Narrowband and Wideband Speech Codecs

There are two main classes of speech codecs, narrowband and wideband. Narrowband speech codecs use 8 kHz sampling frequency when converting the analog signal to digital form. This allows for encoding sound frequencies up to 4 kHz. Furthermore, the input signal is often filtered with an IRS [97] or a ModIRS filter [101] to make the sound signal compatible with traditional land-line circuit switched telephone systems (PSTN). This reduces the bandwidth to 300-3400 Hz.

Wideband speech codecs use 16 kHz sampling frequency in the analog to digital conversion process. The signal is further filtered to an audio bandwidth of 100-7000 Hz [101]. The advantages with wideband speech codecs, compared with narrowband speech codecs, are improved sound quality, intelligibility, naturalness and presence [123].

0 0

Post a comment