Subject: Re: Robust method of fundamental frequency estimation From: Matt Flax <flatmax@xxxxxxxx> Date: Fri, 2 Feb 2007 15:10:17 +1100 List-Archive:<http://lists.mcgill.ca/scripts/wa.exe?LIST=AUDITORY>I guess this topic and thread is now starting to cross between : a] Mathematical techniques of estimating fundamental frequency b] Perception of fundamental frequency and salience c] Psychoacoustic abstraction of experimental evidence for fundamental frequency salience d] Physiological mechanisms of hearing the fundamental frequency and salience. Locked away in the physical inner ear is the truly amazing property for the perception of a fundamental frequency ... even when it is not present ! I draw attention again to the afferent/efferent non-linear cross-over : http://mffmpitch.sourceforge.net/ You will require www.octave.org to run it out of the box. This previous model needs to be updated with a delay and filter between the periphery and the CNS, but that is not too difficult ... reviewed in a lumped model here : http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=1417513 http://www.assta.org/sst/2004/proceedings/papers/sst2004-315.pdf Salience and fundamental frequency generation must all be a part of the same circuit and machinery. Neither a mathematical ACF approach, nor a different psychoacoustical model will have the simplicity of a correct physiological model. As my supervisor once said ... an accurate lumped model will describe at least 70% of the auditory features we perceive... in my opinion, I also agree ... and what a model it will be once we all create it ! Matt -- http://www.flatmaxstudios.com/ http://www.flatmax.org Public Projects : http://sourceforge.net/search/?type_of_search=soft&words=mffm http://www.psysound.org