Filters








1,747 Hits in 4.7 sec

Speech Denoising Using Only Single Noisy Audio Samples [article]

Qingchun Li, Jiasong Wu, Yilun Kong, Chunfeng Yang, Youyong Kong, Guanyu Yang, Lotfi Senhadji, Huazhong Shu
2021
In this paper, we propose a novel Single Noisy Audio De-noising Framework (SNA-DF) for speech denoising using only single noisy audio samples, which overcomes the limi-tation of constructing either noisy-clean  ...  Experimental results show that the proposed SNA-DF not only eliminates the high dependence on clean targets of traditional audio denoising methods, but also outperforms the methods using multiple noisy  ...  Is the performance of speech denoising using only single noisy audio samples better than other speech denoising methods?  ... 
doi:10.48550/arxiv.2111.00242 fatcat:jauecqipjbh2hetpxk5vzgvv2m

N-HANS: Introducing the Augsburg Neuro-Holistic Audio-eNhancement System [article]

Shuo Liu and Gil Keren and Björn Schuller
2019 arXiv   pre-print
N-HANS is a Python toolkit for in-the-wild audio enhancement, including speech, music, and general audio denoising, separation, and selective noise or source suppression.  ...  Experimental results indicate that N-HANS achieves outstanding performance, and ensure its reliable usage in real-life audio and speech-related tasks, reaching very high audio and speech quality.  ...  Samples from Librespeech and AudioSet are mixed to create noisy speech samples.  ... 
arXiv:1911.07062v2 fatcat:obi67dqqrrd2piaoszyw3eufyu

Joint-Modal Label Denoising for Weakly-Supervised Audio-Visual Video Parsing [article]

Haoyue Cheng, Zhaoyang Liu, Hang Zhou, Chen Qian, Wayne Wu, Limin Wang
2022 arXiv   pre-print
Specifically, we sort the losses of all instances within a mini-batch individually in each modality, then select noisy samples according to relationships between intra-modal and inter-modal losses.  ...  noisy labels dynamically.  ...  Single-modal label denoising vs. joint-modal label denoising. "Audio only" or "Visual only" denotes that label denoising is performed only for audio or visual track.  ... 
arXiv:2204.11573v2 fatcat:vxvw2bpb25g2tetzv5zxfdh4c4

Improving Deep Speech Denoising by Noisy2Noisy Signal Mapping [article]

Nasim Alamdari, Arian Azarang, Nasser Kehtarnavaz
2020 arXiv   pre-print
A fully convolutional neural network is trained by using two noisy realizations of the same speech signal, one used as the input and the other as the output of the network.  ...  This paper presents a deep learning-based approach to improve speech denoising in real-world audio environments by not requiring the availability of clean speech signals in a self-supervised manner.  ...  During the actual operation or testing of the developed deep speech denoising approach, only a single channel is used to feed noisy speech signal frames into the trained deep neural network with the output  ... 
arXiv:1904.12069v2 fatcat:67qqj3aikrgvxing7l3pjxjmzy

Handling Background Noise in Neural Speech Generation [article]

Tom Denton, Alejandro Luebs, Felicia S. C. Lim, Andrew Storus, Hengchin Yeh, W. Bastiaan Kleijn, Jan Skoglund
2021 arXiv   pre-print
Placing a denoising preprocessing stage when extracting features and target clean speech during training is shown to be the best performing strategy.  ...  However, the performance of such models drops when the input is not clean speech, e.g., in the presence of background noise, preventing its use in practical applications.  ...  Having access to only a single melspectrum frame of lookahead likely makes it difficult to determine whether a noisy frame is an actual speech sound or a transient background noise.  ... 
arXiv:2102.11906v1 fatcat:gzl6ro53rbb23nxvi5keo3en7y

Noisy-to-Noisy Voice Conversion Framework with Denoising Model [article]

Chao Xie, Yi-Chiao Wu, Patrick Lumban Tobing, Wen-Chin Huang, Tomoki Toda
2021 arXiv   pre-print
In this paper, to explore VC with the flexibility of handling background sounds, we propose a noisy-to-noisy (N2N) VC framework composed of a denoising module and a VC module.  ...  Leveraging crowd-sourced speech data in training is more economical.  ...  This resulted in a total number of 204 audio samples: 48 audio samples per system and 12 samples from noisy ground-truth target speech.  ... 
arXiv:2109.10608v1 fatcat:3sgwff62dvbz7nc5pckj5bjaku

Example-based cross-modal denoising

D. Segev, Y. Y. Schechner, M. Elad
2012 2012 IEEE Conference on Computer Vision and Pattern Recognition  
Can vision assist in denoising another modality? As a case study, we demonstrate this principle by using video to denoise audio.  ...  In testing, cross-modal input segments having noisy audio rely on the examples for denoising. The video channel drives the search for relevant training examples.  ...  Acknowledgments We thank Israel Cohen and Lihu Berman for useful discussions, Marina Alterman for playing the xylophone and Sharon Gannot for enabling us to use his lab.  ... 
doi:10.1109/cvpr.2012.6247712 dblp:conf/cvpr/SegevSE12 fatcat:bvy7asovefhvxkg5xxov46aema

A Wavenet for Speech Denoising [article]

Dario Rethage and Jordi Pons and Xavier Serra
2018 arXiv   pre-print
Specifically, the model makes use of non-causal, dilated convolutions and predicts target fields instead of a single target sample.  ...  In order to overcome this limitation, we propose an end-to-end learning method for speech denoising based on Wavenet.  ...  In section 4.3 we experiment with 10% and 20% noise-only training samples. Denoising step The network is presented with a noisy speech fragment and the condition value is set to zero.  ... 
arXiv:1706.07162v3 fatcat:vdkca5b4ifaidfl3ltrlbisb7m

Speech Denoising by Accumulating Per-Frequency Modeling Fluctuations [article]

Michael Michelashvili, Lior Wolf
2020 arXiv   pre-print
Our code and samples are available at github.com/mosheman5/DNP and as supplementary. Index Terms: Audio denoising; Unsupervised learning  ...  Given a noisy audio clip, the method trains a deep neural network to fit this signal.  ...  Supervised Noise estimation algorithms Supervised speech denoising algorithms observe, during training, both the noisy sample and the underlying clean samples and learn to map from noisy samples to clean  ... 
arXiv:1904.07612v3 fatcat:7xx32n6enzf2neomnskm633rre

DESIGN AND IMPLEMENTATION OF DIFFERENT AUDIO RESTORATION TECHNIQUES FOR AUDIO DENOISING APPLICATIONS

Merin K Mathai .
2015 International Journal of Research in Engineering and Technology  
Broadband denoising is done by using spectral subtraction and Click removal is done by using an adaptive filter method as the first step.  ...  Localized distortion includes clipping and clicks where only certain samples are affected and globalized distortions include broadband noise where complete bandwidth is consumed by noise.  ...  Block diagram of the designed broad denoiser is shown in Fig-5 Fig-5: Broadband denoiser Here, speech signals are only considered these are free from musical noise.  ... 
doi:10.15623/ijret.2015.0410014 fatcat:wj4ttvpk5be5lhkamjxqwdxeki

SEANet: A Multi-Modal Speech Enhancement Network

Marco Tagliasacchi, Yunpeng Li, Karolis Misiunas, Dominik Roblek
2020 Interspeech 2020  
We explore the possibility of leveraging accelerometer data to perform speech enhancement in very noisy conditions.  ...  A sample of the output produced by our model is available at https://google-research.github.io/seanet/multimodal/speech.  ...  In Figure 1 we illustrate the case in which a single accelerometer axis is used. The generator produces as output a single-channel waveformxm, which represents the denoised speech.  ... 
doi:10.21437/interspeech.2020-1563 dblp:conf/interspeech/TagliasacchiLMR20 fatcat:c5zreb7f4jewdpmnh3iw7rihk4

SEANet: A Multi-modal Speech Enhancement Network [article]

Marco Tagliasacchi, Yunpeng Li, Karolis Misiunas, Dominik Roblek
2020 arXiv   pre-print
We explore the possibility of leveraging accelerometer data to perform speech enhancement in very noisy conditions.  ...  A sample of the output produced by our model is available at https://google-research.github.io/seanet/multimodal/speech.  ...  In Figure 1 we illustrate the case in which a single accelerometer axis is used. The generator produces as output a single-channel waveformxm, which represents the denoised speech.  ... 
arXiv:2009.02095v2 fatcat:ipmezvzapvf7jbvg4ttmjyhq24

SoundStream: An End-to-End Neural Audio Codec [article]

Neil Zeghidour, Alejandro Luebs, Ahmed Omran, Jan Skoglund, Marco Tagliasacchi
2021 arXiv   pre-print
In subjective evaluations using audio at 24kHz sampling rate, SoundStream at 3kbps outperforms Opus at 12kbps and approaches EVS at 9.6kbps.  ...  We present SoundStream, a novel neural audio codec that can efficiently compress speech, music and general audio at bitrates normally targeted by speech-tailored codecs.  ...  For evaluation we use 1000 samples of noisy speech, generated as described in Section IV-A and compute ViSQOL scores when denoising is enabled or disabled, using clean speech references as targets.  ... 
arXiv:2107.03312v1 fatcat:i3hwsfqtgjgbxb7bcwfb4xouoy

N-HANS: A neural network-based toolkit for in-the-wild audio enhancement

Shuo Liu, Gil Keren, Emilia Parada-Cabaleiro, Björn Schuller
2021 Multimedia tools and applications  
In this regard, we present (the Neuro-Holistic Audio-eNhancement System), a Python toolkit for in-the-wild audio enhancement that includes functionalities for audio denoising, source separation, and —for  ...  This is achieved by the use of two identical neural networks comprised of stacks of residual blocks, each conditioned on additional speech- and noise-based recordings through auxiliary sub-networks.  ...  denoising, and speech source separation, we depict its processing procedure for some noisy speech samples in Figs. 4, 5, and 6, respectively.  ... 
doi:10.1007/s11042-021-11080-y fatcat:gmr6gngvjjaxtdstfihuritiea

A Research on Different Filtering Techniques and Neural Networks Methods for Denoising Speech

2019 VOLUME-8 ISSUE-10, AUGUST 2019, REGULAR ISSUE  
The analysis is done by evaluating the performance of different denoising techniques for different types of speech samples.  ...  This paper intends to provide the best suited noise removal technique for de-noising and retrieving clean speech from a noisy speech signal.  ...  Denoised Speech Signal The adaptive filter using LMS algorithm cancels the noise from a noisy speech signal.  ... 
doi:10.35940/ijitee.i1107.0789s219 fatcat:uw32s2rgq5fo5f5twgfa2jzbyq
« Previous Showing results 1 — 15 out of 1,747 results