An efficient model-based multirate method for reconstruction of audio signals across long gaps

Esquef, Paulo A. A.; Biscainho, Luiz W. P.

doi:10.1109/tsa.2005.858018

Cited by 22 publications

(9 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, the utterances in Erase-50 ms were produced by erasing 50 ms every 250 ms of data, leading to utterances which are successions of 50 ms of silence, 200 ms of speech, 50 ms of silence, etc. We shall note that gaps from 30 ms to 50 ms are already considered very long gaps in the audio restoration literature (Esquef and Biscainho, 2006;Godsill and Rayner, 1998;Maher, 1994).…”

Section: Experimental Settingmentioning

confidence: 91%

“…Various approaches have emerged recently which attempt to analyze the structure of the spectrogram of an acoustical scene Schmidt and Mørup, 2006;Smaragdis, 2004), while on the other side gap interpolation techniques have been the subject of research for many years (Achan et al, 2005;Cemgil and Godsill, 2005;Clark and Atlas, 2008;Esquef and Biscainho, 2006;Godsill and Rayner, 1998;Lu et al, 2003;Wolfe and Godsil, 2005). However, only few models so far try to deal with both issues.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Computational auditory induction as a missing-data model-fitting problem with Bregman divergence

Roux

Kameoka²,

Cheveigné

et al. 2011

Speech Communication

View full text Add to dashboard Cite

Section: Experimental Settingmentioning

confidence: 91%

Section: Introductionmentioning

confidence: 99%

Computational auditory induction as a missing-data model-fitting problem with Bregman divergence

Roux

Kameoka²,

Cheveigné

et al. 2011

Speech Communication

View full text Add to dashboard Cite

“…We propose the use of a Gaussian-shaped function for elsewhere (7) where (8) which results in unity power of the Gaussian-shaped spectrum, i.e.,…”

Section: A Harmonic Noise Model Of Excitationmentioning

confidence: 99%

“…AR models are of particular interest in the restoration of lost speech samples [2]- [7]. Least square error AR (LSAR) and maximum a posteriori (MAP) AR are among the most popular estimation methods [3], [4].…”

Section: Introductionmentioning

confidence: 99%

“…In [6] Kauppinen and Roth replace the missing portion of the excitation with zeros. Esquef et al [7] recently proposed a time-reversed excitation substitution algorithm with a multirate post-processing module for audio gap restoration. The time reversed excitation technique prevents artefacts that may result from discontinuities of the excitation signal.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Interpolation of Lost Speech Segments Using LP-HNM Model With Codebook Post-Processing

Zavarehei¹,

Vaseghi²

2008

IEEE Trans. Multimedia

View full text Add to dashboard Cite

This paper presents a method for interpolation of lost speech segments. The interpolation method can be used for packet loss concealment in voice communication over mobile phones, for voice over IP or for restoration of lost segments in speech recordings. The interpolation method employs a combination of a linear prediction (LP) model of the spectral envelope and a harmonic noise model (HNM) of the excitation of speech. The speech interpolation problem is transformed to the modeling and interpolation of the trajectories of LP parameters and the amplitude, phase and harmonicity of HNM tracks of speech excitation. In particular, the interpolation of harmonicity results in a smooth transition from voiced to unvoiced speech and vice versa. Crucially, the proposed interpolation method does not suffer from the consequences of zero-excitation of conventional autoregressive (AR) interpolation. Different combinations of linear and autoregressive interpolation methods are evaluated for the estimation of the time-varying parameters of LP-HNM tracks. Furthermore, a post-processing codebook mapping, employed to enhance the interpolation of the spectral envelope of speech, results in improved output quality for longer length speech gaps. For different packet loss rates and patterns of distributions of missing speech gaps, the proposed interpolation methods are evaluated and compared with popular AR-based interpolation methods and the speech packet recovery method specified in the ITU G.711 standard, as a reference. The evaluation results show that the proposed methods substantially improve the restoration of formants and harmonic tracks and consistently results in significant performance gain and improved perceptual quality of speech.

show abstract

A Real-Time Missing Data Recovery Method Using Recurrent Neural Network for Multiple Transmissions

Lin

Lee

et al. 2018

Recent Advances in Intelligent Information Hiding and Multimedia Signal Processing

View full text Add to dashboard Cite

An efficient model-based multirate method for reconstruction of audio signals across long gaps

Cited by 22 publications

References 18 publications

Computational auditory induction as a missing-data model-fitting problem with Bregman divergence

Computational auditory induction as a missing-data model-fitting problem with Bregman divergence

Interpolation of Lost Speech Segments Using LP-HNM Model With Codebook Post-Processing

A Real-Time Missing Data Recovery Method Using Recurrent Neural Network for Multiple Transmissions

Contact Info

Product

Resources

About