• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06a
• /
• pp.963-970
• /
• 1994

In speech signal processing, it is very important to detect the pitch exactly. The algorithms for pitch extraction that have been proposed until now are not enough to detect the fine pitch in speech signal. Thus we propose the new algorithm which takes advantage of the G-peak extraction. It is the method to find MZCI(maximum zer-crossing interval) which is defined as cut-off bandwidth rate of LPF (low pass filter)and detect the pitch period of the voiced signals. This algorithm performs robustly with a gross error rate of 3.63% even in 0 dB SNR environment. The gross error rate for clean speech is only 0.18%. Also it is able to process all course with speed.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.5
• /
• pp.683-687
• /
• 2010

This paper studied some methods which use frequency warping method that is the one of the speech transformation method to develope the robust speech recognition system for the emotional variation. For this purpose, the effect of emotional variations on the speech signal were studied using speech database containing various emotions and it is observed that speech spectrum is affected by the emotional variation and this effect is one of the reasons that makes the performance of the speech recognition system worse. In this paper, new training method that uses frequency warping in training process is presented to reduce the effect of emotional variation and the speech recognition system based on vocal tract length normalization method is developed to be compared with proposed system. Experimental results from the isolated word recognition using HMM showed that new training method reduced the error rate of the conventional recognition system using speech signal containing various emotions.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.4 no.2
• /
• pp.149-154
• /
• 2004

In this research, we selected the speech recognition to implement the electric wheelchair system as a method to control it by only using the speech and used DTW (Dynamic Time Warping), which is speaker-dependent and has a relatively high recognition rate among the speech recognitions. However, it has to have small memory and fast process speed performance under consideration of real-time. Thus, we introduced VQ (Vector Quantization) which is widely used as a compression algorithm of speaker-independent recognition, to secure fast recognition and small memory. However, we found that the recognition rate decreased after using VQ. To improve the recognition rate, we applied ART2 (Adaptive Reason Theory 2) algorithm as a post-process algorithm to obtain about 5% recognition rate improvement. To utilize ART2, we have to apply an error range. In case that the subtraction of the first distance from the second distance for each distance obtained to apply DTW is 20 or more, the error range is applied. Likewise, ART2 was applied and we could obtain fast process and high recognition rate. Moreover, since this system is a moving object, the system should be implemented as an embedded one. Thus, we selected TMS320C32 chip, which can process significantly many calculations relatively fast, to implement the embedded system. Considering that the memory is speech, we used 128kbyte-RAM and 64kbyte ROM to save large amount of data. In case of speech input, we used 16-bit stereo audio codec, securing relatively accurate data through high resolution capacity.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.6
• /
• pp.578-586
• /
• 2021

We propose a semi-supervised learning method based on Variational AutoEncoder (VAE) and Unsupervised Data Augmentation (UDA) to improve the performance of an end-to-end speech recognizer. In the proposed method, first, the VAE-based augmentation model and the baseline end-to-end speech recognizer are trained using the original speech data. Then, the baseline end-to-end speech recognizer is trained again using data augmented from the learned augmentation model. Finally, the learned augmentation model and end-to-end speech recognizer are re-learned using the UDA-based semi-supervised learning method. As a result of the computer simulation, the augmentation model is shown to improve the Word Error Rate (WER) of the baseline end-to-end speech recognizer, and further improve its performance by combining it with the UDA-based learning method.

• Proceedings of the IEEK Conference
• /
• 2000.06d
• /
• pp.92-97
• /
• 2000

An AMR(Adaptive Multi-Rate) speech coding algorithm has been adopted as a standard speech codec for IMT-2000. It is based on the algebraic CELP, and consists of eight speech coding modes having the bit rate from 4.75 kbit/s to 12.2 kbit/s. It also contains the VAD(Voice Activity Detector), SCR (Source Controlled Rate) operation, and error concealment scheme for robustness in a radio channel. The bit rate of AMR is changed on a frame basis depending on the channel condition. In this paper, we introduced AMR speech coding algorithm and performed the real-time implementation using TMS320C6201, i.e., a Texas Instrument's fixed-point DSP. With the ANSI C source code released from ETSI and 3GPP, we convert and optimize the program to make it run in real time using the C compiler and assembly language. It is verified that the decoded result of the implemented speech codec on the DSP is identical with the PC simulation result using ANSI C code for test sequences. Also, actual sound input/output test using microphone and speaker demonstrates its proper real-time operation without distortions or delays.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.2E
• /
• pp.39-44
• /
• 1996

The performance of two-channel adaptive noise canceller is ofter degraded by the weights perturbation due to the speech signal. In this paper, an adaptive noise canceller employing a speech detector and two adaptation algorithms which are switched according to the speech detector is proposed. When highly correlated speech signal is detected, the tap weights of the adaptive filter are adapted by the sign algorithm. On the other hand, the weights are adapted by the NLMS algorithm when silence is detected or when the characteristics of the noise propagation channel is changed. The employed speech detector utilizes the power ratio of the input and the output of an adaptive linear prediction-error filter. According to the computer simulation, the proposed method yields better performance than conventional ones.

• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.503-506
• /
• 2001

In this paper, we describe the implementation of a microphone input level control algorithm and the speech improvement with this level controller in personal computer environment. The volume of speech obtained through a microphone affects the speech recognition rate directly. Therefore, proper input volume level control is desired fur better recognition. We considered some conditions for the successful volume controller implementation firstly, then checked its usefulness on our speech recognition system with common office environment speech database. Cepstral mean subtraction is also utilized far the channel-effect compensation of the database. Our implemented controller achieved approximately 50% reduction, i.e., improvement in speech recognition error rate.

• MALSORI
• /
• no.51
• /
• pp.71-84
• /
• 2004

In this paper, we propose the automatic speech database verification method(or called automatic verification) based on confidence measure for a large speech database. This method verifies the consistency between given transcription and speech using the confidence measure. The automatic verification process consists of two stages : the word-level likelihood computation stage and multi-level likelihood ratio computation stage. In the word-level likelihood computation stage, we calculate the word-level likelihood using the viterbi decoding algorithm and make the segment information. In the multi-level likelihood ratio computation stage, we calculate the word-level and the phone-level likelihood ratio based on confidence measure with anti-phone model. By automatic verification, we have achieved about 61% error reduction. And also we can reduce the verification time from 1 month in manual to 1-2 days in automatic.

• Speech Sciences
• /
• v.12 no.2
• /
• pp.183-191
• /
• 2005

This paper describes the automated call routing for call center system based on speech recognition. We focus on the task of automatically routing telephone calls based on a users fluently spoken response instead of touch tone menus in an interactive voice response system. Vector based call routing algorithm is investigated and normalization method suggested. Call center database which was collected by KT is used for call routing experiment. Experimental results evaluating call-classification from transcribed speech are reported for that database. In case of small training data, an average call routing error reduction rate of 9% is observed when normalization method is used.

• Speech Sciences
• /
• v.14 no.4
• /
• pp.131-144
• /
• 2007

In an effort to improve the performance of mel frequency cepstral coefficients (MFCC), we investigate the effects of varying the parameters for the filter banks and their associated windows on speech recognition rates. Specifically, the mel and bark scales are combined with various types of filter bank windows. Comparison and evaluation of the suggested methods are performed by two independent ways of speech recognition and the Fisher discriminant objective function. It is shown that the Hanning window based on the bark scale yields 28.1% relative performance improvements over the triangular window with the mel scale in speech recognition error rate. Further work on incorporating PCA and/or LDA would be desirable as a postprocessor to MFCC extraction.