How do you convert an iPhone voice memo to MP3, and which bitrate should you pick?

What a voice memo file is

Apple's iPhone User Guide for iOS 27 says that recordings are exported from Voice Memos in .m4a format by default, "so layered recordings are flattened and Spatial Audio becomes stereo" (Export a Voice Memos recording to Files on iPhone, read 2026-10-22). The same page says you can instead tap Options and choose Editable to export "a file where the effects and layers can be adjusted later (QuickTime Audio format)". The sharing page describes the same two choices: Rendered, "a single .m4a file with all audio effects and layers mixed in", and Editable (Share a recording in Voice Memos on iPhone, read 2026-10-22). For conversion you want the Rendered .m4a; we had no Editable file and don't know whether a browser can decode one.

An .m4a is an MP4 container, and the sound inside is usually AAC. Apple's Voice Memos app also has an Audio Quality setting with two choices: "Compressed (for lower-quality, smaller files) or Lossless (for high-quality, larger files)" (Change Voice Memos settings on Mac, Voice Memos User Guide, read 2026-10-22). The iPhone guide pages we read do not describe this setting, and none of Apple's pages we found name the codec, bitrate or sample rate of either choice. So we don't know what bitrate a Compressed memo is recorded at, and we tested a range of them instead. Lossless on an Apple device means ALAC (Apple Lossless), which is the case the converter cannot open.

How we tested

• Speech: a paragraph we wrote (meeting notes, a shopping list, a train time), read by the open-source eSpeak NG 1.50 synthesizer at 22.05 kHz, cut to 60 s and resampled to 44.1 kHz and 48 kHz mono 16-bit WAV. Synthetic speech, so the numbers describe the files, not an iPhone microphone.
• Stand-in memos: that speech encoded with FFmpeg 4.4.2's built-in AAC encoder, mono, at 32, 48, 64, 96 and 128 kbps nominal (the files came out at 38.7 to 135.4 kbps), at 44.1 kHz, plus 64 kbps at 48 kHz, a stereo 128 kbps copy, a copy 18 dB quieter, and two ALAC files (44.1 and 48 kHz). Long files: the speech looped to 30, 60, 90 and 120 minutes at 64 kbps.
• Browser: Chromium 153 driven by Playwright 1.63, opening a local copy of this site's M4A to MP3 converter on a 4-core server (Intel Xeon E5-2699 v4, 9,941 MB of memory, 3,596 to 3,765 MB free before each short job). One fresh tab per file; the page's own Download button saved each result.
• Score: signal-to-error ratio in dB against the 44.1 kHz speech WAV after lining the files up: higher is closer. It measures how far the waveform moved, not how it sounds. A second score compares each MP3 with the decoded stand-in it was made from, which isolates what the converter itself changed.

Which bitrate to pick

Same 64 kbps mono stand-in, every setting the converter offers. The stand-in itself scored 30.4 dB against the original speech, so no MP3 can beat that.

At 128 and 160 kbps the converter's encoder changed the waveform about as much as the AAC recording had (25.8 dB against the decoded memo). At 256 kbps that drops to 61.4 dB, far below anything the memo itself had lost, and the score against the original speech is back to the memo's own 30.4 dB. 192 kbps is the middle: 3.4 dB under the memo at 1.37 MB a minute. The encoding time did not depend on the bitrate; a WAV needs no encoder and took half a second.

The MP3 came out 60.029 s long instead of 60.000 s, and the speech starts 1,105 samples (25.1 ms) later than in the memo, because the MP3 encoder pads the start and end to whole frames. A WAV keeps the exact length and timing. The converter's encoder is mono-aware: a mono memo gives a mono MP3, but a mono MP3 at 128 kbps is the same 960,470 bytes as the stereo one, so mono does not make the file smaller.

Low-bitrate memos: the MP3 can't be better than the recording

Every stand-in converted at the smallest MP3 setting, 128 kbps. The second column is the recording before conversion.

A 32 kbps recording converted at 128 kbps is three times the size and no better than it was: 16.3 dB against 16.9 dB. The MP3 setting puts a ceiling of about 25 dB on this speech at 128 kbps, whatever went in; a recording below that ceiling keeps its own score, one above it is pulled down to the ceiling. If you need the file small, keep the .m4a; if a player insists on MP3, the extra bytes buy nothing you can measure.

Lossless memos, 48 kHz, quiet recordings and trimming

• ALAC (Lossless): both files were listed with "length unknown" instead of a duration, and pressing Convert gave "Your browser couldn't decode this file." after 0.26 and 0.27 s, at 44.1 and at 48 kHz. Chromium has no ALAC decoder. If you have Lossless memos and a Mac or PC, FFmpeg converts them: ffmpeg -i memo.m4a -c:a aac -b:a 128k memo-aac.m4a gives a file the converter opens. We don't know a way to do it on the phone without installing anything.
• 48 kHz memos: the 48 kHz stand-in converted to a 44.1 kHz MP3 (the default) at 24.1 dB and to a 48 kHz MP3 at 22.6 dB; both were 128 kbps and 960,384 or 960,470 bytes. Leave the rate at 44.1 kHz. Side note: with a 48 kHz input and 44.1 kHz output the finished row read "0:60.0" instead of "1:00.0"; the file was 60.029 s.
• Volume slider: the 18 dB quieter stand-in came out 3.53 dB louder at 150% and 6.03 dB louder at 200% (the maths says 3.52 and 6.02), with no clipped samples because its peak was 18.5 dB below full scale. The normal stand-in at 200% clipped 7,444 of 2,646,000 samples and hit 0 dBFS. The page warns "Above 100%, loud parts can clip." as soon as you pass 100; believe it unless the memo is clearly quiet.
• Trimming: each file row has start and end fields. 10 and 25 on the 60 s memo gave a 15.000 s WAV of exactly 661,500 samples. For a cut with a waveform to look at, use the audio cutter instead; it opens .m4a too.
• Memory: every 60 s job peaked at 620 to 653 MB of Chromium memory against 434 to 445 MB with the page open and no file.

Long memos: an hour-long lecture

Mono files take half the memory of stereo, so a memo can be longer than a song before the browser gives up. We looped the speech into 30, 60, 90 and 120 minute mono AAC files at 64 kbps.

Up to 90 minutes converted, and the time grew in step with the length: about 2.8 s per minute of speech on this machine, so an hour of lecture is a three-minute wait, and the tab needs about 2.4 GB of memory for it. The 120-minute file was refused with the same message as the ALAC files, after 15.7 s and without the tab crashing; the kernel log shows no process killed, so this was not the machine running out of memory. It matches what we found for the cutter, which rejects mono 44.1 kHz files between 101 and 102 minutes long: see how long a file an online MP3 cutter can open. For a memo longer than that, split it first on a computer (ffmpeg -i memo.m4a -c copy -f segment -segment_time 3600 part%d.m4a) and convert the parts. The page shows "long file: converting can take several minutes, and very long files can run out of memory" next to any file over 30 minutes.

Step by step

1. In Voice Memos, tap the recording, tap the More button, then Share. Under Options keep Rendered. Choose Save to Files and pick a folder (Apple's export page, linked above).
2. Open the M4A to MP3 converter in a browser, choose the .m4a from Files (or from wherever you copied it on a computer).
3. Set MP3, 192 or 256 kbps, 44.1 kHz. Leave Volume at 100% unless the recording is quiet. Type start and end times if you only need part of it.
4. Press Convert, wait for the row to say done, press Download. The memo itself is not changed; the MP3 is a new file.

If you have several memos, pick them all at once; they convert one after another. To join them into one file first, see how to merge voice memos and MP3 files.

Questions

Does converting a voice memo to MP3 lose quality?

Some, always, because the AAC is decoded and encoded again. How much depends on the bitrate you pick: in our test 128 kbps lost 6.2 dB from a 64 kbps memo, 192 kbps lost 3.4 dB, and 256 kbps lost nothing measurable (0.0 dB). It cannot gain anything: no MP3 scored above the memo it came from.

Why is my MP3 bigger than the voice memo?

Because the smallest MP3 setting is 128 kbps and MP3 has no mono discount, while a Compressed memo is a low-bitrate mono AAC. Our 64 kbps stand-in grew 1.65 times at 128 kbps and 3.30 times at 256 kbps.

Can I convert on the iPhone itself?

The converter is a web page, so it should run in Safari once the .m4a is in Files, but we tested in Chromium on a server, not on an iPhone or in Safari, so we can't say how long a memo an iPhone's memory will take.