BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Generative Speech Separation based on Pitch Information - Dr Xiang
  Li\, Cambridge University Engineering Department
DTSTART:20211011T110000Z
DTEND:20211011T120000Z
UID:TALK163663@talks.cam.ac.uk
CONTACT:Dr Jie Pu
DESCRIPTION:*Abstract*: Monaural speech separation aims to separate concur
 rent speakers from a single-microphone mixture recording. Inspired by audi
 tory scene analysis mechanisms\, a generative speech separation framework 
 based on pitch information will be presented in this talk. The prominent a
 dvantage of this framework is that both the permutation problem and the un
 known speaker number problem existing in general models can be solved by u
 sing pitch contours to indicate the target speaker to be separated. In add
 ition\, the generative approach is applied instead of traditional time-fre
 quency mask based approach\, to improve the perceptual quality of separate
 d speech. Specifically\, the proposed framework can be divided into two ph
 ases: pitch extraction and speech separation. The former aims to accuratel
 y extract pitch contour candidates for each speaker from the mixture\, whe
 re a two-stage approach is presented. Any pitch contour can be selected as
  the condition at the second phase\, and a conditional generative adversar
 ial network (CGAN) is used to separate the speaker corresponding to the gi
 ven pitch condition. The proposed framework is evaluated in terms of pitch
  extraction as well as speech separation.\n\n*Bio*: Xiang Li is a Research
  Associate in the Speech Group of the Machine Intelligence Laboratory\, En
 gineering Department of Cambridge University\, worked with Prof. Mark Gale
 s. She recently received her PhD from Peking University\, supervised by Pr
 of. Xihong Wu. This talk is about her PhD thesis. Her research interests i
 nclude speech enhancement/separation\, perception and natural language pro
 cessing.
LOCATION:Zoom: https://us06web.zoom.us/j/87426783837?pwd=akx4ZWZMYVZML2ZoO
 WRlYzdRaTd6dz09
END:VEVENT
END:VCALENDAR