[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
On MOS Labeled speech databases
I am working in the area of speech quality
estimation and the research depends, on a great deal,
on the availablity of subjectively evaluated speech
databases. In the past I have been trying to contact
various people for the purpose of getting some
databases and apart from getting some favors from some
people, all efforts were futile. People working in
academia and industry do not lend/share their
databases for certain obvious reasons. Nonetheless,
what may the reasons be, the unavailability of such
data creates a feeling of sheer frustration, as one
does not have the right paraphernalia to apply the
ideas. For instance, today I received a rejection for
a paper where a major reason was that the my
approach/model had not been verified on enough data.
The reviewer was kindfully also aware about the
hardships of collecting the databases. Given this, I
at times repent that either I should not have started
my research in this area or else should have started
it with a subjective assessment lab, for instance.
I shall be very grateful to anyone working in this
field and who may possess enough databases that may be
required to validate the authenticity of a proposed
model, AND who may be also be willing to colloborate
on a research project to kindly contact me in person.
I shall be open about discussing my own ideas, and
going through other related formalities. I shall
remain grateful in return.
Looking forward to hearing from someone.
Yahoo! oneSearch: Finally, mobile search
that gives answers, not web links.