DrumGAN
Generate the right drum sounds quickly
and effortlessly using AI
DrumGAN - AI drum tech at your fingertips
DrumGAN is an AI model that utilizes deep-learning algorithms to produce drum sounds. With DrumGAN you can effortlessly generate drum sounds for your music tracks in no time. DrumGAN has already been integrated into the cutting-edge drum sound design software Backbone, from sound technology leader Steinberg. You can see for yourself how DrumGAN works in the web-based demo app. DrumGAN is developed by Music Team at Sony Computer Science Laboratories - Paris (Sony CSL - Paris).
DrumGAN overview
DrumGAN, developed by Sony Computer Science Laboratories (Sony CSL), is based on a Generative Adversarial Network (GAN). The DrumGAN AI model was trained on a wide range of drum sounds in three categories: kick drum, snare drum, and cymbal.
DrumGAN is designed to synthesize high-quality drum kit sounds quickly, with a high degree of originality, and to a professional standard.
In developing DrumGAN, Sony CSL used professional musicians to assess the quality as well as metrics to objectively evaluate the quality and diversity of generated drum sounds.
The DrumGAN technology has been integrated as into commercially available music software, such as drum sound design software Backbone, from Steinberg.
Why use Sony's DrumGAN
Cutting-edge, innovative algorithms developed by Sony.
Quickly generate original, professional-quality drum sounds to streamline your digital music production.
Ready to integrate into your own application or you can re-train the model with your own dataset.
Watch this video of DrumGAN integration into Backbone from Steinberg
DrumGAN is integrated into the drum sound design software Backbone 1.5 from Steinberg Media Technologies GmbH (Steinberg), for generating drum sounds to be used in Digital Audio Workstations (DAW).
This short video gives a brief explanation of the GAN AI model used to build DrumGAN. It then shows how Backbone uses DrumGAN to generate different kick drum, snare drum, and cymbal sounds based on the position of adjustable sliders.
The drum sounds generated can be further tailored to the musician's taste by using the variation controls.
Read more on the Sony CSL - Paris Music Team site.
Chainsaw Man - Behind the Scenes Using DrumGAN tech to make an "impossible" soundtrack
This video shows how Dr. Javier Nistal Hurle, from Sony Computer Science Laboratories - Paris is using the technology of DrumGAN to make “impossible” musical ideas reality.
The tech behind DrumGAN is highly adaptable. What you get out depends on how the model is trained. To create the soundtrack for the acclaimed 2022 anime film Chainsaw Man, composer/producer Kensuke Ushio collaborated with Dr Nistal Hurle and took the DrumGAN model and trained it on chainsaw sounds to create ChainsawGAN.
In the video Mr. Ushio describes how ChainsawGAN is used to supplement the intricate drum part and used to fill in the gaps in the music that "humans would not be able to think of".
The music he created shows how AI assists the creativity of the musician – it does not replace it. AI is used to bridge the gap between the composer's ideas and their realization.
Who will use DrumGAN
- Companies who produce drum sample libraries
- Record labels
- Sound engineers
- Musicians, DJs, Producers
- Film and games producers
- Audio software developers/Product owners
- Music and technology researchers
The research behind DrumGAN
One of the creators of DrumGAN is Javier Nistal Hurle, Sony Computer Science Laboratories - Paris (Sony CSL - Paris) - a research laboratory working on artificial intelligence applied to such areas as language, music, creativity, and sustainability.
Javier’s research focuses on creating interactive tools that Deep Learning can bring into music. His work focuses on using the latest technology to broaden the sonic and musical palette available to producers and other music creatives.
For more information about Sony CSL - Paris, please Contact Us.
