Hi Christian, On 11/05/2012 01:15 PM, Hoene wrote: > we at University of Tübingen/Symonics. I am expecting about 100 participants > in the tests. Give me some more time to write down the testing plan. I will > be happy to read your comments then.
I'll be happy to read your test plan. Note that if you really have 100 willing participants and are going to use MUSHRA, you can actually make many tests because MUSHRA typically only needs 10-20 listeners to produce statistically significant results. This would make it possible to actually help improve the existing implementation, as opposed to only adding one more data point. > As the tests are not intended for the codec developers but for people > outside of the IETF, I will focus on well standardized testing procedures > like MUSHRA and the guidelines given in > "practical procedures of subjective testing" > http://www.itu.int/opb/publications.aspx?media=electronic&parent=T-HDB-QOS.0 > 2-2011 About MUSHRA, I suggest you review the exact methodology because your previous test had issues with that. For example, your slides ( http://wiki.tools.ietf.org/wg/codec/trac/raw-attachment/wiki/TestingResults/summary.pdf ) indicated "Participants were not informed about the presence of hidden references" which goes against the MUSHRA methodology. Cheers, Jean-Marc _______________________________________________ codec mailing list [email protected] https://www.ietf.org/mailman/listinfo/codec
