Nothing Special   »   [go: up one dir, main page]

How Amp/Pedal Capture Works

Record

Input/Output samples of the amp or pedal are recorded for creating a digital model.

Train

Machine learning algorithms are used to train a real-time digital model of the tone from the recordings.

Make some noise!

Models are easily loaded into the Proteus plugin for unlimited tone possibilites on your guitar. Check out the Tone Library to see what amps and pedal captures are available. Proteus Knob and Snapshot models are now compatible with the Two Notes GENOME plugin!

Plugin Downloads

Proteus

Proteus is the new standard in GuitarML device capture, maintaining sound quality while dramatically decreasing CPU usage from the SmartPedal. Create your own Knob captures of the full range of a Drive / Gain / EQ knob from your amp or pedal, or quickly create snapshot models of a single setting. Download the Proteus Capture Utility and check out the tutorial videos for the easiest way to capture Amplifiers and Distortion/Overdrive/Boost pedals, and check out the growing ToneLibrary for pre-made tones from popular and rare guitar gear.
Special thanks to John Stutts for the hand sketched graphic and Stefan Schmidt for the 3D pedal and waves Blender model.

Version 1.2 adds a 3-Band EQ and default IR with an updated look.

Version 1.2 Available formats (from installer):
-Windows 10/11 64-bit (VST3, AAX)
-Mac 10.11 and up (AU, VST3, AAX)
-Linux (VST3, AAX)

Epoch Amp

Epoch Amp is a amp modelling plugin designed to bring realism and versatility to your digital studio. Containing four amp models ranging from clean to edge of breakup, crunchy to high gain, you will constantly be discovering new sounds to add to your sonic arsenal. The full range of the gain knob has been meticulously captured for each of the four amp modes. The plugin also features a mix control for mixing in your input signal with the dirty signal. It has built in Impulse Responses hand picked for each amp mode, which you can turn off to use your own. Epoch Amp is intended to be used as the foundation of your signal chain to add any number of other effects to.
Special thanks to John Stutts for the 3D amp model used to create the plugin graphics.
Version 1.0

Available formats (from installer):
-Windows 10/11, 64-bit (VST3, AAX)
-Mac 10.15 and up (AU, VST3, AAX)
-Linux (VST3, LV2)

NeuralPi

NeuralPi is a DIY guitar pedal using neural networks to emulate real amps and pedals on a Raspberry Pi 4. Models trained from recordings of real amps and pedals can be loaded into the plugin for endless possiblities on your guitar. GuitarML provides the cross compiled VST3 plugin to build your own NeuralPi, as well as Windows and Mac installers (VST3, AU, Standalone) to control the NeuralPi over WiFi. The Win/Mac plugins are fully functional guitar plugins that allow you to try GuitarML's most advanced amp/pedal models without building the NeuralPi. Version 1.3 has the ability to load conditioned models, in addition to snapshot models. This type of model more accurately replicates the full range of a gain/drive knob. See the GitHub ReadMe and Build Guide for more information.

v1.3 Available formats (from installer):
-Windows 7 and up, 32-bit or 64-bit (Standalone, VST3)
-Mac 10.11 and up (Standalone, AU, VST3)
-Linux (Standalone, VST3, LV2)

NeuralPi 3-D Printed Case

Three .stl files are included for 3-D printing your own NeuralPi compatible case. Includes box with all necessary RaspberryPi4/HiFiBerry DAC + ADC input and output holes, removable cover, and RCA output cover.

The Prince

The Prince is a plugin of my homebuilt transparent overdrive style pedal, cloned using neural networks. The graphics were created from actual photos of my pedal using a "stop motion" technique. The plugin features three GuitarML neural network models conditioned on the Gain and Tone knobs, one each for Overdrive, Boost, and Distortion modes. The Prince should be used with an impulse response plugin (such as Pulse) to emulate playing the pedal through an amplifier. Use two instances of the Prince to emulate a highly sought after dual overdrive.
This plugin is intended as an example of how a pedal/amp builder (hobbyist or professional) can faithfully recreate their work in the digital world using neural networks.

Version 1.0

Available formats (from installer):
-Windows 10/11, 64-bit (VST3, AAX)
-Mac 10.15 and up (AU, VST3, AAX)
-Linux (VST3, LV2)

  • Source Code

    • TS-M1N3

    TS-M1N3 is a guitar plugin clone of the TS-9 Tubescreamer overdrive pedal. Machine learning was used to train a model of both the drive and tone knobs for an accurate recreation of the pedal in all possible configurations. Version 1.2

    Available formats (from installer):
    -Windows 10/11, 64-bit (VST3, AAX)
    -Mac 10.15 and up (AU, VST3, AAX)
    -Linux (VST3, LV2)

    Chameleon

    Chameleon is a guitar plugin using neural networks to create three distinct sounds from a vintage style amp head. Chameleon's core sound comes from a neural net inference engine which allows the plugin to disguise itself as a high end tube amplifier. While SmartAmp and SmartAmpPro are more experimental in nature, Chameleon uses the state-of-the-art RTNeural engine which brings CPU usage down to compete with other commercial plugins. Version 1.2



    Available formats (from installer):
    -Windows 10/11 64-bit (VST3, AAX)
    -Mac 10.11 and up (AU, VST3, AAX)
    -Linux (VST3, AAX)

    SmartAmp

    The SmartAmp is a guitar plugin that uses machine learning to mimic all the nuances of a tube amplifier. It features a clean and lead channel with EQ and gain controls. Version 1.3

    Available formats (from installer):
    -Windows 10/11 64-bit (VST3, AAX)
    -Mac 10.11 and up (AU, VST3, AAX)
    -Linux (VST3, AAX)

    SmartPedal

    The SmartPedal uses the same WaveNet model architecture as SmartAmp, and you can load your own trained models from the PedalNetRT code. Capture your own devices and load the models using the SmartPedal. Download the Capture Utility zip file, which includes a Google Colab script and wav file, for the easiest way to create your own models. Version 1.5

    Available formats (from installer):
    -Windows 10/11 64-bit (VST3, AAX)
    -Mac 10.11 and up (AU, VST3, AAX)
    -Linux (VST3, AAX)

    Support

    GuitarML believes in advancing music technology by releasing it as free and open source for anyone to use and learn from. Because this software is free, your support is extremely valuable in making sure this work continues to grow and improve. Funding will go directly back into making GuitarML better.

    I'd like to personally thank our supporters, without you this work could not continue and grow!

    About

    GuitarML uses machine learning and DSP technolgies to create amazing open source plugins and pedals. While using A.I./machine learning was a novel concept when I first started working on GuitarML in 2020, it is now available in many commercial and open source products, including NeuralDSP's Quad Cortex, ToneX, and NAM Neural Amp Modeller. It has been exciting to see this technology grow and mature. In addition to neural net modelling, GuitarML works on and promotes other open source solutions for creating music, including PCB design for analog and digital guitar pedals. GuitarML is working to make this technology easy and accessible to musicians, developers, and A.I. researchers.

    About the author: My name is Keith Bloemer, an engineer and guitar player whose search for great guitar tone led me to create GuitarML. By making these projects free and open source, I hope to encourage people to get their hands dirty with coding, engineering, art and music. GuitarML is only possible due to the generosity of open source developers and researchers. In that same spirit, I hope these projects inspire others to follow their curiosity and push the boundaries of what's possible in music technology.

    Contact

    We would love to hear about your experience using GuitarML's plugins or pedals! You can also email your own trained models and they may be included in the ToneLibrary.