Extract the Audio From a Video on Android
You want the sound from a lecture recording or a video someone sent you, not a 300 MB upload to a website. Here is how to pull the audio out on the phone.
The audio in a video file is already a separate track. An MP4, MKV or MOV is a container holding a video stream, an audio stream and an index that ties them together, so getting the sound out means one of two things: copying that audio track out of the container untouched, or re-encoding it to MP3 — both small jobs that run fine on a phone, with no need to hand a 300 MB file to a website.
Convertlix for Android does it on the device: you pick the video through Android's own file picker, the work happens on the phone's own processor, and the file is never uploaded, so there is no server copy of your lecture to retain. It needs Android 8.0 or later and one sign-in — there is no guest mode — after which the tools keep working with the network off, which is the honest test. The free tier is metered by a monthly allowance the app shows you and ceilings video at 100 MB per file; Premium raises that to 2 GB, and results land in the phone's Downloads folder.
You are not imagining how roundabout this feels
Search for this and almost every result wants the upload before it will do anything. A 45-minute lecture recorded on a phone is commonly 300 MB to 900 MB, which on a typical home connection means minutes of waiting before any conversion starts, then the same wait coming back — for an operation the phone already performs every time it plays the video. The same instinct applies to documents you would rather not hand to a stranger's server: Merge PDFs Without Uploading Them Anywhere.
What you can fix is the format. What no converter can fix is the recording: if the microphone was three rows from the lecturer, you get exactly the audio that was captured, room echo and air conditioning and all. And if the audio track is silent or was never recorded, there is nothing to extract.
What a container is, and what "extract" really means
MP4, MKV and MOV are containers, not formats of sound or picture. Each holds a video track, an audio track and an index — in an MP4 that index is the moov atom, a table of timestamps and byte offsets saying where every frame and sample lives. Extraction just means telling a tool to ignore the video track and write the audio one somewhere else. That same index is why a downloaded file sometimes plays on a television but refuses to skip: Downloaded Video Won't Seek on Your TV? The Fix.
Two operations get called by the same name:
A stream copy lifts the audio bytes out and drops them into a new container, usually an .m4a. Nothing is decoded or re-encoded, so what you get is exactly the track that was inside the video, and it finishes in seconds even for an hour of footage. Almost every phone video carries AAC audio, typically 128 to 256 kbps, so a stream copy gives you an AAC file.
A conversion decodes that AAC and encodes it again as MP3, WAV, FLAC or OGG. It costs one encoding generation — a small, real quality loss you are unlikely to hear on speech — and it is what most destinations still ask for, because MP3 plays on essentially anything.
Bitrate decides the size, not the length
Size for a compressed format is bitrate multiplied by duration, and nothing else: 96 kbps for 10 minutes is about 7 MB, 192 kbps about 14 MB. Speech stays clear at 64–96 kbps in mono; music wants 192–256 kbps in stereo. A one-hour lecture therefore lands around 30–45 MB instead of the 600 MB the video was.
| Output format | Best for | Size per 10 minutes | What you lose |
|---|---|---|---|
| MP3 | Anything that just has to play — car stereos, upload forms | ~7 MB at 96 kbps, ~14 MB at 192 kbps | One encoding generation; bigger than AAC at similar quality |
| AAC (.m4a) | Keeping the exact track already inside the video | 9–19 MB, whatever the source bitrate was | Nothing on a stream copy, but some forms and older head units refuse .m4a |
| WAV | Editing, and tools that insist on uncompressed input | ~100 MB (stereo, 44.1 kHz, 16-bit) | Nothing in the samples; portability, and it restores nothing discarded |
| FLAC | Archiving a concert recording you care about | 50–60 MB | Device support, especially in cars; same caveat on discarded detail |
| OGG | Small speech files you will play yourself | ~7 MB at 96 kbps | Compatibility — many televisions, stereos and web forms will not open it |
How to extract the audio from a video on Android
- Check the file size first. Long-press the video in Files or your gallery and read its size. Over 100 MB is above the free per-video ceiling, so shorten it with the gallery's own trim first, or use Premium, which takes files up to 2 GB.
- Decide what the destination wants. A form, an upload page or a car stereo means MP3; an editor may insist on WAV. If you simply want to keep the sound, the AAC track already inside the video is the most faithful result.
- Convert the video to MP3. Open Convertlix, choose Video Converter and pick the file: it takes MP4, MKV, MOV or AVI, writes MP4, MKV, MOV or MP3, and can trim. It runs on the phone, so a long recording keeps processing while you switch to another app.
- Run a second pass if the format has to be different. Audio Converter handles MP3, AAC, WAV, FLAC, OGG and M4A, plus trim, so an MP3 can become a WAV for a tool that demands one — though no conversion puts back detail the first encoder discarded.
- Play thirty seconds of the result. Open it from Downloads and listen to the start, the middle and the end before you delete the video. A silent or truncated source shows up immediately.
When this will not work
Videos over 100 MB are refused on the free tier, and the ceiling applies to the file you hand over, so trimming inside the same job does not rescue a 600 MB recording — shorten it elsewhere first, or move to Premium. DRM-protected downloads from a streaming service are encrypted; nothing here opens those. A video whose audio track is missing or silent yields a silent file. If a transcription service demands WAV, note that video to MP3 to WAV is two encodes and a far larger file, so feed it the MP3 if it accepts one. And a WhatsApp voice note is already audio, usually OGG or M4A, so it needs a format conversion rather than extraction.
FAQ
Does extracting audio lose quality?
Copying the audio track out of the container without re-encoding gives you the same audio that was in the video, bit for bit — that is what tools mean by a stream copy. Converting to MP3 is different: the AAC gets decoded and encoded again, which costs one generation. On speech at 96 kbps that difference is very hard to hear; on music at 128 kbps or below, a careful listener can find it.
MP3 or M4A — which should I pick?
Pick MP3 if the file has to be accepted by something you do not control: upload forms, older car stereos and course platforms take it without argument. Pick M4A (AAC) if the destination is your own phone or computer and you want the track that was already inside the video, since that path involves no re-encoding. AAC also holds up better at very low bitrates, which matters for long spoken recordings.
How big will the MP3 be?
Multiply the bitrate by the duration: 96 kbps for 10 minutes is roughly 7 MB, and 192 kbps for the same 10 minutes is roughly 14 MB. A one-hour lecture at speech-appropriate settings comes out around 30–45 MB, a fraction of the video it came from. Duration drives the size, not the video's resolution, since the picture is no longer in the file.
Can I do this without installing anything?
On a computer, yes, and the standard tools beat a phone comfortably for very large files. ffmpeg does it in one line — ffmpeg -i in.mp4 -vn -c:a copy out.m4a to copy the track out unchanged, or ffmpeg -i in.mp4 -vn -q:a 2 out.mp3 to re-encode — VLC has Media then Convert/Save, and Audacity is the right choice if you also need to cut or level the result. For a 2 GB file that desktop is the better machine.
Does Convertlix upload my video?
No. The video is opened through Android's file picker, decoded and written on the phone itself, and the result is saved to Downloads; nothing is transmitted, so there is no copy on a server to delete. The app's network calls carry identity, licence, quota, remote configuration and payment data only — never a file. Switch on airplane mode after signing in and the conversion still completes, which is the easiest way to check.
If the sound is all you needed, the Video Converter turns the recording into an MP3 on the phone and leaves the original video where it is.