I'm Sabian Hibbs
CTO, AI Audio Architect & Systems Innovator
Specializing in the complex intersection of artificial intelligence, advanced audio processing, and large-scale system architecture.
# EUCLID system architecture class MultiHeadBinaryClassifier:
About Me
Visionary Technology Leader
"My journey into audio technology began with a fascination for how sound shapes our experiences. From mixing my first track to building AI systems that can understand audio at a deep level, I've always been driven by the intersection of creativity and cutting-edge technology."
As Co-Founder and CTO of Uhmbrella Ltd., I lead technical strategy for cutting-edge solutions in music copyright protection and AI-generated content detection. My expertise combines audio domain knowledge with machine learning innovation, exemplified by the EUCLID system for synthetic audio detection.
My passion for audio began in game development, where I discovered the power of sound to create immersive experiences. This led me to explore the intersection of audio and artificial intelligence, where I found my true calling in designing systems that can understand and generate sound in ways that mimic human perception.
AI/ML Systems
Expert Designer
Audio Processing
Signal Specialist
System Architecture
End-to-End Solutions
Problem Solving
Critical Thinker
Professional Journey
A timeline of my professional experience and educational background
Work Experience
Co-Founder & CTO
Uhmbrella Ltd.
Leading technology strategy at a startup focused on AI audio detection and copyright protection. Developing multi-head binary classification systems for synthetic audio detection.
AR/VR Research Panel Member
Meta
Contributing to AR/VR research and testing remotely. Providing expertise on audio integration and spatial audio for immersive environments.
Game Audio Course Coordinator
ThinkSpace Education Ltd
Supported ongoing game development using Unity, Unreal Engine 4, and Wwise. Created and delivered course material to master's students via webinars and assignments, liaised with third-party contractors, and contributed to new course content in collaboration with the University of Chichester.
Software Tool Developer
ThinkSpace Education Ltd
Developed in-house software tools including: Automated Batch Video Processor (Python), Batch Video/Audio Speaker Diarization Tool (Python/CUDA), Batch Video Editing Processor, Apple Silicon Permission Classifier Rectifier, File Corruption Detector, and a Novel Lossless/Lossy Audio Compressor comparable to FLAC without decoding overhead.
Assistant Lecturer – Audio Electronics & Mathematics
University of Derby
Assisted in teaching undergraduate courses, delivered seminars and workshops, and supported Level 4 and Level 5 students in Audio Electronics and Mathematics. Created a supportive learning environment to help students grasp complex technical concepts.
Audio Rigger (Technical Audio Designer)
Frontier Developments PLC
Worked on optimizing Jurassic Park Evolution 2 for multiple platforms. Acted as a bridge between sound designers and audio programmers, improving Wwise project implementation. Gained hands-on experience in Quality Assurance and contributed to NDA projects.
Lead Sound Designer
Dark Matter Games
Oversaw audio design for games in Derbyshire, UK. Managed sound effects creation, music integration, and voice-over recording processes for multiple game projects.
Education
Master's Degree in Audio Engineering
University of Derby
Specialized in advanced audio processing techniques and system design.
BSc in Music Technology
Birmingham City University
Focused on audio technology fundamentals and practical applications.
BTEC Level 3 Extended Diploma in Music Technology
Joseph Wright Derby College
Built a foundation in audio engineering and production techniques.
Technical Expertise
Specialized skills across AI, audio processing, and software development
AI/ML System Design
AI & Machine Learning
Design and implementation of custom AI architectures for audio classification and processing
Speech Recognition
AI & Machine Learning
Fine-tuning and optimization of speech recognition models
Vector Databases
AI & Machine Learning
Implementation of high-performance similarity search for audio content
Digital Signal Processing
Audio Engineering
Advanced audio processing and algorithmic sound manipulation
VST Plugin Development
Audio Engineering
C++ audio plugin architecture and implementation
Spatial Audio
Audio Engineering
Implementation of 3D audio systems for VR/AR applications
Middleware Integration
Game Audio
Implementation of audio systems in game engines using industry-standard middleware
Game Engine Audio
Game Audio
Native audio integration with leading game engines
Python Development
Software Engineering
Specialized Python development for audio analysis and machine learning
System Architecture
Software Engineering
Design of scalable, high-performance systems for audio processing
C++ Programming
Software Engineering
Basic VST audio signal processing
Unreal Engine Development
Game Development
Creating games and systems within Unreal Engine
Value Proposition
I am a systems thinker and technical innovator who understands how to build complex, functional, and high-performing AI solutions for challenging real-world problems, particularly in the audio domain.
Bridging the gap between theoretical concepts and practical, scalable implementation for cutting-edge audio technology.
Technologies & Tools
Featured Projects
Highlighting key achievements and innovative solutions
NewGate Video Processor
A video processing tool that utilizes audio-based triggering and frame manipulation techniques alongside data management algorithms for efficient content processing.
View Project
BatchLMT2
A fine-tuned model based on OpenAI's Whisper for improved speech recognition accuracy, with batch processing capabilities.
View Project
SpaceEchoVST3Plugin
A professional-grade VST3 audio plugin that recreates classic space echo effects with modern digital signal processing techniques.
View Project
Project Celcius
A brutalist puzzle game inspired by NaissanceE, featuring level design and Wwise audio implementation with VR support via Unreal Engine.
Unreleased
Motion Extraction Posy
A tool for extracting and analyzing motion data from video for applications in animation, gaming, and movement research.
View Project
Synthetic-Audio-Detection
A comprehensive system for detecting AI-generated audio content using multi-head classification techniques and advanced neural networks.
View ProjectCase Studies
Detailed exploration of selected projects showcasing approach and outcomes
EUCLID AI System
Multi-head binary classification for synthetic audio detection
Challenge
Develop a state-of-the-art system capable of detecting AI-generated audio with high accuracy across multiple models and generation techniques.
Approach
Designed a novel multi-head classification architecture that analyzes audio across different frequency bands and temporal segments using specialized neural network heads. Each segment undergoes feature extraction through a custom pipeline that combines traditional DSP techniques with deep learning embeddings.
Technologies
Results
- 98.6% accuracy on detecting state-of-the-art synthetic voices
- Successfully identified manipulation techniques not seen during training
- System implemented in production environments for content authentication
Game Credits
Professional game development projects I've contributed to
Jurassic World Evolution 2
Frontier Developments
Responsibilities
Technical audio implementation, Wwise optimization, audio rigging across multiple platforms, and collaboration with sound designers and programmers.
Project Celsius
Dark Matter Games
Responsibilities
Led the audio design team, created and implemented game audio systems, supervised sound effect creation, and managed voice-over recording sessions.
Research Papers
Published academic and technical papers on audio technologies, AI systems, and signal processing
Multi-Head Binary Classification System for Synthetic Data Detection
This whitepaper describes EUCLID (Enhanced Utility for Classification and Identification of Data), an innovative multi-head binary classification architecture that distinguishes between authentic and synthetic audio data.
Read PaperEffects of Lossy Compression Codecs on Perception of Immersive Audio
This research investigates if Vorbis lossy compression on ambisonic audio affects directional accuracy when binauralized within Audiokinetic Wwise and auditioned through Unreal Engine 4.
Read PaperEnhancing Speech Recognition in Panopto Videos
A comparative analysis of OpenAI's Whisper and traditional automatic speech recognition systems for improving transcription accuracy and accessibility in educational videos.
Read PaperVideo Portfolio
Selected video demonstrations of various projects and technologies I've worked on.
Ocean Sim - Short Demo (Wwise, Unreal Engine)
Ocean simulation with custom audio implementation using Wwise and Unreal Engine.
Watch on YouTubeDesert Shooter Vehicle Demo Work In Progress
Vehicle audio implementation work in progress using Wwise and Unreal Engine.
Watch on YouTubeC++ Audio Programming - My Space Echo VST
Code review and demonstration of a custom Space Echo VST plugin implemented in C++.
Watch on YouTubeProject Celcius - Initial Proof of Concept
Early proof of concept for Project Celsius, demonstrating interactive audio implementation.
Watch on YouTubeVR Ambisonic Testing Demo
Testing of ambisonic audio implementation for virtual reality environments.
Watch on YouTubeGet In Touch
Interested in collaboration or have questions about my research papers? Let's connect!
Contact Information
Connect on Social Media
Find me online
Follow me for updates on my latest work and research in AI audio systems and technology.