GuitarML Services

Digital Modelling Capabilities

GuitarML specializes in modelling analog audio hardware using open source machine learning technologies. Here is a list of current capabilities:

High quality real-time models of analog hardware using machine learning:

• Distortion/Overdrive/Boost
• Amplifiers (clean or overdriven)
• Most EQ characteristics
• Saturation
• Rig Captures (model of multiple FX limited to the above device types)

• Snapshot models (single sound)
• Parameterized models (models of the full range of a control or multiple controls like Gain, EQ, etc.)

Note: Typically the training data for creating models (audio input/output recordings) will be provided by the client.

* Cannot currently model time based effects (Reverb, Delay, Flange, Phaser, Wah)
* Machine Learning and A.I. is not a "one size fits all" solution. An initial test will be done to determine any limitations or challenges with modelling your particular audio device.

Software Development

• Real-time plugins using neural net/machine learning models
• Experience with Impulse Responses and other audio DSP
• JUCE framework plugin development (Windows/Mac/Linux)
• Multiple format plugins through JUCE (Standalone / VST3 / AU / AAX / LV2)
• Limited User Interface Design/Development (main focus is the DSP side)
• Experience with embedded linux devices/OS such as Raspberry Pi, Elk Audio OS, Patchbox OS
• Software Dev Experience includes: c++ / Python / PyTorch / Tensorflow / Keras / Git / Continuous Integration

Testimonials

Keith is a brilliant programmer and literally a genius for advancing Machine Learning in the field of audio processing. He is fantastic to work with, very supportive and quick to respond. Looking forward to doing many more projects with him!

Keith's innovative, open-sourced initiative in digital guitar engineering is as effective as it is efficient, simplifying access to an important, emerging technology.