Submitted in partial fulfilment of the requirements for the degree of M.A., University of London.
[This is an HTML version of the thesis just as it was written and submitted in 1962. Beyond converting it to HTML and adding some internal hyperlinks, I have made no attempt to update it. In retrospect I feel quite proud of the statistical sophistication I showed in making sure that my speaker sample was homogeneous: in these days of instant computerized ANOVAs, bear in mind that in 1962 working out standard deviations involved hours of repetitive labour with a whirring mechanical calculating machine. Note that SI units had not yet been generally adopted: cps (cycles per second) would nowadays be called Hz, and kc/s kHz. Daniel Jones's spelling tamber has not after all succeeded in displacing the spelling timbre. --JCW, February 2001]
- Final note: inverse filtering