Initializing... drag & drop files here
Supports: 3GPP
This walks you through pulling the audio track out of an old .3gpp mobile-phone video and saving it as a standalone AAC file. It is written for anyone who has a clip recorded on a 2000s-era feature phone and wants just the sound — a voice memo, a song fragment, an interview — without the low-resolution video wrapped around it. By the end you will have an AAC file that plays in almost any modern player, plus a clear sense of how good that audio can realistically be.
This is an extract, not a remux that keeps the picture: the video stream is discarded and only the audio is written out. If you want to keep the video, convert to MP4 instead.
.3gpp clip onto the page or click "+ Add Files" to pick it from your computer. You can queue several clips and they will all use the same settings.The single most useful thing to understand here is that 3GPP clips from old phones almost always carry AMR narrowband audio — a speech codec that samples at 8 kHz, covers only the 200–3400 Hz telephone band, and runs at 4.75–12.2 kbit/s. That is voice-call fidelity. Converting it to AAC makes the file portable, but AAC cannot add detail the phone never recorded — there is no high end to restore. So spending a high bitrate on AMR-sourced audio is wasted space.
3GPP is a container, so what comes out depends entirely on what the handset put in. Identify your source in this table and you will know what the extracted AAC can and cannot sound like before you run the conversion.
| Audio in the 3GPP | What it is | Sampling / band | Channels | Typical source rate |
|---|---|---|---|---|
| AMR-NB | 3GPP narrowband speech codec, the feature-phone default | 8 kHz, roughly 200–3400 Hz | Mono | 4.75–12.2 kbit/s |
| AMR-WB (ITU-T G.722.2) | Wideband speech, later and less common | 16 kHz, roughly 50–7000 Hz | Mono | 6.6–23.85 kbit/s |
| AAC-LC | General-purpose lossy codec on later handsets | Full band, limited by the encode | Mono or stereo | Commonly 64–128 kbit/s |
| No audio track | Video-only capture or a stripped clip | — | — | — |
Every one of those inputs is already lossy, so the extraction is a lossy-to-lossy step in all cases. A high AAC bitrate buys file size, never fidelity.
AAC is a codec, not a file type, and it ships inside several different wrappers. This page writes a .aac file — a raw ADTS stream, meaning bare AAC frames with lightweight sync headers and no real container around them. The table below is the map if a downstream tool wants something else.
| Extension | What it really is | Carries metadata / chapters | Best for |
|---|---|---|---|
.aac |
Raw AAC in an ADTS stream, no container | Minimal | Feeding encoders, quick playback in VLC or a modern player |
.m4a |
AAC inside an MPEG-4 container | Yes | Music libraries, iTunes/Apple Music, tagging |
.mp4 |
AAC alongside a video track in MPEG-4 | Yes | Keeping picture and sound together |
.3gp / .3gpp |
The mobile profile of the same MPEG-4 container | Limited | The original phone recording you started with |
If a clip is DRM-protected, corrupted, or only partially downloaded from a phone backup, extraction can fail or produce a truncated file — the audio track has to be intact and readable. A clip that was already silent in the original recording will produce a silent AAC file; there is nothing to recover. And if you actually want the recording at its best possible quality, there is no codec trick for that: an AMR source is permanently limited to telephone-band speech, and the honest ceiling is simply a clean, faithful copy of what the phone captured.
Your file is uploaded over an encrypted connection, the audio track is extracted on our servers, and both the original and the AAC output are deleted automatically a few hours after conversion. There is no sign-up, no watermark, and your files are never shared or made public.
Because the original recording was made with a phone-call codec. Most 3GPP clips from feature-phone-era handsets store AMR narrowband audio: mono, 8 kHz sampling, limited to the 200–3400 Hz speech band. AAC faithfully carries that across, but the treble and stereo width were never recorded, so they cannot reappear in the output.
No — and no online converter can. Conversion changes the container and codec so the file plays everywhere, but it cannot add information the source lacks. The realistic goal is a faithful, portable copy of the original audio, not an upgrade. In our testing, an AMR-sourced 3GPP clip extracted to AAC was indistinguishable from the original by ear, which is exactly the intended result.
No. They are the same 3GPP container — based on the MPEG-4 Part 12 (ISO base media) file format and defined by the 3rd Generation Partnership Project — just written with the longer extension. You can upload either here, or use the dedicated 3GP to AAC page; the extraction is identical.
Keep it modest. Because AMR narrowband tops out around 12.2 kbit/s and telephone-band fidelity, an AAC bitrate of roughly 64 kbps is plenty — higher settings only inflate the file. If you know your clip held AAC originally, match that bitrate instead of raising it, to avoid extra generation loss. Enter the figure under Constant Bitrate, or choose a band under Variable Bitrate; the Quality Preset dropdown is the hands-off alternative and the least predictable control here, and nothing is pre-selected in it, so leaving it on the "Highest" shown at load hands the choice to the server's own default.
AAC (Advanced Audio Coding) is the lossy audio standard defined in ISO/IEC 13818-7 (MPEG-2 Part 7) and ISO/IEC 14496-3 (MPEG-4 Part 3). It was designed as the successor to MP3 and generally sounds better than MP3 at the same bitrate. It plays natively across iOS, Android, YouTube, and most modern players, which makes it a reliable target for rescuing old mobile audio.
You get a .aac file — a raw ADTS stream rather than an MPEG-4 container. VLC, mpv, foobar2000, and most current desktop and mobile players open it directly, and encoders and command-line tools take it happily. What a bare .aac is not good at is library duties: it has almost nowhere to store artwork, track numbers, or chapter marks, so a music app may import it without tags. If you want a taggable file for a music library, convert the same clip to MP3, which every library application indexes cleanly.
Yes, by default. Audio Sample Rate and Audio Channel both start on "Original", so the extraction matches whatever the phone recorded — for an AMR-NB source that means 8 kHz mono, which is correct and not a fault. You can override either control to force a specific rate or downmix, but there is no benefit in resampling an 8 kHz speech recording up to 44.1 kHz: it inflates the file without adding any frequency content, because the extra bandwidth was never captured. Setting "Stereo" on a mono source is the same story — it duplicates one channel into two.
Not on this page — it is an audio-only extract. To keep the picture, convert to MP4. If you simply prefer a different audio format, 3GPP to MP3 outputs MP3 from the same source.