• The Journal of the Acoustical Society of Korea
• /
• v.19 no.6
• /
• pp.53-61
• /
• 2000

In the speech recognition under the noisy environments, reducing the mismatch introduced between training and testing environments is an important issue. Spectral subtraction is widely used technique because of its simplicity and relatively good performance in noisy environments. In this paper, we introduce histogram method as a reliable noise estimation approach for spectral subtraction. This method has advantages over the conventional noise estimation methods in that it does not need to detect non-speech intervals and it can estimate the noise spectra even in time-varying noise environments. Even though spectral subtraction is performed using a reliable average noise spectrum by the histogram method, considerable amount of residual noise remains due to the variations of instantaneous noise spectrum about mean. To overcome this limitation, we propose a new over-estimation technique based on distribution characteristics of histogram used for noise estimation. Since the proposed technique decides the degree of over-estimation adaptively according to the measured noise distribution, it has advantages to be few the influence of the SNR variation on the noise levels. According to speaker-independent isolated word recognition experiments in car noise environment under various SNR conditions, the proposed histogram-based over-estimation technique outperforms the conventional over-estimation technique.

• Maxillofacial Plastic and Reconstructive Surgery
• /
• v.28 no.4
• /
• pp.310-319
• /
• 2006

Submucosal type cleft palate is subdivision of the cleft palate. It is very difficult to find submucosal cleft, because when we exam submucosal type cleft palate patients, it seems to be normal. But in fact, there are abnormal union of palatal muscles of submucosal type cleft palate patients. Because of late detection, the treatment - for example, the operation or the speech therapy - for the submucosal type cleft palate patient usually becomes late. Some patients visited our hospital due to speech disorder nevertheless normal intraoral appearance. After precise intraoral examination, we found out submucosal cleft palate. We evaluated the speech before and after surgery of these patients. In this study, we want to find the objective characteristics of submucosal type cleft palate patients, comparing with the normal and the complete cleft palate patients. Experimental groups were 10 submucosal type cleft palate patients and 10 complete cleft palate patients who got the operation in our hospital. And, the controls were 10 normal person. The sentence patterns using in this study were simple 5 vowels. Using CSL program we evaluated the Formant, Bandwidth. We analized the spectral characteristics of speech signals of 3 groups, before and after the operation. In most cases, the formant scores were higher in experimental groups (complete cleft palate group and submucosal type cleft palate group) than controls. There were small differences when speeching /a/, /i/, /e/ between experimental groups and control groups, large differences when speeching /o/, /u/. After surgery the formant scores were decreased in experimental groups (complete cleft palate group and submucosal type cleft palate group). In bandwidth scores, there were no significant differences between experimental groups and controls.

• Journal of the Institute of Electronics Engineers of Korea TE
• /
• v.37 no.5
• /
• pp.62-70
• /
• 2000

In this study, speech recognition for Korean digit is performed using filter bank which is programmed discrete HMM and DTW. Spectral analysis reveals speech signal features which are mainly due to the shape of the vocal tract. And spectral feature of speech are generally obtained as the exit of filter banks, which properly integrated a spectrum at defined frequency ranges. A set of 8 band pass filters is generally used since it simulates human ear processing. And defined frequency ranges are 320-330, 450-460, 640-650, 840-850, 900-1000, 1100-1200, 2000-2100, 3900-4000Hz and then sampled at 8kHz of sampling rate. Frame width is 20ms and period is 10ms. Accordingly, we found that the recognition rate of DTW is better than HMM for Korean digit speech in the experimental result. Recognition accuracy of Korean digit speech using filter bank is 93.3% for the 24th BPF, 89.1% for the 16th BPF and 88.9% for the 8th BPF of hardware realization of voice dialing system.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.1
• /
• pp.30-37
• /
• 2022

This paper compares and evaluates model performance from two perspectives according to the learning target and network structure for training Deep Neural Network (DNN)-based speech enhancement models in the frequency domain. In this case, spectrum mapping and Time-Frequency (T-F) masking techniques were used as learning targets, and a real network and a complex network were used for the network structure. The performance of the speech enhancement model was evaluated through two objective evaluation metrics: Perceptual Evaluation of Speech Quality (PESQ) and Short-Time Objective Intelligibility (STOI) depending on the scale of the dataset. Test results show the appropriate size of the training data differs depending on the type of networks and the type of dataset. In addition, they show that, in some cases, using a real network may be a more realistic solution if the number of total parameters is considered because the real network shows relatively higher performance than the complex network depending on the size of the data and the learning target.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.8 no.2
• /
• pp.60-69
• /
• 1998

This paper proposes a speech synthesis method using Fuzzy VQ, and then study how to make choice of fuzziness value which optimizes (controls) the performance of FVQ in order to obtain the synthesized speech which is closer to the original speech. When FVQ is used to synthesize a speech, analysis stage generates membership function values which represents the degree to which an input speech pattern matches each speech patterns in codebook, and synthesis stage reproduces a synthesized speech, using membership function values which is obtained in analysis stage, fuzziness value, and fuzzy-c-means operation. By comparsion of the performance of the FVQ and VQ synthesizer with simmulation, we show that, although the FVQ codebook size is half of a VQ codebook size, the performance of FVQ is almost equal to that of VQ. This results imply that, when Fuzzy VQ is used to obtain the same performance with that of VQ in speech synthesis, we can reduce by half of memory size at a codebook storage. And then we have found that, for the optimized FVQ with maximum SQNR in synthesized speech, the fuzziness value should be small when the variance of analysis frame is relatively large, while fuzziness value should be large, when it is small. As a results of comparsion of the speeches synthesized by VQ and FVQ in their spectrogram of frequency domain, we have found that spectrum bands(formant frequency and pitch frequency) of FVQ synthesized speech are closer to the original speech than those using VQ.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.342-350
• /
• 2022

In this paper, actual drone noise samples are collected for speech processing in disaster environments to build noise-corrupted speech database, and speech enhancement performance is evaluated by applying spectrum subtraction and mask-based speech enhancement techniques. To improve the performance of VoiceFilter (VF), an existing deep neural network-based speech enhancement model, we apply the Self-Attention operation and use the estimated noise information as input to the Attention model. Compared to existing VF model techniques, the experimental results show 3.77%, 1.66% and 0.32% improvements for Source to Distortion Ratio (SDR), Perceptual Evaluation of Speech Quality (PESQ), and Short-Time Objective Intelligence (STOI), respectively. When trained with a 75% mix of speech data with drone sounds collected from the Internet, the relative performance drop rates for SDR, PESQ, and STOI are 3.18%, 2.79% and 0.96%, respectively, compared to using only actual drone noise. This confirms that data similar to real data can be collected and effectively used for model training for speech enhancement in environments where real data is difficult to obtain.

• Journal of Digital Convergence
• /
• v.19 no.2
• /
• pp.447-456
• /
• 2021

This study analyzed 28 domestic and foreign studies related Prompts for Restructuring Oral Muscular Phonetic Targets treatment techniques from 1984 to 2020 to prepare basic data for the development of PROMPT intervention programs and examination tools. According to the analysis, continuous research has been conducted since 1984 when the prompt study was first started, and the method of research was 16 intervention studies, with the highest number of speech disorders, and the target age being 3 to 5 years old, the most frequently conducted for infancy. The treatment was the most frequent in the 16th sessions, and the activities were based on the Motor Speech Hierarchy(MSH), except for the subjects of the non-verbal autism spectrum disorder. According to the analysis of the dependent variables, 'speech production' was the most common, followed by 'speech motor control', 'articulation', and 'speech intelligibility' were highest. Combined with all these studies, it suggests that PROMPT, which are directly useful for exercise spoken word production, are effectively being used outside the country and that it is necessary to develop a PROMPT program that can be applied domestically, in Korea.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.8
• /
• pp.383-388
• /
• 2006

This Paper presents a novel approach to single channel microphone speech enhancement in noisy environments. Widely used noise reduction techniques based on the spectral subtraction are generally expressed as a spectral gam depending on the signal-to-noise ratio (SNR). The well-known decision-directed(DD) estimator of Ephraim and Malah efficiently reduces musical noise under the background noise conditions, but generates the delay of the a prioiri SNR because the DD weights the speech spectrum component of the Previous frame in the speech signal. Therefore, the noise suppression gain which is affected by the delay of the a priori SNR, which is estimated by the DD matches the previous frame rather than the current one, so after noise suppression. this degrades the noise reduction performance during speech transient periods. We propose a computationally simple but effective speech enhancement technique based on the sigmoid type function for the weight Parameter of the DD. The proposed approach solves the delay problem about the main parameter, the a priori SNR of the DD while maintaining the benefits of the DD. Performances of the proposed enhancement algorithm are evaluated by ITU-T p.862 Perceptual Evaluation of Speech duality (PESQ). the Mean Opinion Score (MOS) and the speech spectrogram under various noise environments and yields better results compared with the fixed weight parameter of the DD.

• Phonetics and Speech Sciences
• /
• v.4 no.3
• /
• pp.103-110
• /
• 2012

In this paper, a target signal detection method using multiple signal classification (MUSIC) algorithm is proposed. The MUSIC algorithm is a subspace-based direction of arrival (DOA) estimation method. Using the inverse of the eigenvalue-weighted eigen spectra, the algorithm detects the DOAs of multiple sources. To apply the algorithm in target signal detection for GSC-based beamforming, we utilize its spectral response for the DOA of the target source in noisy conditions. The performance of the proposed target signal detection method is compared with those of the normalized cross-correlation (NCC), the fixed beamforming, and the power ratio method. Experimental results show that the proposed algorithm significantly outperforms the conventional ones in receiver operating characteristics (ROC) curves.

• ETRI Journal
• /
• v.23 no.1
• /
• pp.16-22
• /
• 2001

Efficient quantization methods of the line spectrum pairs (LSP) which have good performances, low complexity and memory are proposed. The adaptive quantization range method utilizing the ordering property of LSP parameters is used in a scalar quantizer and a vector-scalar hybrid quantizer. As the maximum quantization range of each LSP parameter is varied adaptively on the quantized value of the previous order's LSP parameter, efficient quantization methods can be obtained. The proposed scalar quantization algorithm needs 31 bits/frame, which is 3 bits less per frame than in the conventional scalar quantization method with interframe prediction to maintain the transparent quality of speech. The improved vector-scalar quantizer achieves an average spectral distortion of 1 dB using 26 bits/frame. The performances of proposed quantization methods are also evaluated in the transmission errors.