Digital Audio Data

Sampling

Sampling Rate

Nyquist Frenquency

Nyquist frequency = 1⁄2 × Sampling rate

Nyquist frequency = 1⁄2 × 44,100 Hz = 22,050 Hz = 22.05 kHz

Other sample rates

48 kHz - DVD format
96 kHz
192 kHz

Quantization and sample resolution

Bit depths

Digital Audio File Formats

Compressed
- AIFF - .aiff
- WAVE - .wav
- Apple’s core audio format - .caf

Once we convert to digital, how is it stored?

In uncompressed file formats, digital audio is stored as a series of 16- or 24-bit amplitude values or as 32-bit floating-point amplitude values—32-bit floating-point values are used for internal calculations within most audio recording software, but recording (ADC) and playback (DAC) are done with 16 or 24 bits. Chunks of data written at the beginning of the file indicate information such as the sampling rate, the resolution, the number of channels (two for stereo, six for 5.1 surround, etc.), and the length of the file. Some formats include other information as well, such as loop points for use by hardware or software samplers. Digital audio that is used by audio editing/processing/mixing software is stored in an uncompressed file format.

44.1 kHz, 16-bit, stereo

44,100 samples/second × 2 bytes/sample = 88,200 bytes/second ≈ 90 kB/second

90 kB/second × 60 seconds/minute = 5,400 kB/minute ≈ 5 MB/minute

5 MB/minute × 2 channels (left and right) = 10 MB/minute

Compressed

Lossless
- mp4
- Ogg
- Free Lossless Audio Codec - .flac
Lossy
- mp3
- mp4 with AAC

Lossless compression reduces the size of audio files in such a way that the original data can be perfectly recovered. An example of lossless compression from the non-audio world is the .zip file, in which all the data from the zipped files can be fully recovered.

Lossy compression, on the other hand, reduces audio file sizes by permanently and unrecoverably eliminating some of the audio information. Lossy compression typically achieves far greater reduction in audio file sizes than lossless compression.

Most lossy compression schemes utilize perceptual encoding to achieve file size reduction without a substantial loss in quality.

The term “perceptual encoding” encompasses a variety of techniques that use knowledge of the human auditory system to selectively remove elements that we wouldn’t hear anyway. As an example, one of the auditory phenomena utilized by perceptual encoders is frequency masking. {{/note %}}