Audio processing involves the study of mathematical and signal processing techniques to understand or alter the nature of audio signals. The different kind of audio signals under study include speech, music, environmental audio and computer audio. Audio is analyzed in the temporal or spectral domain by applying various filters.
Key concept is to transform the audio into PCM format so you have access to the raw audio curve. Each channel will have its own curve.
Digital audio is represented by a series of points on this curve. Each point is called an audio sample. Numerical value of each sample can be represented in either integer or floating point.
Be aware to map each audio sample numerical value to memory typically requires several bytes of storage. One byte can store only 2^8 distinct values (256) which will result in noticeable distortion. High quality audio is typically stored using at least two bytes of storage per audio sample. When we use two bytes this gives us 2^16 possible values of the raw audio curve height as the audio wobbles up and down. The more bytes we use for storage the higher fidelity we gain as this reduces the gap between each distinct curve height measurement. This called bit depth. CD quality audio uses two bytes per audio sample per channel. The other fundamental aspect of digital audio is Sample Rate with determines the number of samples per second of time.
