Forensic and legal voice archives — Long-term retention policies for call center recordings often standardize on AMR-NB precisely because the codec is patent-stable, mature, and produces predictable file sizes per hour.
• Ham radio and SDR voice exchange — AMR's narrow 200-3400 Hz bandwidth aligns with SSB voice channels, making it a natural archive format for QSOs and digital voice modes.

AMR vs MP3 vs Opus — Format Comparison

AMR-NB Bitrate Mode Quick Guide

Frequently Asked Questions

Why does my AMR file sound so muffled compared to the original video?

AMR-NB is a narrowband codec — it deliberately discards every frequency above ~3400 Hz, which is where sibilants (s, sh, t), music harmonics, and brightness live. The codec was tuned for the 1990s GSM voice channel, not for music or wideband speech. If the video had music or wideband audio, that information is gone after the conversion. For wideband speech that still sounds natural, use AMR-WB (sometimes called AMR+ or G.722.2) instead, or pick Opus at 24 kbit/s.

Should I pick 12.2 kbit/s or a lower rate?

12.2 kbit/s is the default because it matches GSM Enhanced Full Rate (GSM-EFR) and is the only AMR-NB mode 3GPP labels as "toll-quality" out of the eight modes. The audible difference between 12.2 and 7.40 kbit/s on a clean voice recording is small; below 7.40 kbit/s artifacts become obvious. Pick 7.40 kbit/s if you're targeting voicemail-style storage, 4.75 kbit/s only if total file size is more important than intelligibility.

Will my MP4's stereo and 48 kHz audio survive the conversion?

No, and that's by design. AMR-NB is mono only and sampled at exactly 8 kHz — the encoder downmixes stereo to mono and resamples 44.1/48 kHz down to 8 kHz before encoding. If you need to preserve stereo or higher sample rates, convert to MP3, AAC, or Opus instead. See Convert Video to MP3 for music-grade extraction.

Can I play AMR files on iPhone or Android?

Android has had native AMR-NB playback support since Android 1.0 (it's part of the baseline media framework). iOS will play AMR through QuickTime and most third-party players, but iMessage and the stock Music app don't expose AMR in the share sheet. For cross-platform sharing, MP3 or AAC is friendlier. To go the other direction once it's on your phone, use AMR to MP3 or AMR to WAV.

What's the difference between .amr and .3ga file extensions?

Both wrap the same AMR-NB or AMR-WB bitstream — .amr is the original 3GPP storage extension, and .3ga is the audio-only variant of the 3GP container that Samsung phones started using around 2010. ffmpeg and VLC treat them identically. If a player rejects an .amr file, renaming to .3ga (or vice versa) sometimes fixes the issue without re-encoding.

Is AMR still patented? Are there licensing fees?

The AMR-NB patent pool is administered by VoiceAge, with patents from Nokia, Ericsson, NTT, and others. Per VoiceAge's published terms, decoder licenses are free for personal-computer software, while encoder licenses run roughly $0.30-$0.40 per unit for hardware shipments with volume discounts. The original AMR-NB patents have largely expired (most filings date to 1998-1999, putting them past the 20-year mark), but pool-administered licensing programs may still apply to certain implementations — check VoiceAge directly if you're shipping commercial hardware.

What's the maximum file size I can upload?

Free uploads are limited to 400 MB per file; Basic raises the video cap to 1 GB and Pro to 4 GB. AMR output files are tiny (a 10-minute video typically produces a 1-2 MB AMR), so the cap matters mostly for the source video upload, not the result.

How do I extract only the spoken section from a long video?

Use the Trim controls in Advanced Options before converting — set the start time in HH:MM:SS.ms format and the duration of the section you want. The video is decoded, the selected audio range is extracted, then encoded to AMR. Alternatively, trim the video first with the Video Cutter and feed the shorter clip into this converter.

Why isn't AMR-WB (16 kHz wideband) available as an option?

This page targets AMR-NB (.amr extension) because it's the dominant variant in deployed voicemail, MMS, and legacy 3GPP systems. AMR-WB uses a different bitstream and .awb extension. If your destination system specifically requires AMR-WB (sometimes labeled G.722.2 or HD Voice), let us know — meanwhile, Opus at 16 kHz is a more modern wideband-speech alternative with no licensing concerns.