• The Journal of the Acoustical Society of Korea
• /
• v.19 no.4
• /
• pp.52-57
• /
• 2000

In this paper, we designed the 16 Kbps speech coder that has compatibility with CS-ACELP algorithm(G.729). The speech signal is sampled at rate of 16 KHz, divided into two narrowband signal by QMF filterbank, and decimated to rate of 8 KHz. The lower-band signal is encoded by CS-ACELP and the upper-band signal is encoded by Adaptive Transform Coding(ATC) algorithm. At the receiver, two band signals are synthesized by decoder of CS-ACELP and ATC, respectively. The reconstructed output is obtained by passing the QMF synthesis bank. The proposed wideband coder is evaluated with ITU-T G.722 coder through the Mean Opinion Score(MOS) test.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.3 no.3
• /
• pp.607-616
• /
• 1999

This paper proposed an improved 4.8kbps VSELP in order to keep the good quality in band-limited channel. In the most cases, it is difficult to keep the good quality at the low bit rate. In order to solve the problems, many methods are proposed, but they are not suitable to the Korean language structure because they are designed for being suitable to the foreign language structure. In experiment, we use the noseless Korean voice data. We show that the proposed 4.8kbps VSELP is not excellent to the 8kbps VSELP in SEGWSNR(Segmentally Weighted SNR), but it is the superior to the 8kbps VSELP in the MOS(Mean Opinion Score) test.

• 한국정보통신설비학회:학술대회논문집
• /
• 2005.08a
• /
• pp.329-332
• /
• 2005

In this paper, we describe implementation of personal media gateway (PMG) for applications over BcN networks. PMG is a TV based set-top terminal, which enables transmission of Full D1 high quality video and audio at the speed of maximum 2Mbps. It supports SIP protocol and QoS for the BcN networks. The hardware of the PMG consists of host module, audio/video codec processing module, DTMF module, and remote control I/O module. H.263 and MPEG4 software are implemented in DSP as codec for hi-directional communication and streaming, respectively. G.711 and Ogg-Vorbis are implemented as audio codec. We examined the quality of video using the Video Quality Test Equpment, which was developed by KT Convergence Lab. The experimental results show the video quality of MOS 4.1 and audio quality of MOS 4.3. We expect that PMG will be prospective business models, and create new customer value.

• ETRI Journal
• /
• v.46 no.3
• /
• pp.550-563
• /
• 2024

A novel inductance simulation circuit employing only two dual-output voltage-differencing buffered amplifiers (DO-VDBAs) and a single capacitance (grounded) is proposed in this paper. The reported configuration is a purely resistor-less realization that provides electronically controllable realized inductance through biasing quantities of DO-VDBAs and does not rely on any constraints related to matched values of parameters. This structure exhibits excellent behavior under the influence of tracking errors in DO-VDBAs and does not exhibit instability at high frequencies. The simple and compact metal-oxide semiconductor (MOS) implementation of the DO-VDBAs (eight MOS per DO-VDBA) and adoption of grounded capacitance make the proposed circuit suitable for on-chip realization from the perspective of chip area consumption. The function of the pure grounded inductance is validated through high pass/bandpass filtering applications. To test the proposed design, simulations were performed in the PSPICE environment. Experimental validation was also conducted using the integrated circuit CA3080 and operational amplifier LF-356.

• The Journal of the Acoustical Society of Korea
• /
• v.27 no.2
• /
• pp.75-79
• /
• 2008

In this paper, we propose a novel speech enhancement technique using global soft decision (GSD) based on the probabilistic outputs of support vector machine (SVM). Generally, speech enhancement algorithms applied soft decision gain modification and noise power estimation have bettor performance than those employing hard decision. Especially, global speech absence probability (GSAP), which is known as an effective measure of the speech absence in each frame, has been adopted to SD-based speech enhancement methods. For this reason, we introduce a new GSAP estimated from the probabilistic output of SVM using sigmoid function. The performance of the proposed algorithm is evaluated by the PESQ and MOS test under various noise environments and yields better results compared with the conventional GSD scheme.

• Speech Sciences
• /
• v.10 no.2
• /
• pp.77-84
• /
• 2003

In this research, we seek the beginning of the speech and detect the stationary speech region using lip information. Performing running average of the estimated speech signal in the stationary region, we reduce the effect of musical noise which is inherent to the conventional MlMSE (Minimum Mean Square Error) speech enhancement algorithm. In addition to it, SFM (Spectral Flatness Measure) is incorporated to reduce the speech signal estimation error due to speaking habit and some lacking lip information. The proposed algorithm with Wiener filtering shows the superior performance to the conventional methods according to MOS (Mean Opinion Score) test.

• Proceedings of the IEEK Conference
• /
• 2000.09a
• /
• pp.819-822
• /
• 2000

최근 유아기 아이들을 위한 교육매체의 발달로 각종 CD-ROM이나 테이프 등에서 성우의 목소리로 동화를 읽어주는 시스템이 많이 나와 있고, 또한 Web Book이 점차적으로 보편화가 되 가고 있다. 하지만 이런 획일적이고 균일화된 매체들은 아이들에게 금새 실증을 내게 하기 때문에 흥미 유발을 위해 동화의 주인공을 자기자신이나 친근한 사람의 이름 등으로 바꾸어 발성해 준다면 더욱 친근감 있게 받아들일 것이다. 본 논문에서는 기존의 성우가 발성하는 동화의 주인공 이름을 Test화자가 새로운 이름으로 발성을 해주면 기존 성우의 목소리패턴으로 바꾸어 동화를 읽어주는 시스템에 대해서 제안하고자 한다. 우선 Test화자가 발성한 목소리를 성우의 목소리로 바꾸어 주기 위해서 기존의 성우가 발성한 동화주인공 이름과 Test화자가 발성한 이름과의 운율패턴을 비교하여 성우의 운율패턴에 일치시키고 성우의 목소리 패턴으로 변경된 새로운 주인공의 이름만을 기존의 동화 DB에 삽입하였다. 또한 에너지 패턴조절은 기존의 성우가 발성한 기준패턴에 근사화 시켰고 끝점을 스므딩 시킴으로써 자연스런 발성이 되게 만들어주었다. 결과적으로 Mos Score가 3.873로 비교적 좋은 결과를 얻을 수 있었다.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.2
• /
• pp.112-118
• /
• 2009

This paper proposes a rapid detection algorithm of a salient region for music video browsing system, which can be applied to mobile device and digital video recorder (DVR). The input music video is decomposed into the music and video tracks. For the music track, the music highlight including musical chorus is detected based on structure analysis using energy-based peak position detection. Using the emotional models generated by SVM-AdaBoost learning algorithm, the music signal of the music videos is classified into one of the predefined emotional classes of the music automatically. For the video track, the face scene including the singer or actor/actress is detected based on a boosted cascade of simple features. Finally, the salient region is generated based on the alignment of boundaries of the music highlight and the visual face scene. First, the users select their favorite music videos from various music videos in the mobile devices or DVR with the information of a music video's emotion and thereafter they can browse the salient region with a length of 30-seconds using the proposed algorithm quickly. A mean opinion score (MOS) test with a database of 200 music videos is conducted to compare the detected salient region with the predefined manual part. The MOS test results show that the detected salient region using the proposed method performed much better than the predefined manual part without audiovisual processing.

• IE interfaces
• /
• v.12 no.4
• /
• pp.496-505
• /
• 1999

Quality of Service(QoS) is the collective effect of service performances and has a direct impact on customer satisfaction. Although QoS is subjective, network performance parameters contributing to QoS can be measured physically. Therefore overall customer satisfaction for each test condition of the performance parameters is evaluated by asking respondents to indicate his or her opinion on a five-category rating scale i.e., excellent, good, fair, poor, and unsatisfactory. The opinion data resulting from the test can then be used to measure and analyze QoS from the customers' viewpoints. In this papaer, we consider two methods for analyzing the opinion data: MOS method and Cumulative Probability Curve method. The former evaluates an arithmetic mean of the opinion scores which quantify the surveyed opinions of respondents. The latter uses graphical and analytical models which are based on the distribution of the opinions rather than an arithmetic mean. The advantages, disadvantages, and an alternative of each method are discussed, together with future directions of research.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.21 no.12
• /
• pp.3246-3255
• /
• 1996

A threshold voltage extraction scheme which does not need matched replica of the MOSFET under test is proposed. In contrast to alternative methods, the accuracy of the proposed scheme does not depend on the matching of the test transistors. The proposed scheme has been implemented in a matching-free way using a switched-capacitor subtracting ampliier and a dynmic current mirror. Nonideal effects associated with these circuits, such as non-zero offset voltages and finite gains of op-amps, capcitor mismateches, and charge injection of MOS switches, are investigated and compensated. The circuit has been designed using ISRC 1.5.mu.m CMOS process parameters andfabricated at Inter-University semiconductor Research Center, and its performance has been evaluated.