Virtual Pipe Organ Project 

 

Project 
Organs 
Features 
Free stuff 
Contacts 
History
Screenshots
Pictures
Demos
Updates
Order

                        

The Quiz - results, 2nd part

This part seemed to lack the interest of public a bit, so the total number or submitted questionaries was only 32. This unfortunately reduces its significance. The right choises are marked in red:

Results per composition
  authentic virtual no answer % of right answers
Composition 1 20 10 2 62,5%  !
Composition 2 15 15 2 47%
Composition 3 14 16 2 50%
Composition 4 17 12 3 37,5%  !
Composition 5 13 17 2 41%
Composition 6 10 20 2 31%  !

The diversity of answers is larger than in the first part. This is most probably the effect of limited number of questionaries submitted. But the surprising reality is that the overall ability to make right choise decreased compared to the first part of the quiz. This might mean:

  1. the sample set sounds more convincing than real life organ since people attribute the authenticity easier to the virtual recording having the countrepart available. I think my ideas about the "degree of reality" of a sample set might be helpful for understanding this.
  2. Only the unexperienced part of audience answered the second part of the quiz contributing to the bad result. This is not likely, since the on the other hand the Composition no. 1 was well recognized to be authentic by two thirds of the audience what was not the case in the 1st part of the quiz.
  3. too few people answered the test in this part and so the outcome is not fully statistically relevant. This is most probably the case. I expect that if more people answered the test, the results would look similar to the 1st part.
Right answers per user (see below the graphical representation)
no decision 0/6 1/6 2/6 3/6 4/6 5/6 6/6
2 2 3 7 9 4 5 0

The results follow approximately the expected Gaussian curve, which centers around the middle value (3 good answers out of 6) = an ideal "pure guess" value. This time it is slightly shifted towards left (meaning that people were likely to make the right attribution in less than 50% of cases) There are 5 users who got almost all the right answers. There are two users who gave all wrong answers. These probably hear the difference in sound well but attribute it wrongly. This could again mean that the sample set sound more "perfect" than the real organ and so the users can be misled, expecting the perfection to be the quality of the real organ.

The surprising result is that having both authentic and virtual recording of the same piece does not help to the solution of the audition quiz and again it is more guessing than knowing.

Follow to the analysis of the reasons given for the discernment of the virtual/authentic recording.

 


A note to the complementary recording for the Composition no. 4. If you listen carefully to the complementary recording available for the Composition no. 4, you will most probably hear a "strange pause" just before the last chord. Indeed, this is a cut, which I had to introduce since the very last chord of the real recording was corrupted by heavy bench crunches and so it would disqualify the comparison. Unfortunately, I do not have another version of the authentic recording (when doing it I was not thinking of publishing it for the quiz and so I did not pay attention to this). I had to use the last chord from the virtual recording. However, no user was mislead by this, so I am sure that this nasty trick did not influence the results of this part of the quiz. All the users, who had 5 out of 6 good answers attributed the Composition no 4 to the virtual recording anyway! Two users only reported, that there is something "strange" happening at this place in the complementary recording. I congratulate to these two persons for the supreme listening capabilities!