[linux-audio-user] Re: Digital Fidelity

Paul Davis paul at linuxaudiosystems.com
Tue Feb 28 08:20:39 EST 2006


On Tue, 2006-02-28 at 12:38 +0100, Carlo Capocasa wrote:
> Double blind tests can be very useful for a great number of things.
> However, they have a bias: They neglect 'placebo'. However, we are

this is completely wrong. the purpose of a double blind test is to
destroy any possibility of a placebo effect. placebo effects can be seen
whenever the test subject can sustain a belief in a particular outcome
(e.g. the tablet they have given me will make me feel better, the
smaller loudspeakers have worse bass reproduction that those monster
units, the fostex converters can never sound as good as the apogees).

in unblind and single blind tests, nothing exists to prevent these pre-
existing beliefs from influencing the outcome (though single blind
removes the beliefs held by the test subject, it does not remove the
influence of the beliefs held by the tester). a double blind test
removes the placebo effect because:

	* the test subject is exposed to several different conditions
	   (this is not medical testing)
        * the test subject has no idea what equipment or signals
           are used for each condition

the same holds true for the experimenter. they are both "blind" about
what is being tested at any point in time. no placebo effect.





More information about the Linux-audio-user mailing list