Skip to content

Academic projects

A list of my academic projects, papers, talks, etc. For an overview of the different projects I worked on throughout my Master's in Music Technology, visit

Master's Thesis

My major project in grad school was titled "Music demixing with the sliCQ transform," which led to a publication and the thesis I submitted for my Master of Arts degree in 2022 from the McGill Music Technology department.

It attempts to replace the STFT with the sliCQT, a nonuniform time-frequency transform, in Open-Unmix, a PyTorch AI model (deep learning neural network) for music source separation.

I worked on it during the AICrowd Music Demixing Challenge 2021 (MDX21). The code for the project and related files are on GitHub:,

Papers, posters, and publications

Cadenza Challenge Workshop 2023,

    title={Better music demixing with the sliCQ transform},
    author={Hanssian, Sevag},
    booktitle={1st Cadenza Workshop on Machine Learning Challenges to Improve Music for People with Hearing Loss},

Master's Thesis @ McGill University 2022,

  title={Music demixing with the sliced constant-Q transform},
  author={Hanssian, Sevag},
  school={McGill University},

MDX21 workshop @ ISMIR 2021,

    title={Music demixing with the sliCQ transform},
    author={Hanssian, Sevag},
    booktitle={MDX21 workshop at ISMIR 2021},

    title={Music demixing with the sliCQ transform},
    author={Hanssian, Sevag},
    journal={arXiv preprint arXiv:2112.05509},

Poster at ISMIR 2021 (MDX workshop):

Poster at CIRMMT The Future of Sound 2021:

Videos, talks, and presentations

15 min talk at the Cadenza Challenge 2023 Workshop about the updated xumx-sliCQ:

30 min talk at the MILA Music AI reading group about xumx-sliCQ:

Short 5-minute talk at the MDX21 town hall about xumx-sliCQ: