To accurately detect a consonant section even under an environment where a noise level is relatively high.
A consonant section detection device 110 includes: a framing unit 120 which segments an input signal by a preliminarily determined frame unit to generate a framed input signal; a spectrum generation unit 122 which converts the framed input signal from a time domain to a frequency domain to generate a spectrum pattern; an average deriving unit 126 which derives band-specific average energy being average energy in each of continuous and preliminarily determined band widths, in the spectrum pattern; and a consonant determination unit 128 which compares the derived band-specific average energies with each other and determines whether the framed input signal includes a consonant or not by detecting that band-specific average energy of a first frequency band is higher than that of a second frequency band being a frequency band lower than the first frequency band.
YAMABE TAKAO
| JP2010164971A | 2010-07-29 | |||
| JPH05346797A | 1993-12-27 | |||
| JP2010164971A | 2010-07-29 |
| WO2001033548A1 | 2001-05-10 | |||
| US20070288238A1 | 2007-12-13 | |||
| WO2001033548A1 | 2001-05-10 | |||
| US20070288238A1 | 2007-12-13 |
JPN6015011233; Ｃｌｉｆｆｏｒｄ Ｊ． Ｗｅｉｎｓｔｅｉｎ， ｅｔ ａｌ．: 'Ａ ｓｙｓｔｅｍ ｆｏｒ ａｃｏｕｓｔｉｃ-ｐｈｏｎｅｔｉｃ ａｎａｌｙｓｉｓ ｏｆ ｃｏｎｔｉｎｕｏｕｓ ｓｐｅｅｃｈ' ＩＥＥＥ Ｔｒａｎｓａｃｔｉｏｎｓ ｏｎ Ａｃｏｕｓｔｉｃｓ， Ｓｐｅｅｃｈ， ａｎｄ Ｓｉｇｎａｌ Ｐｒｏｃｅｓｓｉｎｇ Ｖｏｌ． ＡＳＳＰ-２３， Ｎｏ． １, 197502, ｐｐ．５４-６７, ＩＥＥＥ