Grafit Pro

Visual programming for real-time data processing

Publisher

Grafit Pro is developed and published by SIA SPACE, a company registered in Latvia (European Union).

Company	SIA SPACE
VAT	LV40203141856
Address	"Padambji", Vārpu ciems, Baldones pag., Ķekavas nov., LV-2125, Latvia
Website	siaspace.eu
E-mail	work@siaspace.eu

Banking details

Bank	Swedbank AS
BIC / S.W.I.F.T.	HABALV22
Account	LV72HABA0551044987418

Author

Mārtiņš Dāboliņš — creator and lead developer
work@siaspace.eu

Technology

Grafit Pro is a native macOS application built with Swift 6, Metal, SwiftUI, and AppKit. All GPU rendering uses the Metal framework. The app is optimized for Apple Silicon and requires no external runtime or cloud services.

Signal Processing

All signal processing algorithms in Grafit Pro — including FFT spectral analysis, Welch's method for power spectral density estimation, magnitude-squared coherence, hemispheric asymmetry index, and IIR filtering — are original implementations written in Swift and Metal compute shaders. These algorithms are based on standard digital signal processing methods documented in academic literature:

Welch's Method — P. D. Welch, "The use of fast Fourier transform for the estimation of power spectra," IEEE Transactions on Audio and Electroacoustics, 1967
EEG Band Powers — Standard frequency band definitions (Delta 0.5–4 Hz, Theta 4–8 Hz, Alpha 8–12 Hz, Beta 12–30 Hz, Gamma 30–100 Hz) as used across clinical and research neuroscience
Coherence — Standard magnitude-squared coherence formula: Coh_xy(f) = |S_xy(f)|² / (S_xx(f) · S_yy(f))
Asymmetry Index — Standard hemispheric asymmetry: AI = (R − L) / (R + L)
IIR Filters — Butterworth and notch filters using Audio EQ Cookbook biquad formulas, processed via Apple's Accelerate framework (vDSP)

No third-party signal processing libraries are used. The GPU compute path implements a Cooley-Tukey Radix-2 DIT FFT in Metal shading language.

Machine Learning

Grafit Pro includes two on-device machine-learning analysis nodes (introduced in 2.6). Both run entirely on your Mac via Apple Core ML — audio is never uploaded anywhere.

Pitch Detection uses Spotify's open-source basic-pitch model (Apache License 2.0), converted to Core ML format and bundled with the app. Reference: Bittner et al., "A Lightweight Instrument-Agnostic Model for Polyphonic Note Transcription and Multipitch Estimation", ICASSP 2022.
Audio Beat Detection is an original in-house neural network, trained on synthetic drum data and fine-tuned on the Groove MIDI Dataset (see attribution below). The streaming analysis frontend is test-verified bit-identical to the training pipeline.

Open Source Acknowledgments

Basic Pitch

Polyphonic pitch detection machine-learning model used by the Pitch Detection node. The bundled model is converted from Spotify's released basic-pitch model to Apple Core ML format.
Copyright 2022 Spotify AB · Apache License, Version 2.0

Apache License 2.0 — summary & link

Licensed under the Apache License, Version 2.0 (the "License"); you may not use the licensed files except in compliance with the License. You may obtain a copy of the License at apache.org/licenses/LICENSE-2.0. Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. The full license text is also included inside the app: Settings → About → Open Source Acknowledgments.

Groove MIDI Dataset — ML training data

The drum onset detection model used by the Audio Beat Detection node was trained in part on the Groove MIDI Dataset by Google Magenta (Jon Gillick, Adam Roberts, Jesse Engel, Douglas Eck, David Bamman — "Learning to Groove with Inverse Sequence Transformations", ICML 2019), magenta.tensorflow.org/datasets/groove, licensed under CC BY 4.0. The dataset was modified: MIDI events were converted to onset labels and the audio was used as model training input. This attribution does not imply endorsement by Google. No dataset audio or MIDI is included in the application.

VGER

2D GPU rendering library used for the node graph editor canvas, cable drawing, and UI primitives.
Copyright (c) 2021 Audulus LLC

MIT License

MIT License

Copyright (c) 2021 Audulus LLC

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Legal

Grafit Pro and the Grafit logo are trademarks of SIA SPACE. macOS, Metal, Swift, SwiftUI, AppKit, Apple Silicon, and App Store are trademarks of Apple Inc.