Why your voice recording won't upload, and what converting it really does

Audio is the file type people think about least until it goes wrong. A phone records a lecture, an interview or a meeting without any fuss, and then the first time you try to send it somewhere the problems begin. The upload form rejects the format. The converter you find produces a file that is much larger than the original. Or the file refuses to play after a rename that seemed harmless.

This post looks at what people actually run into, explains the three formats you will meet and why they behave so differently, and then shows measured results from converting a real two minute recording. By the end you should know which format to choose, which bitrate to pick, and when not to convert at all.

What people actually run into

Six threads over twelve years and the same questions come back: which format will the site accept, how do I get there from what my phone made, and why is the new file a different size. The answers are scattered, and a few of them contradict each other. The way through is to understand the three formats first.

Three formats, three different jobs

• WAV. Raw, uncompressed sound. Nothing is thrown away, which makes it ideal for editing and enormous for everything else. A minute of ordinary stereo takes around ten megabytes.
• MP3. The universal lossy format. It throws away detail the ear is least likely to miss to make the file small, and nearly every site, player and device accepts it. This is why forms ask for it.
• M4A. A container that usually holds AAC, another lossy format that is more efficient than MP3, so it sounds as good at a smaller size. It is what iPhones produce. Some sites do not accept it.

Two of the three are lossy, which matters more than it sounds. When you convert one lossy file to another lossy format, the second step has to compress already compressed sound. It cannot recover what the first step discarded, and it throws away a little more. That is the point the Audacity thread's replies keep making.

The arithmetic that explains the sizes

An audio file is a rate multiplied by a length, just like video. The rate is how many thousand bits per second are used, called the bitrate, and everything about the size follows from it.

That middle line is the surprise. If your voice memo uses 64 kbps and you convert it to an MP3 at 128 kbps, the file doubles in size, and it cannot sound any better, because all the detail the converter has to work with was already thrown away when the memo was recorded. A higher bitrate only spends more bytes to store the same sound.

What we measured

To see this with real numbers, we made a 114 second speech recording and saved it two ways: as an M4A at 64 kbps mono, which is the kind of file a phone voice memo produces, and as a CD quality stereo WAV, which is the kind of file a recorder or editor exports. We then converted them with our Audio converter at different settings and measured every result.

The two groups tell opposite stories. Starting from an efficient M4A, every MP3 was bigger, and the higher the setting the more it grew. Starting from a WAV, the MP3 was eleven times smaller. The converter did the same job in both cases. The difference is where the file started. Going from a big uncompressed file to a lossy one saves space, and going from one lossy file to another usually does not.

Notice also that the MP3 at 128 kbps came out at the same size from both starting points, even though one was mono and the other stereo. An MP3's size depends on its bitrate and its length and on nothing else.

Which setting to pick

1. 1Check what the place actually asks forSome forms want MP3, some accept M4A, some say WAV. Read the accepted file types before converting anything, and if M4A is accepted, send it as it is.
2. 2For speech, 96 or 128 kbps is plentyA voice carries far less information than music. Our tool's own guidance calls 128 kbps good for speech, and the measurements above show that 96 gives a smaller file still. Use 192 and above for music, where the extra detail is worth the extra size.
3. 3Never go higher than the originalIf the source is a 64 kbps memo, a higher setting only makes a bigger file. Choose a rate close to what the recording already used, and no higher than you need.
4. 4Convert once, from the originalEach lossy step loses a little more. Keep the first recording, convert a copy, and if you need a different version later, go back to the first one.

What to do on a phone and a Mac

The threads show the platform side of this, and it is worth knowing before you search for a converter. According to the 2022 thread, an iPhone's Voice Memos app does not let you change the format it records in. The 2025 thread reports that dragging a recording out of the Mac app can produce a .qta file, while copying it and pasting it into Finder gave a normal .m4a. And the 2024 thread shows the long way round, sending the memo to a Mac, importing it into the Music app and converting there. All of that works, and all of it takes steps that a single tool in the browser can do in one.

How long it takes

Converting in a browser is not instant, so it helps to know what to expect. On the test computer, the 114 second M4A took about 14 seconds to turn into an MP3, and the larger 19 MB WAV took about 26 seconds, most of it spent reading and decoding the big file. A one hour interview would take several minutes, roughly seven or eight by the same ratio. The tool refuses files above 400 MB, and a phone has far less memory than a computer to work in, so very long recordings are better handled on a laptop.

Making it smaller, other than converting

• Trim the silence and the false starts. Size is bitrate multiplied by length, so every minute removed is a saving with no quality cost at all.
• Pick a lower bitrate for speech. The step from 192 to 96 kbps halves the file, and for speech the difference is usually small.
• Record shorter pieces. Several short clips are easier to upload, and one failed upload loses less.
• Keep a lossless copy only if you need one. FLAC, mentioned in the 2024 thread, keeps everything and is smaller than WAV, but check that the place you send it to accepts it.

A note on privacy

A voice recording is one of the most personal files you own. It can hold a conversation, an address, a medical detail or an interview with someone who trusted you to keep it private. Many free online converters work by uploading your file to their server, which is a poor trade for that kind of recording. A converter that runs in your browser decodes and encodes on your own device, and you can check that by watching your network activity while it works.

Mistakes that cost people time

• Renaming the extension. It does not convert anything and may leave a file that will not play.
• Converting to a higher bitrate than the source. The file grows and the sound does not improve.
• Converting an MP3 again. Each lossy pass loses a little more, and the damage adds up.
• Converting to WAV for safety. It is eleven times bigger than the memo it came from, and very often unnecessary.
• Not checking what the site accepts. If it takes M4A you can skip the whole job.
• Using an online converter for a private recording. Many upload the file.

A short routine to keep

Find out the accepted formats. If your format is accepted, send it as it is. If not, convert one copy of the original, to MP3 at 128 kbps for speech or 192 and above for music, trim it first if it is long, and listen to the result before you upload it. Keep the original recording. That is all there is to it.