Accelerating Automatic Differentiation of Direct Form Digital Filters
Fuente:
arXiv
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866912723984449536 |
|---|---|
| author | Yu, Chin-Yun Fazekas, György |
| author_facet | Yu, Chin-Yun Fazekas, György |
| contents | We introduce a general formulation for automatic differentiation through direct form filters, yielding a closed-form backpropagation that includes initial condition gradients. The result is a single expression that can represent both the filter and its gradients computation while supporting parallelism. C++/CUDA implementations in PyTorch achieve at least 1000x speedup over naive Python implementations and consistently run fastest on the GPU. For the low-order filters commonly used in practice, exact time-domain filtering with analytical gradients outperforms the frequency-domain method in terms of speed. The source code is available at https://github.com/yoyolicoris/philtorch. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_14390 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Accelerating Automatic Differentiation of Direct Form Digital Filters Yu, Chin-Yun Fazekas, György Systems and Control Audio and Speech Processing Signal Processing We introduce a general formulation for automatic differentiation through direct form filters, yielding a closed-form backpropagation that includes initial condition gradients. The result is a single expression that can represent both the filter and its gradients computation while supporting parallelism. C++/CUDA implementations in PyTorch achieve at least 1000x speedup over naive Python implementations and consistently run fastest on the GPU. For the low-order filters commonly used in practice, exact time-domain filtering with analytical gradients outperforms the frequency-domain method in terms of speed. The source code is available at https://github.com/yoyolicoris/philtorch. |
| title | Accelerating Automatic Differentiation of Direct Form Digital Filters |
| topic | Systems and Control Audio and Speech Processing Signal Processing |
| url | https://arxiv.org/abs/2511.14390 |