1ABXTEST(1)                    MPEG Audio Decoder                    ABXTEST(1)
2
3
4

NAME

6       abxtest - double-blind ABX comparison testing script
7

SYNOPSIS

9       abxtest [-n min] [-m max] [-g goal] A-cmd B-cmd
10

DESCRIPTION

12       abxtest  is a tool for conducting listening (or other subjective) tests
13       to determine whether a listener can discern a  difference  between  two
14       subjects under test, denoted A and B.
15
16       The listener conducts several trials in which an unknown subject, X, is
17       randomly chosen to be either A or B.  The probability that X is A or  B
18       is the same. The listener must decide whether X sounds the same as A or
19       the same as B.  The listener may repeat A, B, and X as  many  times  as
20       necessary before reaching a verdict.
21
22       In  order  to  prove  the hypothesis that a difference can be discerned
23       between A and B, the listener must disprove the null hypothesis that  A
24       sounds  the same as B.  This is accomplished by correctly identifying X
25       often enough that randomly guessing correctly the same number of  times
26       would be statistically improbable.
27
28       The A and B subjects for abxtest are actually arbitrary shell commands.
29       The normal mode of use  would  be  to  give  commands  which  play  two
30       slightly  different sounds. The standard output and standard error from
31       each command are suppressed so as not to reveal the identity of X.  The
32       listener  may  interrupt  an executing command (usually by pressing ^C)
33       without terminating the test.
34

OPTIONS

36       -n min Set the minimum number of trials.  Trials  will  continue  after
37              this number if the statistical probability goal has not yet been
38              met. The default is 10.
39
40       -m max Set the maximum number of trials. Trials  will  end  after  this
41              number  even  if  the  statistical probability goal has not been
42              met. The default is 20.
43
44       -g goal
45              Set the minimum statistical probability goal. This is the proba‐
46              bility that the number of correct identifications is the same as
47              random guesses. In order to disprove the  null  hypothesis,  the
48              result  of  the test must be a probability less than or equal to
49              this goal. The default is 0.05.
50

FILES

52       /dev/random
53              By default this file is read to generate a random sequence of X.
54

ENVIRONMENT

56       RANDOM_FILE
57              If set, the file named by this variable will be read instead  of
58              /dev/random to generate a sequence of X.
59

NOTES

61       abxtest  uses  the  binomial  distribution  to accurately calculate the
62       probability that the result is the same as random guesses.
63

AUTHOR

65       Robert Leslie <rob@mars.org>
66
67
68
69MAD                            16 November 2000                     ABXTEST(1)
Impressum