Digital modelling of guitar audio effects

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Elektrik ve Elektronik Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2015

Öğrenci: ENGİN ZEKİ

Danışman: TOLGA ÇİLOĞLU

Özet:

Digital audio effects are used by many electric guitar players. These effects help players to find their desired tones and sounds. Digital audio workstations (DAW) and Virtual studio technologies (VST) advance the development of guitar audio effects softwares. The aim of this thesis is to implement main linear guitar effects (delay, reverb, wah-wah, flanger) real-time and enhance the main nonlinear guitar effects (distortion, overdrive) using system identification methods. After the verification of the effect models with MATLAB, real-time implementations of these effects will be done using C/C++/C# software languages. For the modelling of main nonlinear guitar effects, distortion and overdrive, this thesis investigates current methods of static modelling and dynamic nonlinear state space solutions. After discussion of previous models, this thesis introduces a new method of distortion modelling with system identification called Enhanced Modelling of Guitar Distortion. Enhanced Modelling of Guitar Distortion algorithm will use neural network system identification method ANFIS (Adaptive-Network-Based Fuzzy Inference System). ANFIS is used as a system identification tool in Enhanced Modelling of Guitar Distortion algorithm. This algorithm takes the guitar output signal and pre-amplifies the input with 12AX7 vacuum tube amplifier simulation model to obtain clean channel. ANFIS system identification block is trained using desired distortion effect input output pair. This training and learning results into a ANFIS (Adaptive-Network-Based Fuzzy Inference System) structure that can be used for processing future inputs. Using clean channel output as an input to the ANFIS structure, lead channel output is obtained. Real-time implementations of distortion and overdrive will be done using C/C++/C# software languages. The results of the Enhanced Modelling of Guitar Distortion algorithm will show if it is applicable to model distortion and overdrive effects using system identification approach. Listening tests will be done to provide evidence that the simulation models are successful. Strict real-time constraints are going to be satisfied and the simulation of a guitar amplifier will be presented.