Skip to content

This is the home directory to speaker diarization module being developed for Hetergeneous News data in RedHen Labs as a GSOC Project

Notifications You must be signed in to change notification settings

ksingla025/Speaker_Dia_RedHen

Repository files navigation

Speaker Diarization Module : Heterogeneous news data

Detailed Documentation : http://web.iiit.ac.in/~karan.singla/redhenproposal.pdf

####Using ALIZE Now user can also use ALIZE speaker identification system instead of LIUM just use ALIZE_spk_seg/spk_det.sh

####Using PyCASP User can also use PyCASP based diarization system : Check "pycasp" folder for instructions.

####Using LIUM

Below Modules are primarily for running LIUM diarization system :

####### Module 1 This will extract 16kz audio date from Mp4 files and divide them according to respective news network Data-PreProcessing

Requirements : 1. Python (ofcourse, wrapper is in python) 2. FFMPEG (https://www.ffmpeg.org/download.html)

python 01_pre-processing.py |path-to-mp4-data| |output_name|

#######Module 2 This module takes input the data folder created in previous step and do produces diarization output for each audio with data organized similarly to audio folder with name _seg

python 02_single_show_diarization.py |path-to-audio-data|

#######Module 3 Cross-show diarization : Still under development

Note : "Scripts" folder contains supporting/utility scripts for the pipeline. Change the poitner to "lium.jar" in scripts/diarization.sh, if needed

About

This is the home directory to speaker diarization module being developed for Hetergeneous News data in RedHen Labs as a GSOC Project

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published