Henrik Åkesson

A mixed analog-digital hybrid for speech enhancement purposes

The paper presents and evaluates a hybrid implementation of a low complexity algorithm for speech enhancement, the adaptive gain equalizer (AGE). The AGE is a subband based time domain method for instantaneous boosting of speech. By a combination of digital analysis and analog synthesis, the main advantages of the digital and analog domains are utilized. The hybrid solution is implemented on a printed circuit board and evaluated in real-time. The development time of the proposed solution was very short, the solution is flexible, robust, has high signal bandwidth in the signal chain and does not require a voice activity detector (VAD). Furthermore, the solution is not restricted by quantization errors in the signal chain and does not require a high speed digital signal processor (DSP) for analysis. Informal listening tests and signal-to-noise ratio (SNR) measures verify excellent speech enhancement performance and quality.

Analog circuit implementation for speech enhancement purposes

Human speech is the main method for personal communication. However, interfering noise could degrade the intelligibility of speech, eventually resulting in errors. Thus, efficient speech enhancement algorithms are needed for example in hand held battery powered hearing aids. This paper presents an implementation of a time domain method for speech enhancement purposes: the adaptive gain equalizer. The implementation is carried out on a printed circuit board using common analog electronic components, and evaluated in real-time. The proposed solution benefits from high system bandwidth, it neither quantizes nor digitalizes data, and it is likely to have more efficient power consumption as opposed to many digital signal processor (DSP) based solutions. The evaluation proves the speech enhancement performance of the analog circuit implementation.

Analog versus Digital Control of Boring Bar Vibration

In workshops where metal cutting is performed, the machining processes frequently introduces productivity degrading vibration problems and annoying sound, sometimes almost at unbearable levels. Besides producing disturbing noise, the vibrations affect the surface finish of the workpiece and the tool life. Two different approaches based on feedback control are investigated, both applied for the control of an active boring bar. The first approach is based on a digital adaptive feedback controller; the feedback filtered-X LMS algorithm. The second approach is based on an analog controller; a feedback controller with gain and phase orthogonally adjustable, thus flexible for the control of systems with different dynamic properties. Based on open loop frequency response function estimates, robustness and stability of both the digital and the analog feedback control system are discussed. A comparison of the two controllers concerning their boring bar vibration attenuation performance shows that the analog controller attenuates the vibration in same order of magnitude as the digital controller which is approximately by up to 40 dB.

Modeling a Clamped Boring Bar using Euler‐Bernoulli Beam Models with Various Boundary Conditions

This paper addresses modeling of a clamped boring bar using Euler‐Bernoulli beam theory. Euler‐Bernoulli beams with a number of different boundary conditions were used to model a clamped boring bar. Estimates of the boring bar’s natural frequencies and mode shapes were produced with each of the boring bar models. The estimates produced by the distributed‐parameter system models are compared with eigenfrequencies and mode shapes estimated based on experimental modal analysis of the actual boring bar clamped in a lathe.