convert sounds to pictures and vice versa
Jupyter Notebook Python
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
.gitignore
README.md
SeiLeiseIchWillDasSehen.ipynb
main.py
mydeafdog.py
samplingfunctions.ipynb
whistle.wav

README.md

CODENAME: how to yell at your deaf dog

FAQ:

  • Why would anybody want to yell at their dog? No idea, ask them. Also: Don’t shoot the messenger, this is just the software to do it.
  • What is this about? The central idea is to find a mapping from the space of 1-dimensional signals (e.g. sounds) onto some 2-dimensional, compact subspace of R^2. Furthermore it should be robust to noise (small changes in the Fourier coefficients should not result in big changes of the image). Robustness w.r.t. perturbations in the image cannot be achieved since there is no continuous, injective mapping from R^2 to R, but we find that the family of Peano curves does a good job in coming as close to that invariance as possible.
  • What does this have to do with dogs? Since the DFT in use outputs complex values (or equivalently there are coefficients for sine and cosine) they are mapped to only two of the three RGB channels: (Re(z) to G and Im(z) to B, which is basically just what dogs see all the time anyways. So while (from a technical point of view) it’s perfectly ok to use this software to yell at your deaf relatives, it might confuse them since changes in the R channel do not result in different tones! Dogs, however, wouldn’t notice those changes.