A Systematic Literature Review on Inverse Synthesis and Sound Matching

Gaw˛ecki B, Łukasik E. A Systematic Literature Review on Inverse Synthesis and Sound Matching. In: Convention Paper. Audio Engineering Society; 2026.pp. 10306:1-. Available from: https://aes.org/publications/elibrary-page/?id=23199

@inproceedings{Gawecki2026_23199,
  author = {Gaw˛ecki, Bruno and Łukasik, Ewa},
  title = {{A Systematic Literature Review on Inverse Synthesis and Sound Matching}},
  pages = {10306:1-},
  year = {2026},
  month = may,
  publisher = {Audio Engineering Society},
  url = {}
}

TY  - CPAPER
TI  - A Systematic Literature Review on Inverse Synthesis and Sound Matching
AU  - Gaw˛ecki, Bruno
AU  - Łukasik, Ewa
SP  - 1
M1  - Paper 10306
PY  - 2026
DA  - 2026/05/28
UR  - https://aes.org/publications/elibrary-page/?id=23199
PB  - Audio Engineering Society
LA  - en
AB  - This paper presents a systematic literature review on inverse synthesis and sound matching, which focus on predicting synthesizer parameters to recreate a target audio waveform. Automating this process via machine learning is impeded by non-linear parameter mapping, non-differentiable engines, data scarcity, and metric misalignment. Existing approaches are categorized into non-differentiable synthesizer methods, utilizing evolutionary algorithms and deep learning, incorporating techniques to bypass gradient limitations such as neural proxies or generative models. In contrast, differentiable synthesizer methods, enable the integration of audio loss functions into training pipelines via custom signal processing environments. The analysis identifies a critical reliance on spectral representations for evaluating perceptual similarity, given that parameter based metrics frequently fail to align with human hearing. The findings indicate that while deep learning has reduced inference times, the field lacks a unified production solution. Future progress requires the establishment of standardized benchmarks to evaluate models, the implementation of novel advancements in generative models not yet applied to this problem, and the development of hybrid architectures to simultaneously address these distinct technical challenges.
ER  -