next up previous contents index
Next: Speech research Up: Speech standards Previous: Introduction

Signal theory recalls

When digitising signals, the following parameters have to be considered:

These settings are those of the A/D converter when the data are recorded and must be used when they are played. They vary along the uses of different communities. Usually, the recorded signal is stored in files. When there are N channels, samples may respectively pertain to different files or sometimes to a unique one: in such a case samples or frames of samples are interleaved. The bytes order for each sample in the file (little or big endian) is also a relevant factor in reading the files (for example HI-LO order may be reversed from DOS to some UNIX environment). Data compression whenever used, implies the inverse uncompressing scheme to be applied before using the data.

EAGLES SWLG SoftEdition, May 1997. Get the book...