The frames were taken from the video at every crucial point in which there was a substantial change in the motion (approximately every half a second).
The black marks represent motion. The motion is repetitive and expands and contracts. The motion doesn’t just disappear but rather leaves traces of its existence. Even in the frames that are seemingly “clean”, traces of the black smudges can still be seen.
In some cases the traces fade away over a longer span of time, but many of the traces remain as a background for the movements that follow.
* Traces of an occurrence which remain as part of the scenery once the occurrence has passed.
The “breaths” of the sound (ie. the fraction of a second when no sound is heard) come at intervals when the motion changes drastically.