When people say "convert video to MP3," they usually mean one of two different operations that produce very different quality results. Understanding the difference prevents unnecessary quality loss.
Operation 1: Stream copy (lossless extraction)
If the video contains an audio stream that is already in your desired format, the extraction tool can copy the audio bytes directly from the video container to a new audio file. No decoding or encoding happens. The audio data in the output file is bit-identical to the audio data in the video file.
Example: An MP4 file contains AAC audio at 256 kbps. You extract the audio as an M4A file. The M4A contains the exact same AAC data. Zero quality loss. The extraction takes seconds regardless of file length because no processing occurs.
This is the ideal operation whenever possible. The limitation is that the output format must match the source codec. You cannot stream-copy AAC audio into an MP3 file because MP3 and AAC are fundamentally different codecs with different data structures.
Operation 2: Transcode (lossy conversion)
If you want the audio in a different format (for example, the source has AAC audio but you want MP3 output), the tool must decode the AAC audio to raw PCM audio and then encode that raw audio into MP3. This decode-then-encode process is called transcoding.
Transcoding introduces quality loss because you are compressing already-compressed audio. The AAC encoder already discarded some audio information that it deemed inaudible (this is how lossy compression works). The MP3 encoder then analyzes the decoded audio and discards more information according to its own psychoacoustic model.
The severity of the quality loss depends on the bitrates involved:
Source 256 kbps AAC to 320 kbps MP3: The MP3 encoder has enough bits to preserve nearly all the information that survived the AAC compression. Quality difference is inaudible.
Source 256 kbps AAC to 128 kbps MP3: The MP3 encoder must discard significant additional information. Quality loss is audible on good headphones, particularly in high-frequency content (cymbals, sibilance, string overtones).
Source 128 kbps AAC to 64 kbps MP3: Significant quality degradation. Audible artifacts including warbling, metallic timbre, and loss of stereo separation.
The "double compression" problem
Transcoding from one lossy format to another is sometimes called double compression. Each compression pass removes information, and the information removed by the second pass is different from the first. The cumulative loss is greater than either single pass.
This is why audiophiles prefer to work with lossless source files (FLAC, WAV) and compress only once to the final delivery format. But for video audio extraction, you rarely have a choice: the video already contains lossy audio, so any format conversion involves at least one re-encode.
When to stream copy vs. transcode
Stream copy when: - You are fine with the original audio format (AAC, Opus, Vorbis) - You want maximum quality preservation - Speed matters (stream copy is near-instant)
Transcode when: - You need a specific format for compatibility (MP3 for older devices, WAV for audio editing) - The source format is not supported by your playback device or service - You want to reduce file size (transcoding to a lower bitrate)
If you plan to edit the extracted audio (noise reduction, volume adjustment, trimming), extract to WAV (uncompressed) or FLAC (lossless compressed). This gives you a perfect representation of the audio without compression artifacts. Edit the WAV/FLAC file, then compress to MP3 or AAC only as the final step. This minimizes the number of lossy compression passes. The MediaBrew extract audio tool lets you choose the output format, and for lossless extraction it stream-copies when the source codec matches the output.