Image Processing Reference
In-Depth Information
Fig. 11.7 Virtual source positions, (
plus symbol
): originally intended, (
bullet symbol
): measured
at different positions,
A
N
and
B
N
aliasing frequency can change at each listening position. The spatial aliasing may
cause extra artefacts to the listener, e.g. colouration artefacts. Overall, the typical
results shown as examples are mostly valid within the view angle.
In addition to the objective evaluation method, subjective evaluation methods
are also available for audio rendering system validation as shown in earlier other
works [
23
,
24
]. The subjective evaluation can follow the ITU-R BS.1534-1 guide-
lines [
25
]. During evaluation the participants are asked to move around the whole
room including outside the viewing angle. Subjects record the virtual sources origin
and depth using a visual place marker and questionnaire.
The subjective evaluation was conducted in the studio illustrated in Fig.
11.5
.
Considering previous listening experiences, the listening panel was chosen for the
test such as (a) Electronic engineer and recording musician, (b) Electronic engineer,
(c) Audio/visual engineer with experience in spatial audio, (d) Broadcast and media
professional, (e) DJ and musician. In Table
11.1
it can be seen that the participants
accurately localised the audio objects origins over the majority of the listening
space. It is also shown that non-focused sources (Table
11.1a
) can be reproduced
more accurately than focused (Table
11.1b
), especially for the perception of depth.
It is found that the destructive acoustic errors of the focused sound source can be
reduced when the listeners are within the calculated viewing angle. Another finding
shows that the sound field generated from multiple virtual audio objects improve
perceptual sensitivity, with a higher directional accuracy by approximately 6-14 %.
This improvement can be explained using the acoustic side effects produced when
