Reports on Progress:

Predictor B

Predictor B allows the user to predict the audio quality as a function of down-mix algorithm (Mono, Stereo, 2/1, 1/2, 2/2, 3/0, 3/1, FR).

Example 1

figure: Spat

In this example we simulate that the original 5-channel recording has been down-mixed to Mono. The predicted audio quality at the centre listening position is Fair. It might seem as a very surprising result, however it should be noted that in this example we simulated a full bandwidth Mono down-mix (quite often people associate Mono with bad quality because it is often subconsciously associated with a sound produced by a small loudspeaker having very limited bandwidth and large distortions).

Example 2

figure: Spat

This example shows the result of the predicted quality of the down-mix to Stereo evaluated at the centre listening position. The predicted quality for F-B programme material is Excellent. However, the predicted audio quality at the off-centre listening position is significantly lower - see Example 3.

Example 3

figure: Spat

Please note, that the Predictor B requires the user to declare whether the Picture accompany the audio presentation. The reason for this is the fact, that according to the results of the subjective tests, the evaluation of audio quality at the off-centre listening position may be affected by picture presence for some down-mix algorithms (see Zielinski 2003d). The Predictor B takes this audio-visual interaction into account.