Skip to content
Use pre-trained deepdream on audio spectrograms.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

HAMR 2015: DeepDreamEffect


  • Christian Dittmar
  • Stefan Balke


I used Google's DeepDream processing as an audio effect. Therefore, I export music magnitude spectrogram as RGB channels of PNG images and apply so-called 'Gradient Ascent' with pre-trained networks to these images. Afterwards, I convert the resulting images to spectrograms again and resynthesize them using Griffin and Lim's method.

More information and further demos can be found here:


You need to have a running version of caffe ( on your machine There are hundreds of tutorials describing how to install it (it takes some time...). Anyhow, the notebook with an example is best viewed here:

Music Copyright

The used music is copyright free to our best knowledge. However, we want to thank the creator. If any legal problems occur, please contact us.

You can’t perform that action at this time.