The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
In last few decades Sentiment Analysis is one of the most researched area in the field of Speech Recognition. Different methods like use of textual data, face expressions, voice signals, body movements and physiological signals were adopted for sentiment detection. Presented paper discuss about emotions/Sentiment and proposed an approach for analyzing these sentiment based on speech signals of an...
This paper presents a 2.4 kbps harmonic+noise coder with improved transient speech performance. The coder operates in either the steady mode or the transient mode. A predictive phase model is proposed for the efficient encoding of phase information in steady mode. In this model, the normalized frequency deviations are used to minimize the phase prediction errors. A polynomial is fitted to the set...
Waveform-matching coders preserve the shape of the target waveform and the time synchrony between the original and the synthesized signal. In parametric coders the shape of the encoded waveform is often changed, and the time-synchrony between the input and synthesized speech is not preserved. These two issues, time synchrony and waveform shape, are major obstacles in representing different speech...
Speech coding at very low bit rates has many applications such as answering machines, IP telephony, mobile communications, military communications etc. In this paper we describe a speech coder capable of operating at both 2.4 and 1.2kb/s, and produces good quality synthesised speech. The basic principle of the coder is based on frequency domain vocoding method where the LP excitation is classified...
This paper presents a new method for musical noise attenuation induced by spectral subtraction. The proposed technique consists in combining a wavelet packet decomposition with Wiener filtering. Simulations demonstrate how this artifact is reduced without affecting the original speech intelligibility.
This paper provides a way to classify vocal disorders for clinical applications, thanks to the idea of geometric signal separation in a feature space. It is well known that the human voice source generates complex signals including subharmonics and toroidal oscillations. Typical chaotic quantities — like the entropy and the dimension of the attractor — together with autocorrelation function, power...
This paper proposes a robust speaker-independent, connected digit recognition system for mobile applications. The system requires a small amount of ROM and low computational cost with high recognition accuracy. In addition, the system can be efficiently implemented on most currently available 32-bit fixed-point DSP chips. To reach these goals, we combined robust speech parameter processing technologies...
In noisy environments, a robust speech/non-speech detection is necessary for speech recognition. This paper presents a new method for speech/non-speech detection using third-order moments. The analysis of the energy third-order moment behaviour gives useful information on energy distribution. The new algorithm is compared to the one based on noise and speech statistics presented in [5]. The results...
A new algorithm to control the step-size of a frequency domain echo cancellation system is presented. The step-size is controlled via the estimate of the coherence between the microphone signal and the output of the echo cancellation filter. The use of coherence allows an independent step-size control for each frequency bin. Additive local noise correlated with the echo signal and thus corrupting...
Single-channel noise reduction for speech enhancement is often applied in cellular and in hands-free telephones. For speech distortions to be minimal, single-channel systems based on spectral subtraction cannot entirely eliminate environmental noise. A relatively high spectral noise floor has to remain in the speech signal. For further reduction of annoying noise components, pitch-adaptive post-filtering...
This paper discusses the effectiveness on the use of Hidden Markov Model tool kit (HTK) for recognizing speech, speaker and emotion from the emotional speeches using Mel frequency cepstral coefficients (MFCC) as a feature. Emotion independent speech recognition, speaker independent speech recognition, emotion independent speaker recognition and speaker independent emotion recognition systems were...
The voice is most prominent & primary mode of communication among the human beings. With this speech human can communicate with machine, thus this technique is used in education, military and medical sectors. Though this is not the new area, from last few decades researchers are working on the improvement of accuracy in voice recognition system. The design of that system concerns major issues...
This paper describes the effect of 0dB and 10dB babble noise on stuttered speech. 100 samples are collected from the subjects (stutterer), among which 80 samples are used to prepare database dictionary and 20 samples for testing. The samples are segmented and pass through Mel Frequency Cepstral Coefficient (MFCC) feature extraction and statistical estimation to prepare dictionary, then testing samples...
Improvement in quality as well as intelligibility is the prime requirement of speech enhancement devices such as hearing aids. Understanding speech signals can be very difficult for hearing impaired as well as normal hearing listeners especially when background noise is present. Majority of the traditional speech enhancement algorithms have reported improvement in quality but not necessarily intelligibility...
Wavelet transform is an important tool used in many application areas. This paper proposed the analysis of noise reduction techniques which are based on wavelet transform. In this paper three noise reduction techniques based on wavelet transform are described. These methods are Wavelet Split Coefficient, Hard Thresholding and Soft Thresholding. MATLAB GUI is developed for visualization of results...
In degraded listening conditions, speakers are known to adapt their speech via the Lombard reflex to make it more comprehensible. This characteristic has been used in previous work to modify speech recorded in quiet before it is rendered in a noisy environment. The spectral modifications used have been found to be effective in low-pass noise such as babble noise. In this work, we investigate intelligibility...
Speech recognition and speaker recognition have wide range of applications in security systems and smart home designs. In this paper we discuss a method by which text dependent speaker recognition can be used to control gear shifting in light motor vehicles which could be helpful for people who lost one hand in accidents to drive cars. Speaker recognition involves two processes namely feature extraction...
Speech Enhancement refers to the improvement in the intelligibility and or the quality of the degraded speech signal using signal processing techniques. Till recent days speech enhancement is a very difficult problem because the noise content in the speech signals varies its nature and characteristics with time and application to application. Using speech enhancement techniques the quality and intelligibility...
This paper proposes a robust speaker direction estimation method based on a microphone array for voice based interaction with smart TV. The proposed method uses speech basis and associated weights of non-negative matrix factorization for finding the speaker independent utterance direction from input signal with noise. The experimental results of the speaker direction estimation in real acoustic environment...
This paper presents a wind-noise suppressor with wind-burst detection based on a stationary noise estimate. Input that exceeds an estimated stationary noise level is detected as wind burst. Upon detection, the spectral gain is reduced to apply maximum suppression determined by a floor gain. False detection of a speech onset as wind is minimized by speech-wind discrimination with spectral smoothness...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.