[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

detection/identification



I'm not a psychologist, but I would like to offer some possible
ideas - the psychologists can comment on these types of
experiments.

It seems to me that the issue is not only whether one can
notice the difference between a real and synthesized piano, but
also the question of what characteristics make a synthesized piano
sound realistic.

One possibility is to have as stimuli both real piano notes and
synthesized notes, using different synthesis techniques and
different parameter settings.  Then an ANOVA test could be
constructed in which people make preference ratings and one
could try to identify the factors and crossed factors which
listeners prefer.

Another possibility is a multidimensional scaling experiment
based on difference judgments (how different are these?) for
pairs of stimuli.

There are several algorithms in existence for synthesizing piano
notes and I don't know whether anyone has made any comparisons
in subjective tests.

Linda Seltzer
lseltzer@phoenix.princeton.edu

McGill is running a new version of LISTSERV (1.8d on Windows NT). 
Information is available on the WEB at http://www.mcgill.ca/cc/listserv