Conference Logo

INTERSPEECH 2009

Title: Reconstructing Clean Speech from Noisy MFCC Vectors

Authors: Ben Milner, Jonathan Darch, Ibrahim Almajai

Multimedia files in this directory

SizeViewName
93644 view ETSI_10dB.wav
93644 view ETSI_20dB.wav
93644 view ETSI_5dB.wav
93644 view ETSI_Clean.wav
94702 view ORIG_10dB.wav
94702 view ORIG_20dB.wav
94702 view ORIG_5dB.wav
94702 view ORIG_Clean.wav
93644 view PRED_10dB.wav
93644 view PRED_20dB.wav
93644 view PRED_5dB.wav
93644 view PRED_Clean.wav
93644 view PREDA_10dB.wav
93644 view PREDA_20dB.wav
93644 view PREDA_5dB.wav
93644 view PREDA_Clean.wav
1203 view README.TXT

Contents of the README.TXT file


Title: Reconstructing Clean Speech from Noisy MFCC Vectors
Authors: Ben Milner, Jonathan Darch, Ibrahim Almajai

ETSI_5dB.wav     ETSI reconstructed speech at 5dB.
ETSI_10dB.wav    ETSI reconstructed speech at 10dB
ETSI_20dB.wav    ETSI reconstructed speech at 20dB
ETSI_Clean.wav   ETSI reconstructed speech in clean conditions
ORIG_5dB.wav     Original speech at 5dB
ORIG_10dB.wav    Original speech at 10dB
ORIG_20dB.wav    Original speech at 20dB
ORIG_Clean.wav   Original speech in clean conditions
PREDA_5dB.wav    Reconstructed speech solely from MFCC vectors at 5dB with model adaptation
PREDA_10dB.wav   Reconstructed speech solely from MFCC vectors at 10dB with model adaptation
PREDA_20dB.wav   Reconstructed speech solely from MFCC vectors at 20dB with model adaptation
PREDA_Clean.wav  Reconstructed speech solely from MFCC vectors in clean conditions with model adaptation
PRED_5dB.wav     Reconstructed speech solely from MFCC vectors at 5dB
PRED_10dB.wav    Reconstructed speech solely from MFCC vectors at 10dB
PRED_20dB.wav    Reconstructed speech solely from MFCC vectors at 20dB
PRED_Clean.wav   Reconstructed speech solely from MFCC vectors in clean conditions



Note: these files were provided as-is by the author of the mentioned paper, and the publisher accepts no responsibility for the contents of these files.