Anshuman Ganguly

Parallel feedback Active Noise Control of MRI acoustic noise with signal decomposition using hybrid RLS-NLMS adaptive algorithms

This paper presents a cost-effective adaptive feedback Active Noise Control (FANC) method for controlling functional Magnetic Resonance Imaging (fMRI) acoustic noise by decomposing it into dominant periodic components and residual random components. Periodicity of fMRI acoustic noise is exploited by using linear prediction (LP) filtering to achieve signal decomposition. A hybrid combination of adaptive filters-Recursive Least Squares (RLS) and Normalized Least Mean Squares (NLMS) are then used to effectively control each component separately. Performance of the proposed FANC system is analyzed and Noise attenuation levels (NAL) up to 32.27dB obtained by simulation are presented which confirm the effectiveness of the proposed FANC method.

Real-time active noise control of multi-tones and MRI acoustic noise in fMRI bore with signal decomposition and parallel hybrid RLS-NLMS adaptive algorithms

This paper presents a real-time implementation of a cost-effective adaptive feedback Active Noise Control (FANC) method for attenuating acoustic multi-tone noise and functional Magnetic Resonance Imaging (fMRI) acoustic noise in a fMRI bore test-bed. Periodic property of the signal is used to decompose it into dominant periodic components and residual random components using linear prediction (LP) filtering. After decomposition, a hybrid combination of Recursive Least Squares (RLS) and Normalized Least Mean Squares (NLMS) filters is used to effectively attenuate each of the periodic and random components of noise separately. Real time implementation of proposed FANC method on fMRI test bed is discussed and Noise attenuation levels (NAL) obtained are presented which support the effectiveness of the FANC method in practice.

Exploring feedback active noise control with ambisonics

Excessive undesired acoustic noise interferes with normal daily human activities and hence, it needs to be controlled and alleviated. With efficient and powerful computational capability of digital technology, recent adaptive active noise control (ANC) research has seen its way into consumer electronics, automobiles and recreational applications. Unlike Feedforward ANC methods, Feedback ANC methods require no direct measurement of the source noise making them more cost-effective and more suitable for use in large 3-dimensional acoustic environments with multiple noise sources. In this paper, we propose the use of spatial advances made in Higher Order Ambisonics (HOA) technology to attenuate localized noise sources and extend the quiet zone in a large area using multi-channel Feedback ANC architecture. We show that the use of Ambisonics soundfield microphones improves updating of the adaptive algorithms for parallel processing and real-time implementation in the proposed ANC system. The proposed ANC system can be used to control the noise in large acoustic enclosures such as museums and concert halls.

Non-Uniform Microphone Arrays for Robust Speech Source Localization for Smartphone-Assisted Hearing Aid Devices

Robust speech source localization (SSL) is an important component of the speech processing pipeline for hearing aid devices (HADs). SSL via time direction of arrival (TDOA) estimation has been known to improve performance of HADs in noisy environments, thereby providing better listening experience for hearing aid users. Smartphones now possess the capability to connect to the HADs through wired or wireless channel. In this paper, we present our findings about the non-uniform non-linear microphone array (NUNLA) geometry for improving SSL for HADs using an L-shaped three-element microphone array available on modern smartphones. The proposed method is implemented on a frame-based TDOA estimation algorithm using a modified Dictionary-based singular value decomposition method (SVD) method for localizing single speech sources under very low signal to noise ratios (SNR). Unlike most methods developed for uniform microphone arrays, the proposed method has low spatial aliasing as well as low spatial ambiguity while providing a robust low-error with 360° DOA scanning capability. We present the comparison among different types of microphone arrays, as well as compare their performance using the proposed method.

Improved pre-filtering stages for GCC-based direction of arrival estimation using smartphone

Sound source localization (SSL) is one of the important areas in signal processing especially for hearing aid applications. Having at least two microphones and powerful processing capability makes smartphone as a good sound source locator for people who have hearing impairment. In our previous work, we showed that traditional Generalized Cross Correlation (GCC) method with spatial post-filtering stage would increase the performance of instantaneous estimation of direction of arrival (DOA). In this paper, we combine a spectral pre-filtering stage with our previous work to improve the noise robustness of the earlier method. By combining pre-filtering stage with GCC and post-filtering stage, we obtain robust DOA estimation under various realistic noise types. We experiment with several pre-filtering stages and study their impact on reducing root mean square error (RMSE) and mean absolute error (MAE) for DOA estimation. The real-time implementation of some of the proposed algorithms on smartphone is also pres...

Stereo I/O framework for audio signal processing on android platforms

Smartphones and tablets are attractive cost-effective solutions for implementing multi-channel audio signal processing algorithms. Modern smartphones and tablets have at least two microphones and requisite processing capabilities to perform real-time audio signal processing. However, it is still challenging to integrate the dual-microphone audio capture and the audio processing pipeline in real-time for efficient operation. In this paper, we propose an efficient stereo input/output framework for real-time audio signal processing on Android smartphones/tablets. This framework enables us to perform a large variety of real-time signal-processing applications on Android platform such as Direction of Arrival (DOA) estimation, Audio Compression, and Multi-Channel Speech Enhancement. The implementations and demos of Direction of Arrival (DOA), Audio Compression, and Multi-Channel Speech Enhancement algorithms are also presented to show the effectiveness of proposed framework.

Correction to: Non-Uniform Microphone Arrays for Robust Speech Source Localization for Smartphone-Assisted Hearing Aid Devices

In the original publication of the article, Fig. 1 was incorrect. The correct figure is shown below.

ICA based single microphone Blind Speech Separation technique using non-linear estimation of speech

In this paper, a Blind Speech Separation (BSS) technique is introduced based on Independent Component Analysis (ICA) for underdetermined single microphone case. In general, ICA uses noisy speech from at least two microphones to separate speech and noise. But ICA fails to separate when only one stream of noisy speech is available. We use Log Spectral Magnitude Estimator based on Minimum Mean Square Error (LogMMSE) as a non-linear estimation technique to estimate the speech spectrum, which is used as the other input to ICA, with the noisy speech. The proposed method was tested for machinery, babble and traffic noise types mixed with speech at Signal to Noise Ratios (SNRs) of −5 dB, 0 dB and 5 dB. Objective and subjective results show high quality and intelligibility in the separated speech using the developed method.

Real-time Smartphone implementation of noise-robust Speech source localization algorithm for hearing aid users

Speech source localization has numerous application areas such as hearing aid devices (HAD) and consumer electronics applications. Utilizing the powerful processing hardware of smartphones, we demonstrate that smartphones are capable of instantaneous estimation of sound location. In this paper, we present instantaneous direction of arrival (DOA) by using traditional Generalized Cross Correlation (GCC) followed by a spatial post-filtering stage. A simple voice activity detector (VAD) is used for the post-filtering stage to improve noise robustness in some realistic reverberant noisy environments. Root mean square error (RMSE) is used as an evaluation criterion for the proposed method. Both real recorded data and simulated data under different noise types are used for experiments. A real-time implementation of the method on an Android-based smartphone is also presented.

Parallel feedback architecture for ambisonics based active noise control

Active noise control(ANC) has been widely studied for smaller acoustic cavities, like cars and small enclosures, over past several years. One of the key challenges in implementing ANC systems for larger acoustic environments lies in extending the target “quiet” zone for maximizing the noise cancellation. Traditional ANC systems use two “omnidirectional” microphones to access reference noise signal and “residual” error signal at target zone. The target zone around the error microphone remains to be small. Feedback ANC with Ambisonics is aimed to extend the quiet zone by employing specialized encoding and decoding of the noise signal accessed using a single Soundfield microphone. In this paper, a parallel Feedback ANC architecture is fused with the Ambisonics encoding to improve the noise cancellation performance. The proposed ANC system also features an adaptive signal decomposition technique to effectively handle noise types with dominant predictable components, like machinery noise. Noise attenuation lev...