Estimation and tracking of pitch for noisy speech signals using EMD based autocorrelation function algorithm
| Autor: | K Pratibha, H M Chandrashekar |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Computational complexity theory
Noise measurement Computer science Noise (signal processing) Autocorrelation Speech processing Signal Hilbert–Huang transform Background noise 030507 speech-language pathology & audiology 03 medical and health sciences Computer Science::Sound 0305 other medical science Algorithm |
| Zdroj: | 2017 2nd IEEE International Conference on Recent Trends in Electronics, Information & Communication Technology (RTEICT). |
| Popis: | Pitch is an important parameter in speech processing. It plays a major role in many speech processing applications. Speech signal is affected by background noise and it degrades the performance. Estimation of pitch for noisy speech signal is important task in many applications. This paper includes the estimation and also tracking of pitch using Empirical Mode Decomposition based Autocorrelation Function algorithm is used. Empirical Mode Decomposition is applicable for nonlinear and non-stationary signals and autocorrelation is efficient method for estimation of pitch in noise corrupted signals and as well as noiseless speech signal. Zero-crossing rate and energy based methods are used to separate voiced/unvoiced region. Only voiced region is considered for the estimation of pitch. Experimental analysis and result for different speech signals is explained in detail. This method gives the less computational complexity which is suitable for real time. It also gives the efficient estimation of pitch in speech signal. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|