[LAU] music from image metadata

[Thomas Mayer]

Hi,

On 01.04.2014 15:24, Patrick Shirkey wrote:
> I suppose I could translate it into ascii and then use that. I vaguely
> recall that someone has written an ascii to audio tool.

Use exiftool, as described in one of the links, I have provided. The
Debian package is called libimage-exiftool-perl.

Running exiftool -a -u -g1 <filename> outputs 4kB of text.

>> Here's an example, that uses the returned data from a search, including
>> a description on the algorithm to generate the sound:
>> https://soundcloud.com/residuum/twitter-sonification
>>
> 
> It's an interesting piece.  Quite soothing in it's own way. Did you select
> specific instruments/noises or did you let it automate that process too?

Have you read the description for the piece? It is described quite clearly:

| This is not musically interesting, but only showing a way to sonify
| data, tweets are processed and sonified in the following way:
| - content is splitted into ASCII values, and each value creates a
| sound of 1/10 s length.
| - non-printable characters make a pause, all others are mapped to the
| respective MIDI value (! = 1, " = 2, ... A = 33, B = 34, etc.), and
| then used to drive a sine oscillator
| - the longer the username, the louder the sound.

> Do you have others in that style?

No, not really. I have made a tool for sonifying stock markets and
bitcoin market prices, but nothing fancy and only varying small parameters:
http://www.residuum.org/moneytalks.html
https://github.com/residuum/Money-Talks
https://github.com/residuum/PuRestJson/tree/master/examples

Hth,
Thomas
-- 
"Chaney was aware that anything, however small, can get the eye of the
media if it's repulsive enough." (Robert Anton Wilson - The Universe
Next Door)
http://www.residuum.org/