• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.7C
• /
• pp.962-970
• /
• 2004

This paper presents an efficient implementation of the 2.4 kbps EHSX(Enhanced Harmonic Stochastic Excitation) speech coder on a TMS320C6701$^{TM}$ floating-point digital signal processor. The EHSX speech codec is based on a harmonic and CELP(Code Excited Linear Prediction) modeling of the excitation signal respectively according to the frame characteristic such as a voiced speech and an unvoiced speech. In this paper, we represent the optimization methods to reduce the complexity for real-time implementation. The complexity in the filtering of a CELP algorithm that is the main part for the EHSX algorithm complexity can be reduced by converting program using floating-point variable to program using fixed-point variable. We also present the efficient optimization methods including the code allocation considering a DSP architecture and the low complexity algorithm of harmonic/pitch search in encoder part. Finally, we obtained the subjective quality of MOS 3.28 from speech quality test using the PESQ(perceptual evaluation of speech quality), ITU-T Recommendation P.862 and could get a goal of realtime operation of the EHSX codec.c.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.2
• /
• pp.413-418
• /
• 2009

In a speech coding system using excitation source of voiced and unvoiced, it would be involved a distortion of speech qualify in case coexist with a voiced and an unvoiced consonants in a frame. So, We proposed TS(Transition Segment) including unvoiced consonant searching and extraction method in order to uncoexistent with a voiced and unvoiced consonants in a frame. This research present a new method of TS approximate-synthesis by using Least Mean Square and frequency band division. As a result, this method obtain a high qualify approximation-synthesis waveforms within TS by using frequency information of 0.547kHz below and 2.813kHz above. The important thing is that the maximum error signal can be made with low distortion approximation-synthesis waveform within TS. This method has the capability of being applied to a new speech coding of Voiced/Silence/TS, speech analysis and speech synthesis.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.6
• /
• pp.670-677
• /
• 2019

As the presence of background noise in acoustic signal degrades the performance of speech or acoustic event recognition, it is still challenging to extract noise-robust acoustic features from noisy signal. In this paper, we propose a combined structure of Wasserstein Generative Adversarial Network (WGAN) and MultiTask AutoEncoder (MTAE) as deep learning architecture that integrates the strength of MTAE and WGAN respectively such that it estimates not only noise but also speech features from noisy acoustic source. The proposed MTAE-WGAN structure is used to estimate speech signal and the residual noise by employing a gradient penalty and a weight initialization method for Leaky Rectified Linear Unit (LReLU) and Parametric ReLU (PReLU). The proposed MTAE-WGAN structure with the adopted gradient penalty loss function enhances the speech features and subsequently achieve substantial Phoneme Error Rate (PER) improvements over the stand-alone Deep Denoising Autoencoder (DDAE), MTAE, Redundant Convolutional Encoder-Decoder (R-CED) and Recurrent MTAE (RMTAE) models for robust speech recognition.

• The Journal of the Acoustical Society of Korea
• /
• v.27 no.7
• /
• pp.379-385
• /
• 2008

In this paper, we propose an effective speech reinforcement technique under the near-end and the far-end noise environments. In general, since the intelligibility of the far-end speech for the near-end listener is significantly reduced under near-end noise environments, we require a far-end speech reinforcement approach to avoid this phenomena. Specifically, based on the estimated background noise spectrum of the near-end, we reinforce the far-end speech spectrum by incorporating the more general cases under the near-end with background noise. Also, we propose the novel approach to reinforce the actual speech signal except for the noise signal in the far-end noisy speech signal. The performance of the proposed algorithm is evaluated by the CCR (Comparison Category Rating) test of the method for subjective determination of transmission quality in ITU-T P.800 under various noise environments and shows better performances compared with the conventional method.

• Proceedings of the IEEK Conference
• /
• 1998.10a
• /
• pp.1247-1250
• /
• 1998

In speech synthesis, LF model is widely used for excitation signal for voice source coding system. But LF model does not represent the harmonic frequencies of excitation signal. We propose an effective method which use sinusoidal functions for representing the harmonics of voice source signal. The proposed method could achieve more exact voice source waveform and better synthesized speech quality than LF model.

• Proceedings of the KIEE Conference
• /
• 1987.07b
• /
• pp.881-883
• /
• 1987

This paper decribes a new technique for the generation of excitation sources in LPC system. We synthesize a speech signal using several excitation sources, according to residual signal energy and ZCR(zero Crossing Rate). One of the excitation sources mix the double differentiated glottal wave form source and noise source. As a result, we got improved speech signal than that produced by conventional LPC system.

• Phonetics and Speech Sciences
• /
• v.6 no.3
• /
• pp.63-72
• /
• 2014

The current study assessed the utility of acoustic analyses the most commonly used in routine clinical voice assessment including perturbation, nonlinear dynamic analysis, and Spectral/Cepstrum analysis based on signal typing of dysphonic voices and investigated their applicability of clinical acoustic analysis methods. A total of 70 dysphonic voice samples were classified with signal typing using narrowband spectrogram. Traditional parameters of %jitter, %shimmer, and signal-to-noise ratio were calculated for the signals using TF32 and correlation dimension(D2) of nonlinear dynamic parameter and spectral/cepstral measures including mean CPP, CPP_sd, CPPf0, CPPf0_sd, L/H ratio, and L/H ratio_sd were also calculated with ADSV(Analysis of Dysphonia in Speech and VoiceTM). Auditory perceptual analysis was performed by two blinded speech-language pathologists with GRBAS. The results showed that nearly periodic Type 1 signals were all functional dysphonia and Type 4 signals were comprised of neurogenic and organic voice disorders. Only Type 1 voice signals were reliable for perturbation analysis in this study. Significant signal typing-related differences were found in all acoustic and auditory-perceptual measures. SNR, CPP, L/H ratio values for Type 4 were significantly lower than those of other voice signals and significant higher %jitter, %shimmer were observed in Type 4 voice signals(p<.001). Additionally, with increase of signal type, D2 values significantly increased and more complex and nonlinear patterns were represented. Nevertheless, voice signals with highly noise component associated with breathiness were not able to obtain D2. In particular, CPP, was highly sensitive with voice quality 'G', 'R', 'B' than any other acoustic measures. Thus, Spectral and cepstral analyses may be applied for more severe dysphonic voices such as Type 4 signals and CPP can be more accurate and predictive acoustic marker in measuring voice quality and severity in dysphonia.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.7
• /
• pp.1571-1576
• /
• 2013

This paper proposes an improvement algorithm of signal-to-noise ratios (SNRs) for speech signals under noisy environments. The proposed algorithm first estimates the SNRs in a low SNR, mid SNR and high SNR areas, in order to improve the SNRs in the speech signal from background noise, such as white noise and car noise. Thereafter, this algorithm subtracts the noise signal from the noisy speech signal at each bands using a spectrum sharpening method. In the experiment, good signal-to-noise ratios (SNR) are obtained for white noise and car noise compared with a conventional spectral subtraction method. From the experiment results, the maximal improvement in the output SNR results was approximately 4.2 dB and 3.7 dB better for white noise and car noise compared with the results of the spectral subtraction method, in the background noisy environment, respectively.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.46 no.5
• /
• pp.8-14
• /
• 2009

In the research field of speech recognition, pinpointing the endpoints of speech utterance even with the presence of background noise is of great importance. These noise present during recording introduce disturbances which complicates matters since what we just want is to get the stationary parameters corresponding to each speech section. One major cause of error in automatic recognition of isolated words is the inaccurate detection of the beginning and end boundaries of the test and reference templates, thus the necessity to find an effective method in removing the unnecessary regions of a speech signal. The conventional methods for speech endpoint detection are based on two linear time-domain measurements: the short-time energy, and short-time zero-crossing rate. They perform well for clean speech but their precision is not guaranteed if there is noise present, since the high energy and zero-crossing rate of the noise is mistaken as a part of the speech uttered. This paper proposes a novel approach in finding an apparent threshold between noise and speech based on Lyapunov Exponents (LEs). This proposed method adopts the nonlinear features to analyze the chaos characteristics of the speech signal instead of depending on the unreliable factor-energy. The excellent performance of this approach compared with the conventional methods lies in the fact that it detects the endpoints as a nonlinearity of speech signal, which we believe is an important characteristic and has been neglected by the conventional methods. The proposed method extracts the features based only on the time-domain waveform of the speech signal illustrating its low complexity. Simulations done showed the effective performance of the Proposed method in a noisy environment with an average recognition rate of up 92.85% for unspecified person.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1984.12a
• /
• pp.66-69
• /
• 1984

This paper is a study on the Endpoint Detection for Korean Speech Recognition. In speech signal process, analysis parameter was classification from Zero Crossing Rate(Z.C.R), Log Energy(L.E), Energy in the predictive error(Ep) and fundamental Korean Speech digits, /영/-/구/ are selected as date for the Recognition of Speech. The main goal of this paper is to develop techniques and system for Speech input ot machine. In order to detect the Endpoint, this paper makes choice of Log Energy(L.E) from various parameters analysis, and the Log Energy is very effective parameter in classifying speech and nonspeech segments. The error rate of 1.43% result from the analysis.