How long does it take to convert a whole M4A library to MP3, and are the originals safe?
Short answer: plan on the MP3 encoding, not the file count. In our test the M4A to MP3 converter took 100 tagged M4A songs of 3 minutes each in one batch and converted all of them, one after another:
- Time: 1,862 s (31 min 2 s) for 100 songs at 192 kbps, 18.43 s per song on average (17.46 to 20.75 s). 20 songs took 360.5 s at 192 kbps, 420.6 s at 320 kbps and 510.7 s at 128 kbps; as WAV, 30.1 s.
- Originals: all 100 source files were byte-for-byte the same afterwards (same SHA-256), and nothing was added to or removed from their folder. The MP3s arrive as new downloads with the same name, for example
track_001.m4agivestrack_001.mp3. - Tags: title, artist, album, album artist, year, track number, genre and the cover image matched the source on every one of the 180 files we read back.
To convert a library: open the M4A to MP3 converter, pick MP3 and a bitrate (192 kbps is a sensible default for music), choose all the files at once, press Convert N files, then press Download on each row. Keep the tab open until the last row says done.
Why it takes as long as it does
A browser cannot rewrap AAC audio as MP3. The converter decodes each M4A to plain samples, then encodes those samples again as MP3 with the LAME-based encoder this site uses, which runs in JavaScript. Decoding is quick; encoding is the slow part. That is why saving the same 20 songs as WAV, which needs no encoder, took 30.1 s instead of 360.5 s.
Files are converted one at a time, in the order they appear in the list, so the total is roughly the time for one song multiplied by the number of songs.
How we tested
- Tool: this site's M4A to MP3 converter, run in Chromium 153 through Playwright 1.63 on a local copy of the page, on a 4-core server (Intel Xeon E5-2699 v4) with 9,941 MB of memory, of which 4,178 to 5,264 MB was free before each job (1 MB = 1,048,576 bytes). Each job used a fresh browser, picked all its files at once, waited until every row had read its length, pressed Convert, and then pressed every Download button.
- Library: 100 M4A files we made, each 3:00 of Mozart's Symphony No. 40, first movement, played by the Musopen Symphony, in the public domain on Wikimedia Commons. Every file starts at a different point of the recording. AAC at 256 kbps, 44.1 kHz stereo, made with FFmpeg 4.4.2's built-in AAC encoder, with a full tag (title, artist, album, album artist, year, track number, genre) and a JPEG cover. 100 files, 553.9 MB.
- Checks: SHA-256 of every source file before and after; the file list of the source folder; time from pressing Convert to each row turning done; memory of all Chromium processes, sampled every 0.2 s; every downloaded file read back with Mutagen and FFprobe and compared with its source.
Results
| Batch | Total time | Per 3-min song | Output size | Peak memory (all Chromium) | Tags and cover kept |
|---|---|---|---|---|---|
| 20 songs, MP3 128 kbps | 510.7 s | 24.56 s (23.51 to 25.97) | 55.0 MB | 951 MB | 20 of 20 |
| 20 songs, MP3 192 kbps | 360.5 s | 17.97 s (17.42 to 19.11) | 82.4 MB | 942 MB | 20 of 20 |
| 20 songs, MP3 320 kbps | 420.6 s | 19.76 s (18.95 to 21.52) | 137.4 MB | 1,017 MB | 20 of 20 |
| 20 songs, WAV 16-bit | 30.1 s | 1.15 s (1.03 to 1.28) | 605.6 MB | 1,505 MB | 20 of 20 |
| 100 songs, MP3 192 kbps | 1,862.0 s | 18.43 s (17.46 to 20.75) | 412.2 MB | 1,471 MB | 100 of 100 |
Every MP3 came out 3:00.036 long against 3:00.000 for the M4A: MP3 encoding pads the end to a whole frame. Every WAV was exactly 3:00.000. All 20 or 100 outputs in a batch were the same size to within 2 bytes, because the songs had the same length and the encoder runs at a constant bitrate.
The time per song did not follow the bitrate: 128 kbps was the slowest of the three MP3 settings. To check that this was the encoder and not the batch, we ran the same encoder outside the browser on one of the songs, twice per bitrate: 23.95 and 22.94 s at 128 kbps, 17.73 and 17.19 s at 192 kbps, 18.96 and 18.97 s at 320 kbps. So 128 kbps does not save time with this encoder.
Going from 20 to 100 songs did not slow each song down: 17.97 s on average for 20 at 192 kbps, 18.43 s for 100, or 6.14 s per minute of music. Memory grew from 942 to 1,471 MB at the peak, because every finished MP3 stays in the tab until you close it (412.2 MB for 100 songs). The 100 files were all listed and ready 3.2 s after we picked them, and nothing failed.
Are the original M4A files safe?
Yes. A web page can read a file you pick, but it cannot write to it, rename it or delete it. Our check agrees: for every job, all source files had the same SHA-256 before and after, and the source folder still held exactly 100 files. The MP3s are new files that your browser saves where it saves downloads, named after the source with the extension changed.
That also means converting does not free up any space: you keep both copies until you delete the M4A files yourself. Check a few MP3s first.
Does converting M4A to MP3 lose quality, and is M4A better than MP3?
Converting can only lose: the MP3 is made from the M4A's decoded sound, not from the original recording. To put numbers on it we used the 30 s Mozart excerpt as a 44.1 kHz WAV and measured how far each file's waveform was from that WAV, as signal-to-error in dB (higher is closer; it measures difference, not what you can hear). AAC was made with FFmpeg 4.4.2's built-in AAC encoder, MP3 with this site's encoder.
| Bitrate | AAC (M4A) from the WAV | MP3 from the WAV | MP3 made on the converter from a 256 kbps M4A |
|---|---|---|---|
| 128 kbps | 28.2 dB, up to 16,538 Hz | 15.8 dB, up to 15,999 Hz | 15.7 dB |
| 160 kbps | not made | 17.8 dB | 17.8 dB |
| 192 kbps | 33.6 dB, up to 18,163 Hz | 28.3 dB, up to 15,999 Hz | 27.5 dB |
| 256 kbps | 36.3 dB, up to 19,251 Hz | 37.3 dB, up to 18,126 Hz | 33.9 dB |
| 320 kbps | 38.6 dB, up to 19,251 Hz | 41.7 dB, up to 18,136 Hz | 35.3 dB |
Three things follow. At 128 and 192 kbps the AAC file was closer to the source than the MP3 at the same bitrate, and kept more of the high end; at 256 and 320 kbps the MP3 scored higher on this measure. Converting a 256 kbps M4A to MP3 at 128 to 192 kbps cost little beyond what that MP3 bitrate costs anyway (15.7 against 15.8 dB, 27.5 against 28.3 dB). And going above the M4A's own quality buys nothing: from a 256 kbps M4A, which scored 36.3 dB, the 320 kbps MP3 reached 35.3 dB, below the M4A it came from. Keep the M4A if your player handles it; convert when something needs MP3.
Tips for a large library
- Pick the format and bitrate before you press Convert, and leave them alone while it runs: each file uses the setting in force when its turn starts. 192 kbps gave 4.1 MB per 3-minute song in our test, 128 kbps 2.7 MB and 320 kbps 6.9 MB.
- Select the files in one go. In the file picker, open the folder and press Ctrl+A (Cmd+A on a Mac). All 100 of our files had read their length within 3.2 s.
- Leave the tab open. Closing or reloading it stops the batch and loses the finished files that have not been downloaded.
- Download each row. There is one Download button per file; there is no zip of the whole batch.
- Very long files are a different limit: each file is decoded into memory on its own. The converter's FAQ gives the figure from our earlier test (a 45-minute M4A worked, a 60-minute one crashed the tab with about 3 GB free).
Questions
Is MA4 or MP4A the same as M4A?
They are usually typing slips for M4A. M4A is the file extension; mp4a is the code name the MP4 format uses for AAC audio inside the file. The same converter handles them, as long as the audio is AAC.
Does it work with Apple Lossless M4A files?
Not in the Chromium browser we tested: the converter's FAQ says Apple Lossless (ALAC) M4A files did not decode there. Our library was AAC only.
Are the files uploaded?
No. Decoding and encoding happen in your browser; see how to check whether an online tool uploads your file.