Recognition experiments with the SpeechDat-Car Aurora Spanish database using 8 kHz- and 16 kHz-sampled signals
Tóm tắt
Like the other SpeechDat-Car databases, the Spanish one has been collected using a 16 kHz sampling frequency, and several microphone positions and environmental noises. We aim at clarifying whether there is any advantage in terms of recognition performance from processing the 16 kHz-sampled signals instead of the usual 8 kHz-sampled ones. Recognition tests have been carried out within the Aurora experimental framework, which includes signals from both a close-talking microphone and a distant microphone. Our preliminary results indicate that it is possible to get a performance improvement from the increased bandwidth in the noisy car environment.
Từ khóa
#Databases #Microphones #Frequency #Working environment noise #Testing #Sampling methods #Bandwidth #Speech recognition #Telecommunication standards #Standards developmentTài liệu tham khảo
2000, ETSI ES 201 108 V1.1.2 Distributed Speech Recognition; Front-end
pearce, 2000, Enabling New Speech Driven Services for Mobile Devices: An overview of the ETSI standards activities for DSR Front-ends, Applied Voice Input/Output Society Conf (AVIOS 2000)
0, Baseline Results for Subset of SDC Finnish Database for ETSI STQ WI008 Advanced Front-End Evaluation, STQ Aurora DSR Working Group Document AU/225/00
moreno, 2001, SpeechDat-Car Spanish Database. SpeechDat-Cat Project LE4-8334, UPC
nadeu, 1995, Proc EUROSPEECH, 1381
10.1109/TASSP.1980.1163420
macho, 2000, Spanish SDC-Aurora Database for ETSI STQ Aurora WI008 Advanced DSR Front-End, UPC
nadeu, 2001, Speech Communication - Noise Robust ASR, 34, 93
moreno, 2000, SPEECHDAT-CAR. A Large Speech Database for Automotive Environments, Proc II LREC