Skip to content

Commit d78060d

Browse files
authored
Create README.md
Signed-off-by: smallbutfine <80461402+smallbutfine@users.noreply.github.com>
1 parent a6772e9 commit d78060d

File tree

1 file changed

+62
-0
lines changed

1 file changed

+62
-0
lines changed

README.md

Lines changed: 62 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,62 @@
1+
# GuitarAmpModellingGUI
2+
GUI for easier installation and training of neural network models for guitar amplifiers and pedals, based on the GuitarML Proteus models. These are usable for Proteus, Chowdhury-DSP BYOD and even NeuralPi, on all platforms incl. Linux and RaspberryPi.
3+
4+
What is this?
5+
GuitarML's work on Proteus. NeuralPi and Proteusboard (hardware) is amazing.
6+
Yet, it is not easy to wrap your head around if you are not familiar with programming, AI, machine learning, neuronal networks.
7+
So, Keith Bloemer a.k.a. GuitarML set up a Google Colab script to give people the Opportunity to train their own models online.
8+
Still, I thought that things could be easier, and I wanted a faster way to work with the python scripts.
9+
So I automated some things on my Windows 10 machine.
10+
I assume, that most musicians use this OS.
11+
This repository contains a bunch of Windows batch scripts and a freepascal based GUI solution.
12+
They allow me to quickly install, re-install and test everything and even start a local offline model training.
13+
I tried to make the install of all prerequisites as automated as possible, yet as little intrusive of global OS settings as possible.
14+
Despite of a quiet install of the actual python3 with standard settings, everything is hold inside of one folder or are just temporary.
15+
The python prerequisites are hold in a virtual environment to not collide to any other python installation.
16+
The offline training takes it's time, though, but is managable with a fast PC.
17+
I will provide a convenient binary installer soon, so people will not have to deal with code organization and compilation.
18+
19+
How to build/compile
20+
21+
Building the GUI is quite easy.
22+
Download the repository to a local folder on your Windows 10 PC.
23+
Download and install CodeTyphon 8. You find it here: https://www.pilotlogic.com/sitejoom/index.php/downloads/category/14-codetyphon.html
24+
As soon as you have the development environment IDE running, it is fairly simple.
25+
(Looks complicated, but is pretty straightforward for everyone who has ever used an IDE for programming, no matter what language.)
26+
Open the Lazarus (Pascal) project file from the repository folder and compile it.
27+
This should work out of the box.
28+
You get the binary executable in form of an .exe file right in this repository folder.
29+
30+
Usage:
31+
Once the Lazarus project is compiled, the binary executable can simply be double clicked or started right from the Lazarus IDE.
32+
The GUI is made for simplicity and works without submenu hassle.
33+
It is mostly self-explanatory.
34+
The help menu opens Youtube videos of GuitarML with really good information, how to capture an amp, pedal or VST plugin.
35+
36+
When it comes to the building/training of the AI model, that can be used in VST plugins,
37+
the GUI offers a very easy nearly "mouse only" way to initiate OFFLINE/local training.
38+
The scripts are highly automated and there is as little manual work as it gets.
39+
40+
You need the keyboard only for entering the desired name of the model and to confirm closing of scripts.
41+
(I left that in, so you can see what actually happened and if errors occured. Which is most likely not happening.)
42+
43+
Yes, this works just by simply clicking a few buttons.
44+
The hardest part is capturing the audio files of your gear or plugins.
45+
Even the manual file renaming can be left out. You can import the files from wherever they are, easily.
46+
47+
Once you have these, your job is ... waiting for the scripts to do the hard work on their own.
48+
49+
When the model is trained and ready to use, you can install it with the click of a button to whereever you need it.
50+
(Well, you have to give it a name....)
51+
52+
53+
These are great times for musicians.
54+
Machine Learning and Neuronal Network Models revolutionize the way music is made.
55+
And it is great to see, that people open source their software.
56+
With excellent free plugins and DIY hardware devices, that are by far more affordable than most similar commercial solutions.
57+
They can give you sound quality that is indistinguishable from the expensive hardware they are modelling, in studio and on stage.
58+
59+
Keep making music.
60+
Rock on.
61+
62+
Best regards, Martin

0 commit comments

Comments
 (0)