The most important characteristics of audio files are:
1. The duration of the audio signal.
2. Bit depth: the number of bits with which the sampled signal is stored. The more bits there are used, the more accurate we can store – and reproduce – the original signal.
3. Sampling rate: the number of samples of the original signal that are made per second. This is usually expressed in Hz. According to the nyquist frequency, the original signal can be reproduced exactly when the sampling rate is twice the highest signal frequency.
4. Number of channels: a value indicating the number of unique signals in the audio object, for example 2 (stereo).
In order to save storage space and bandwidth, a number of “lossy” file formats have been conceived. These formats sacrifice some of the sound quality by removing frequencies from the audio track in a smart way so that fewer data need to be stored. This results in a smaller file size. For permanent archiving it is desirable to submit a “lossless” file, in a format with the best quality without data loss. For a usable file size, however, it can be much more user-friendly to offer a lossy data export. Best practice is to assess case by case whether it is desirable to submit lossy formats as well as the lossless original data.
Preferred formats
- BWF (.bwf)
- MXF (.mxf)
- Matroska (.mka)
- FLAC (.flac)
- OPUS
Non-preferred
- WAVE (.wav)
- MP3 (.mp3)
- AAC (.aac, .m4a)
- AIFF (.aif, .aiff)
- OGG (.ogg)