[Frameworks] optical soundtrack generator?

[Eric Theise]

Interesting thread.

Scott (Stark), the Kapotski "Optical Sound generator" you launched the
thread with (http://www.kapotski.be/wp/?p=204) looks to be a simple circuit
that maps values of light-to-dark to frequency. A bit like a theremin but
for light instead of proximity.

(When I was a kid I built circuits like that from scratch using electronic
components (Chicago alert: probably purchased from Allied Electronics on
Western Av (https://chicagology.com/silentmovies/alliedradio/)). I remember
being fascinated by the "electric eye" that opened the doors of a grocery
store on Foster Av, just east of Damen.)

Sandy McLennan's recommendation of Matt McWilliams' soundtrack.optical
looks promising. It doesn't seem as if it's under heavy development  – only
one code commit this year, from January – but it seems quite far along in
its functionality. You say you're unfamiliar with Processing; it's an
environment developed in Java that was designed to be artist-friendly. Been
around for nearly 20 years and it's well-documented and has an active user
base. Roughly speaking, programs are called sketches and the environment
compiles these sketches before displaying results. It's easy to get started
with, there are many tutorials, and McWilliams directs you to one (
https://github.com/processing/processing/wiki/How-to-Install-a-Contributed-Library).
The software I initially used to do the frame-by-frame analysis of the
colors in your To Love or To Die was adapted from a Processing sketch. Even
if you end up not using soundtrack.optical you'd benefit from studying the
code to see what it's doing.

I ended up writing a Python program to analyze To Love or To Die; Python is
best suited to your desire to use a scripting language. I used OpenCV
because I was doing cluster analysis on video attributes but a quick search
suggests that Python libraries like pyAudioAnalysis (
https://github.com/tyiannak/pyAudioAnalysis/) or scikit-sound (
http://work.thaslwanter.at/sksound/html/) would be relevant – or overkill –
to the audio import and analysis aspects of your project. One of Python's
SVG writers, or something in Matplotlib, might be appropriate in generating
your output although, again, they might be overkill.

Personally, I'd start by installing soundtrack.optical and scaling its
output for use with 35mm frames.

Eric


On Thu, Sep 10, 2020 at 8:53 AM <sstark at hi-beam.net> wrote:

> Thank you all for the suggestions of generating an optical track from an
> audio source. I like the idea of finding an optical sound camera (I do have
> a 16mm Auricon that would do this for 16mm, not 35). And the digital angle
> from sixteenmillimeter.com or Scott D’s suggestion sound interesting if a
> bit daunting (I’ve no experience with Processing). Anyway I’ll keep the
> list posted of any progress.
>
>
>
> Scott Stark
>
> www.scottstark.com
> _______________________________________________
> FrameWorks mailing list
> FrameWorks at jonasmekasfilms.com
> https://mailman-mail5.webfaction.com/listinfo/frameworks
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman-mail5.webfaction.com/pipermail/frameworks/attachments/20200910/bc61c593/attachment.html>