Audio analysis and automatic equalizing processes in the context of mixing preparation

Mixing is a crucial stage in the music production process that demands high levels of concentration and can take a significant amount of time. Preparing a mix project, including tasks such as pre-mixing, can also be tiring. To support faster and more efficient pre-mixing, various tools have been developed. Over time, these tools have transformed into automation systems and AI-powered applications capable of performing many tasks independently. Modern software helps complete mixing tasks quickly, offers solutions to problems, and provides new creative perspectives to users. In recent years, with the involvement of amateurs, musicians, and home studio setups alongside professionals, new trends have emerged in the music industry. The tools developed for music production play an important role in these trends. As part of this thesis study, a software tool with analysis and automatic equalizer matching capabilities was developed and named AVE. The program was created using MathWorks MATLAB and primarily utilizes the Fast Fourier Transform (FFT) method. To perform the necessary operations, original algorithms were designed and implemented. Unlike many existing programs where certain parameters are inaccessible to users, AVE provides these parameters with multiple choice options. To compare the equalizer matching functions of AVE with similar softwares, various acoustic instruments were recorded, and an artificial sine wave test signal was generated. Additionally, an equalizer was applied to an acoustic guitar recording to produce a new audio sample, thereby increasing the number of test samples. In the experiments, certain audio samples were selected as references, and equalizer matching was applied to the other samples. The goal was to make the processed samples resemble the reference ones in terms of frequency response and timbre. The results obtained from AVE and other software tools were measured and compared using AVE. Additionally, the accuracy of the key findings and AVE's analytical capabilities were validated by comparing them with results from FabFilter Pro-Q 4, which performed successfully in the tests. AVE's equalizer matching performance was compared step-by-step with eight different software programs. These programs include: iZotope Neutron 5, iZotope Ozone 11, FabFilter Pro-Q 4, Sonible smart:EQ 4, IK Multimedia T-RackS 6 module Master Match X, MeldaProduction MAutoEqualizer, MeldaProduction MFreeformEqualizer, and MeldaProduction MMatcher. According to the results obtained, the AVE tool demonstrated superior performance compared to the other programs in all tests, which is a noteworthy outcome. Within the scope of the thesis, the software's structure, underlying principles, and functions were explained, and its limitations and potential were discussed alongside the experimental findings. The capabilities of the program suggest that AVE can be used not only in the pre-mixing stage, but also during recording, mixing, and mastering processes. Additionally, its open-source nature and the availability of numerous user-selectable parameters make it suitable for educational purposes and further development, highlighting some of the tool's key potential outcomes.