WIDEBAND SPEECH CODING STANDARDS AND WIRELESS SERVICES
WIDEBAND SPEECH CODING STANDARDS AND WIRELESS SERVICES
复制标题
宽带语音编码标准和无线服务
DOI:
--
复制
发表时间:
2006
期刊:
影响因子:
--
通讯作者:
Kari Järvinen
中科院分区:
文献类型:
--
作者:
Simão Ferraz;Kari Järvinen
ost communication systems in use today, such as the public switched telephone network (PSTN), are based on narrowband speech in which the audio bandwidth is limited to about 200‐3400 Hz. This bandwidth limitation dates back to the early days of wireline telephony, around a century ago. At that time, it was difficult to build inexpensive highquality microphones. In addition, the higher frequencies were lost anyway as calls passed over long lengths of copper wire. Narrowband speech is still commonly used today. However, for certain scenarios we are well aware of its limited quality. For many fricatives and plosives (such as /s/, /f/, /p/, /b/, and /t/) most of the energy lies well above 3000 Hz. The resulting inability to distinguish between them limits the intelligibility of narrowband speech. We especially notice this problem when trying to capture unknown words or names over telephone. In addition, many speaker-dependent characteristics disappear when bandwidth is limited to narrowband, making it more difficult to recognize who is talking. Distinguishing a mother from her daughter is not always easy over the phone. In this feature topic the use of the term wideband speech follows the International Telecommunication Union (ITU) definition denoting speech where the bandwidth has been roughly doubled compared to narrowband telephony, with a resulting bandwidth in the range of approximately 50‐7000 Hz. For comparison purposes, the so-called super-wideband extends to 14 kHz bandwidth, and audio covers the full audible spectrum from 20 Hz to 20 kHz.