Variable Audio

In the original eye-tracked film, there is only fixed audio. Adding to the output film, a large database of nearly 60 voice-overs for each of the narratives depicting the internal thoughts of the character at the given point in the story is triggered by the eye-tracking data. Thus, an initial database details when the voiceovers are allowed to be triggered, detailing the name of the audio track, the starting time when the track is allowed to be played, and the duration of when this track is allowed to be played, effectively detailing the ending point. These details are parsed in the patch shown in Figure-6 to create a frame based real-time parsable script that will simply detail the audio file for each of the 4 characters.

Using the output from this patch, the line number is sent to the text file to gather at any given time the corresponding file name of the voice over file to be played as shown in Figure-7. Then, by knowing the source dimension, a simple Max-gate controls which of the four text files are triggered. With the file name, this is sent to a buffer to be output (see Figure-8). Additionally, note how the gate just before this buffer ensures that only 1 sample can be output at any given time. This ensure that the current sample has to finish before a new sample can be played.

Figure 6: A database of about 60 sound clips and the corresponding start times and durations they can be triggered are parsed to create a database for each frame of the video corresponding to a single audio track.

Figure 7: The database created from Figure-6 is parsed in realtime by this patch to gather the triggered sound clip's file name.

Figure 8: The source dimension controls the output of the gate which triggers a file name in the samples database. This file name is sent to be played in the corresponding buffer.