Make audio smaller by lowering its bitrate. MP3, AAC and OGG are re-encoded at the level you pick; WAV and FLAC are lossless, so they have no bitrate to trade. Every file keeps its own format, and if the result is not smaller than your file you keep the original.
Supported input formats — pick one to select a file.
Make an audio file smaller by lowering its bitrate. MP3, AAC and OGG are re-encoded at the level you choose; WAV and FLAC are lossless, so they have no bitrate to trade and are not re-encoded into something worse. Every file keeps the format it arrived in, and if the result is not smaller than your file you keep your original rather than a bigger one.
There is a single setting: how much detail to keep, expressed as a bitrate level, pre-set at a sensible default and movable in either direction. It applies to the formats that are encoded with a bitrate. It cannot apply to the others, because a WAV or a FLAC has no bitrate to trade — and rather than let the control sit there looking live while changing nothing, the panel disables it and describes the reason when every file in the queue is lossless. An empty queue leaves it enabled, since files are still being chosen.
MP3, AAC and OGG are re-encoded at the level you pick, and that is the trade: fewer bits per second, less detail. WAV and FLAC are lossless containers, so a lower level cannot make them smaller — a lossless file is already carrying exactly the audio data it holds, and re-encoding it into a lossy format would change what it is rather than compress it. That is also why a lossless file can come out slightly larger after being written again, and why the page keeps the original when that happens.
When a transcode does not beat the original, the visitor gets their original file back, not a larger one, and the row says so. The download carries the original's name, because naming an untouched file something like '_compressed' would be a second untrue statement stacked on the first. It is the same promise the Image Compressor makes, enforced in the same place — at the moment the file is handed over rather than in the copy.
The transcode runs on the server, so this page cannot work without the audio being uploaded — the dropzone says so on the page. The converted output is held in ephemeral storage only long enough to be downloaded, and a scheduled cleanup job purges converted files when the retention window passes. The optional import-by-URL feature also goes through this site's server, which is the same route the file takes when you upload it directly.
Your audio is uploaded to this server and transcoded there — the dropzone on the page says so, and it is the honest position rather than a disclaimer: a bitrate change needs the encoder, and the encoder runs on the server. The converted result is kept in ephemeral storage only so that you can download it, and a scheduled cleanup job purges converted files when the retention window passes; that window is 24 hours by default on this deployment, and the paid plans advertise a longer download window. URLs you import are fetched through this server too, which is the same path a direct upload takes.
It depends entirely on what the file is and what level you pick, and the page will not quote you a figure. What decides it is the gap between the bitrate the file already carries and the level you set: an MP3 at a high bitrate brought down a long way is a large saving, while a file that was already encoded at a low bitrate has little left to give. A WAV or a FLAC is the extreme case — no bitrate to lower means no saving, and writing it out again can produce a slightly larger file, which is why you would get your original back in that case.
No — there is no target-size field, and the mechanism cannot reach one directly. You choose a bitrate level, and a file's size then follows from that level and its duration: the same bitrate on a ten-minute track and a two-minute track gives you two very different file sizes. A target-size mode would have to keep re-encoding until the number came down, which changes the audio unpredictably from one file to the next. If you need a file under a specific limit, lowering the level and watching the resulting size on the row is the way to get there.
For MP3, AAC and OGG, yes — they are re-encoded at the level you pick, and reducing the bits available is exactly what the level controls, so some detail is not carried over. That is the trade the page is offering, stated plainly rather than hidden. For WAV and FLAC the answer is no, because nothing is re-encoded at a lower quality: those are lossless containers holding the audio data as it is, which is also why lowering the level cannot make them smaller and why the control turns itself off when the queue consists only of them.
Because the transcode could not beat the original, and the page will not switch you to a worse file to make a run look successful. When that happens the row keeps the original, reports its size, and says so, and the download gives you your file back under its original name. In practice it is almost always a WAV or a FLAC — a lossless file has already had the inefficiency squeezed out of it in the sense that matters here, so re-writing it can add a little weight rather than remove any. No setting can change that outcome; a different format would, and that is a job for the Audio Converter rather than this page.
No, every file keeps the format it arrived in. The format picker was removed on purpose, for the same reason the image page removed its own: a separate Audio Converter already covers moving a file between MP3, WAV, FLAC and the rest, and two pages offering the same switch would compete for the same visitor and the same search. Use this page when the format is already right and the file merely needs to be smaller.
Yes. The transcode runs on the server, so the audio has to be uploaded for the page to work at all, and the page says so in the dropzone footnote rather than leaving you to wonder — the encoder is not something a browser tab can run on its own. The result is kept in ephemeral storage only long enough for you to download it, and a scheduled cleanup job removes converted files after the retention window. Importing a file by URL fetches it through this server as well, which is the same path a direct upload takes.
The dropzone lists the exact set it takes, and the page's format chips show the same list, so it is worth reading them rather than assuming. What is worth knowing is that the list is deliberately narrower than what audio players can open: WMA and M4A are refused at upload, and that was measured rather than guessed — with real audio data under each name, an admitted extension reaches the queue while a refused one comes back with a file-type error, which is how the two are told apart. A refused file is skipped with a message rather than failing the whole batch.
