09/06/17 – Random thoughts and continued experiments

Key themes: Chaos, absurdity, subversion, persistence, manipulation.

Basic thoughts on visuals –

  • Camera distortions: heat haze, swirls, colour splitting
  • skybox distortions?
  • objects in the game world with strange behaviors – self multiplying, synchronised movement, orbiting.
  • Perhaps there could be almost pick ups in the game, tiny orbs which orbit the centre point of your vision and divide and multiply depending on the sound and the player’s position.

This, taken from the Audiokinetic Library:

“It is also possible to use it inside the Interactive Music Hierarchy, but this is not recommended as it will affect the timings in an undesirable manner in the context of interactive music.”


Although this raises the question of the need to use the music hierarchy. Experimentation needs to be done in Actor-Mixer Hierarchy.

Been experimenting/bending my mind with modulators, driving RTPC changes with LFO’s, and driving the LFO frequency with an Envelope determined by SlowFast. Approaching limits of comprehension. Threw pitch shift into the mix after realising concrete (possibly terminologically incorrect) rule:

pitch shifter – changes pitch preserving time
time stretch – changes time preserving pitch
playback speed – changes pitch and time simultaneously

If each of these is enacted upon any given track, with values randomised or by or mapped user input control (some degree of the two would be preferable), the results would be interesting.

Issue with Owen’s suggestion of eating own tail: before, it was layered parts of a whole each sliding in and out of place. Now I have a recording of that muddled whole. He suggested layering the best new bits, but initial experiments show it creates a densely muddy and overly-intense new creation not as interesting as what fed it. Perhaps I could try layering several copies of each individual track in a new music segment. Then enacting the weirdness upon it. Another approach would be generating more initial output and recording each track individually. Downside would be they wouldn’t be reacting to the same input, so it becomes tenuously random.

UPDATE (5 mins later) – layering several tracks of the output into the exisiting Bridge1 music segment and messing around with SlowFast DID create some interesting noises. Here is the version where all tracks are stacked thusly in the music segment:


Resulting in this (with random tweaking of SlowFast):

Here with tracks spaced in the segment thusly:


Resulting in:

One obvious problem with the spaced version is that the audio drops altogether occasionally. I can only surmise that this is due to the time stretch on certain tracks finishing them before the the container is looped. A problem with the spaced version is there are noticeable loops. if the pieces were longer this potentially wouldn’t be so much of a problem. Although actually, my favourite of the two is the stacked recording, partially because of the loops.

I went out for a walk to clear my head, and had a thought. The stacked approach would avoid the loop problem with the music segments containing every piece on random sub-tracks. I then added in some tracks of silence (currently seven, the same as how many tracks there are) to make it a little less busy. This does risk an entire loop of silence, so I’m not quite sure what the solution to that is.

Anyway, here’s the Music Editor View showing the first two tracks in the music segment with their random sub tracks:


So here the container is looping between the green and red flags (further testing shows the staggered approach to be unreliably prone to cutting out), on each new loop a random track is played from each track and the effects of SlowFast are applied to it.

Here’s an example. There were one or two cut outs, I’m not totally sure why, but perhaps only a few actual tracks were playing and because of their unique time stretches they finished before the loop did. Anyway, EXAMPLE:

Interestingly I found that so long at it was at a number other than 5, SlowFast could be left untouched and the relationships between the tracks would still be dynamic due to their individual time stretch values.

I think I’ve generated enough weirdness for the day.