Bandwidth extension of narrowband speech using integer wavelet transform

Abstract

Public telephone systems transmit speech across a limited frequency range, about 300-3400 Hz, called narrowband (NB) which results in a significant reduction of quality and intelligibility of speech. This study proposes a fully backward compatible novel method for bandwidth extension (BWE) of NB speech. The method uses integer wavelet transform technique to provide a perceptually better wideband (WB) speech signal. The spectral envelope parameters are extracted from the down sampled frequency shifted version of the high-frequency components of speech signal existing above NB, which are spread by using pseudo-noise codes, and are embedded in the integer wavelet coefficients of NB speech signal. The hearing threshold is calculated in the integer wavelet domain and this threshold is employed as the embedding threshold. The embedded information is extracted at the receiving end to reconstruct the WB speech signal. Theoretical and simulation analyses show that the proposed method is robust to quantisation and channel noises. The comparison category rating listening and log spectral distortion tests clearly show that the reconstructed WB signal gives a much better performance in terms of speech quality when compared with some of the existing speech BWE methods employing data hiding.

Speech Bandwidth Extension Using DWT-FFT-Based Data Hiding

“…The obtained speech quality of the proposed method and conventional speech BE methods [8][9][10]15] is also assessed using absolute category rating (ACR) listening test [34] recommended by international telecommunications union (ITU-T). Speech BE using different data hiding techniques [8,9,[27][28][29][30][31] are used MOS test to evaluate the perceptual transparency. These tests were conducted in a quiet environment using headphones.…”

“…Bit error rate (BER) is used to evaluate the robustness of hidden data against quantization and channel noises. Speech BE using different data hiding techniques [11,13,27,29,30] are used NB-PESQ measure to evaluate the perceptual transparency.…”

“…Smaller LSD value indicates the best quality of the reconstructed HIB signal. Table 5 illustrates the results of the mean LSD values for the traditional techniques [8][9][10]15] and the proposed technique under μ-law coding [8,11,12,[27][28][29][30][31][32]. A clearly improved reconstructed HIB signal quality and thus clearly improved reconstructed WB signal quality of the proposed technique over the traditional techniques is observed from Tab.…”

Towards Robust Speech Super-Resolution

“…Besides source-filter model-based methods are those regarding speech signals as the addition of consecutive amplitude and frequency modulation signals [40], prediction of missing frequencies using error estimation, and non-negative matrix factorization [56]. Additionally, alternative signal transforms like modified discrete cosine transform and the wavelet transform are explored for spectral analysis and extension [42], [47].…”