[Text available only in English] Sounds, images and biometric data of birds are used to recreate their sound and behaviour. Due to interpolation between different kinds and glitches in the learning algorithm the sounds start very natural and get more and more contrived and structured as in modern minimal or electronic music while the images become more and more abstract and generalized. Similar to my other work every performance is very different from the others before. 3 different neural networks are interacting and are used for the recreation of sound using supercollider and image using open frameworks. Everything happens in real time.