Description: This folder or any similar folder such as this created to hold your speech files, should be copied into the root folder C:\ , created on the local hard drive. Then the path will read thus C:\SpeechData\Tope... etc. An alternative is to modify the file setTemplates.m to change the directory as appropriate.
Try to use the same microphone for training as well as for recognition to obtain improved accuracy. You might also need to tweak your sound card s mic input level.
The recognition time is dependent on systems s processing power and has been observed at between 1 and 5 seconds on various x86 platforms. To speed up recognition, you can implement unconstrained DTW (this is the code portion taking most time).
To Search:
- [dtw] - dtw file DTW algorithm is to use speech
File list (Check if you may need any files):
SSBerouti79.m
SSBoll79.m
sound10.wav
SSMultibandKamath02.m
Download