[vorbis] Participate in listening tests

[ff123]

Greg Wooledge wrote:

>ff123 (miyaguch at eskimo.com) wrote:
>
> > FRIEDMAN version 1.21 (Oct 16, 2001) http://ff123.net/
> > Blocked ANOVA analysis
>[...]
>
>Well.  Let me break this thundering silence and reveal my ignorance.
>Are there any online resources which explain the basic math behind
>this methodology?  (I tried, but sites I stumbled across all assumed
>a familiarity with the methodology, and merely showed how to apply it to
>special cases.)
>
> > mpc is better than wma8, xing
> > ogg is better than wma8, xing
>
>Shocking! ;-)

I don't know if it's the best site for explaining ANOVA, but here's one:

http://faculty.vassar.edu/lowry/webtext.html

Scroll down to chapter 13.

Without any math:  you can think of ANOVA as an extension of the 
t-test.  The t-test is used to compare two means, and ANOVA is used to 
compare multiple means.  Specifically, an ANOVA analysis attempts to answer 
the question of whether any two means are significantly different from each 
other.  That is a very limited question, though, and the subsequent 
analysis (Fisher LSD) attempts to separate the means if the ANOVA proves to 
be significant.

The actual method I will use in the final analysis will probably not be 
ANOVA, however.  I am using it now because it is convenient and quick, but 
it assumes things which may not be true, such as normal distribution of 
listener population.  I will probably switch to a resampling-based method 
of analyzing the data once more participants have submitted results.  But 
that statistical method also uses much more computing power (it takes hours 
on my Celeron 800).

ff123

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.