Voice frequency

Abstract: Flexible and skin-attachable vibration sensors have been studied for use as wearable voice-recognition electronics. However, the development of vibration sensors to recognize the human voice accurately with a flat frequency response, a high sensitivity, and a flexible/conformable form factor has proved a major challenge. Here, we present an ultrathin, conformable, and vibration-responsive electronic skin that detects skin acceleration, which is highly and linearly correlated with voice pressure. This device consists of a crosslinked ultrathin polymer film and a hole-patterned diaphragm structure, and senses voices quantitatively with an outstanding sensitivity of 5.5 V Pa−1 over the voice frequency range. Moreover, this ultrathin device (<5 μm) exhibits superior skin conformity, which enables exact voice recognition because it eliminates vibrational distortion on rough and curved skin surfaces. Our device is suitable for several promising voice-recognition applications, such as security authentication, remote control systems and vocal healthcare. Though skin-attachable vibration sensors are promising for voice recognition applications, current technologies do not meet key performance requirements. Here, the authors report a flexible skin-attachable sensor with high sensitivity and flat frequency response over the vocal frequency range.

Abstract: A television telephone system, wherein frequency-modulation with a low modulation index is effected by using a carrier wave of a slightly higher frequency than the maximum frequency of a video signal, and the video signal thus modulated is transmitted through a transmission line. With such system, a wide band video signal and audio signal can be transmitted without cross talk and distortion by using an ordinary telephone cable which is intended to be used only for the transmission of voice frequency signals, and communication can be achieved, with the image of the opposite party or drawing, document or the like being viewed.

Abstract: Departing from past research on managers' responses to employee voice, we propose and examine a nonlinear linkage between promotive/prohibitive voice and managers' evaluations of voicers (i.e., manager-rated voicers' promotability and overall performance). Drawing from social persuasion theory, we theorize that managers tend to give more positive evaluations to employees who engage in a moderate frequency of promotive/prohibitive voice than those who either rarely speak up or speak up very frequently. In Study 1, based on a sample from a Chinese bank, we found that leader-member exchange quality (LMX) moderated the inverted U-shaped linkage of prohibitive voice with manager-rated promotability of voicers, whereas the frequency of promotive voice was not related to promotability, irrespective of levels of LMX. In Study 2, using employee-reported voice frequency, rather than the manager-rated measures adopted in Study 1, we largely replicated the main findings of Study 1 based on a sample from an information technology firm in the United States. In Study 3, using another U.S. sample, from a financial services firm, we found that manager-perceived voice constructiveness mediated the curvilinear interactive effect of prohibitive voice (rather than promotive voice) and LMX on managers' evaluations of employees' overall performance. (PsycINFO Database Record (c) 2018 APA, all rights reserved).

Abstract: Parameter optimization is one of methods to improve accuracy of machine learning algorithms. This study applied the grid search method for tuning parameters in the well-known classification algorithm namely Random Forests. Random Forests was implemented on the voice gender dataset to identify gender based on the human voice’s characteristics. There are two parameters that were tuned to obtain the optimal values. Those parameters are number of variables used in building trees and number of trees that involves in the classifiers. Experimental results on voice gender dataset show that the highest accuracy of Random Forest with parameter tuning is 0.96907 which is higher than the accuracy of the model without parameter tuning (0.9675). The optimal parameter for the best classifier is number of variables is 'sqrt' which is square root of parameters involved in dataset and number of trees is 300. This study shows that the tuning parameter results optimal parameters for developing the best classifier using Random Forests.