A Quick Guide To Lossless Versus Lossy Music Files
- Lossless means that, digitally, regardless of the format the file is stored in, there is no loss of data.
- The raw digital music file format types in common use are AIFF and WAV.
- The most common compressed digital music file format types are FLAC and ALAC.
- FLAC and ALAC compress the file, much like Zip compresses a regular file, but more efficiently for audio information.
- FLAC has different levels of compression, trading off file size for the amount of processing power required to compress and decompress the file.
- On average, a lossless compressed file will be about 50-70% of its original size.
- When de-compressed, no data has been lost.
- The bit rate for a CD-quality AIFF or WAV file is 1411 kbps (kilo-bits-per-second).
- Lossy means that data is removed to make the file smaller and the sound is altered.
- The most common lossy compressed file types are MP3, AAC and OGG.
- Files are compressed by removing the data storing the least audible sounds first.
- This often means any sounds above 16kHz (depending on compression level) are removed, as in music, instruments only produce sounds up to about 12-14kHz for the most part, so a person with good hearing and good equipment might just be able to discern this, but only with effort.
- File sizes end up about 1/10th to 1/3rd the original size.
- Even after decompression, the data removed is still lost.
- The highest bit rate for lossy files is (usually) 320 kbps and the common bit rates are 128 kbps and 192 kbps. Compare this to the lossless figure above.
The LAME Encoder:
LAME is possibly the most famous MP3 encoding software, as over the years the author(s) have been constantly refining it, to the point that someone posted on Head-fi a 128k VBR (variable bit rate) file and the RAW file and some people couldn't tell the difference between them. LAME has a number of settings, including the option to have the encoding rate as a constant bit (encoding) rate -- "CBR" or a variable bit rate -- "VBR". The latter is more effective as it adjusts the rate according to the complexity of the music. It includes a number of presets, from V0 to V9, which are respectively the highest and lowest default options. With the V0 setting, most people would struggle to tell that the music was compressed, without experience or training as well as high quality equipment. With V9, the sound would be like listening to music over a regular analog telephone played by someone holding an old analogue radio to the mouthpiece.