Schröter H, Rosenkranz T, Escalante Banuelos A, Maier A (2020)
Publication Language: English
Publication Status: Published
Publication Type: Other publication type
Future Publication Type: Other publication type
Publication year: 2020
URI: https://github.com/Rikorose/clc-dns-challenge-2020
Open Access Link: https://arxiv.org/abs/2006.13077
Complex-valued processing brought deep learning-based speech enhancement and
signal extraction to a new level.
Typically, the noise reduction process is based on a time-frequency (TF) mask
which is applied to a noisy spectrogram. Complex masks (CM) usually outperform
real-valued masks due to their ability to modify the phase. Recent work proposed to use a complex linear combination of coefficients
called complex linear coding (CLC) instead of a point-wise multiplication with
a mask.
This allows to incorporate information from previous and optionally future
time steps which results in superior performance over mask-based enhancement
for certain noise conditions.
In fact, the linear combination enables to model quasi-steady properties like
the spectrum within a frequency band.
In this work, we apply CLC to the Deep Noise Suppression (DNS) challenge and
propose CLC as an alternative to traditional mask-based processing, e.g. used
by the baseline.
We evaluated our models using the provided test set and an additional validation set with real-world stationary and non-stationary noises. Based on the published test set, we outperform the baseline w.r.t. the scale independent signal distortion ratio (SI-SDR) by about 3dB.
APA:
Schröter, H., Rosenkranz, T., Escalante Banuelos, A., & Maier, A. (2020). CLC: Complex Linear Coding for the DNS 2020 Challenge.
MLA:
Schröter, Hendrik, et al. CLC: Complex Linear Coding for the DNS 2020 Challenge. 2020.
BibTeX: Download