A new high resolution audio filter bank has been built and used to analyze speech samples which show the effects of pitch variation on sustained phonemes. The results of preliminary tests indicate tha the historically accepted -formant- method of representing speech spectra may be of limited usefulness in handling speaker differences and pitch variations in automatic speech recognition.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.