Chameleon is a guitar plugin using neural networks to create three distinct sounds from a vintage style amp head. EQ and gain were added to allow further modification of the three core sounds, named Red (high gain), Gold (crunchy), and Green (crisp and clean). In the same way a real amp head is used with a cabinet and other effects, this plugin is intended to be used in the signal chain along with IR's (cab sim), reverb, and any number of guitar effects.
Chameleon's core sound comes from a neural net inference engine which allows the plugin to disguise itself as a high end tube amplifier. The engine uses a stateful LSTM model, which improves the sound quality of the previous stateless LSTM used in the SmartAmpPro. It also improves CPU usage compared to the SmartAmpPro and SmartGuitarAmp.
Check out sound demos on YouTube: Heavy Demo, Funky Demo
Check out the tech article on Towards Data Science
Chameleon is part of the 2021 KVR Audio Developer Challenge
- Download the appropriate plugin installer (Windows, Mac, Linux) from the Releases page.
- Run the installer and follow the instructions. May need to reboot to allow your DAW to recognize the new plugin.
Re-creation of the LSTM inference model from Real-Time Guitar Amplifier Emulation with Deep Learning
The Automated-GuitarAmpModelling project was used to train the .json models.
GuitarML maintains a fork with a few extra helpful features, including a Colab training script.
The plugin uses RTNeural, which is a highly optimized neural net inference engine intended for audio applications.
# Clone the repository
$ git clone https://github.com/GuitarML/Chameleon.git
$ cd Chameleon
# initialize and set up submodules
$ git submodule update --init --recursive
# build with CMake
$ cmake -Bbuild
$ cmake --build build --config ReleaseThe binaries will be located in Chameleon/build/Chameleon_artefacts/
