Lossy and lossless audio, and what this MP3 keeps

OfflineMP3 writes a lossy MP3. The word lossy means the encoder is allowed to discard parts of the sound. A lossless file is a different kind of file, and this tool does not create one. This page explains the difference in the terms the converter actually uses.

Open the converter

Two different promises

A lossless audio file can be decoded back to the same pulse-code samples that were encoded, bit for bit, when the format is doing what its name says. WAV and FLAC are common examples of that kind of file. People also keep AIFF or ALAC for the same reason. The promise is about the samples, not about a pleasant listening experience. A lossless file of a noisy recording is still noisy.

A lossy file makes a smaller file by throwing information away. The encoder keeps what its model thinks matters more to a listener and spends fewer bits on the rest. MP3, AAC, and Opus are all lossy codecs. Decoding them does not bring back the discarded parts. Encoding the result again, into any of those codecs, can only discard more or, at best, spend bits describing an already damaged signal.

What this tool is willing to read

Inspection accepts AAC-LC inside MP4 or MOV, and Opus inside WebM. Both of those codecs are lossy. A WAV or FLAC file, even a beautiful one, is the wrong container and the wrong codec for this page, and it is refused before encoding. Putting a WAV inside a file named .mp4 does not help, because the extension is not the decision.

So the usual task here is not "make a lossless archive of my master". It is "make a 192 kbps MP3 from a video whose audio is already AAC-LC or Opus". The video may have been shot on a phone, exported from an editor, or saved from a camera. The soundtrack inside it has already been through a lossy encode. The MP3 is a second one.

What the MP3 keeps and what it drops

The encoder keeps a 48 kHz representation of the selected track, at 192 kbps constant bitrate, with one channel or two. A 5.1 layout is reduced first. The downmix uses fixed gains on the front left, front right, centre, back left, and back right, and it leaves out the low-frequency effects channel. That is a deliberate cut, not an accident of a low bitrate. The omitted channel is gone from the stereo MP3.

The page also drops tags, chapters, and cover art. A title stored on the video does not become an ID3 title. The file name is only a cleaned copy of the video's base name. If you need the artwork or the chapter marks, they have to live somewhere else. They are not in this MP3.

The page does not keep the original compressed frames. Someone who wants the AAC or Opus bitstream unchanged needs a tool that copies the packet, not this one. Copying the packet would preserve that codec's own decisions. Re-encoding replaces them with the MP3 encoder's decisions. Those are different files, even when a casual listen makes them seem close.

Why a second encode is not free

Each lossy encode is a one-way cut. Starting from AAC-LC and producing MP3 does not "upgrade" the audio, and it does not get you closer to a microphone feed you never had. If the video is the only copy, the MP3 is a convenience copy for players that want MP3. The video remains the better source if you later want a different encode.

This site will not tell you that 192 kbps is transparent, or that nobody can hear the difference, or that 320 kbps would have been harmless. Those are listening claims. The only claim the product makes about the output is the contract the verifier enforces: one MP3 stream, 48 kHz, the expected channel count, 192 kbps constant bitrate, fully decodable, with no inherited tags or chapters. A file that fails that contract is not saved.

Words that cause confusion

"Extract the audio" sounds like lifting a track out whole. On this site it means decode, then encode. "The original quality" is not a number the page can show you, because the original uncompressed samples are not in an AAC or Opus file to begin with. "CD quality" is not a label the tool uses. A compact disc is 44.1 kHz and 16-bit stereo PCM, which is a different sample rate and a different kind of file from this MP3.

If a page, including this one, says the conversion is lossy, that is the accurate description. The product does not describe the MP3 as a lossless file, and it does not describe the task as a lossless copy of the soundtrack. The FAQ answers the same question in a few lines.