Non-Intrusive Speech Quality Assessment with Transfer Learning and Subject-Specific Scaling

Nessler, Natalia; Cerňak, Miloš; Prandoni, Paolo; Mainar, Pablo

doi:10.21437/interspeech.2021-1685

Cited by 5 publications

(6 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The work in [18]- [57] describes MLbased approaches to NR quality estimation. Some of these NR tools produce estimates of subjective test scores that report speech or sound quality mean opinion score (MOS) [18]- [20], [25]- [28], [31], [36], [40], [42], [43], [45], [49], [50], [57], naturalness [29], [35], [37], listening effort [24], noise intrusiveness [50], and speech intelligibility [21], [33]. The non-intrusive speech quality assessment model called NISQA [53] uses log-mel-spectrograms to produce estimates of subjective speech quality as well as four constituent dimensions: noisiness, coloration, discontinuity, and loudness.…”

Section: A Existing Machine Learning Approachesmentioning

confidence: 99%

“…Sufficient datasets are rare and expensive to generate through laboratory testing, so crowd-sourced tests are becoming common. Joint training or transfer learning can leverage objective FR quality values [26], [31], [50] or impairment categories [49] alongside MOS values to maximize the benefit of those MOS values. Semi-supervised learning [42] is also an effective way to compensate for scarce MOS values.…”

Section: A Existing Machine Learning Approachesmentioning

confidence: 99%

See 1 more Smart Citation

Wideband Audio Waveform Evaluation Networks: Efficient, Accurate Estimation of Speech Qualities

Catellier,

Voran

2023

IEEE Access

View full text Add to dashboard Cite

Speech quality and speech intelligibility can vary dramatically across the wide range of currently available telecommunications systems, devices, and operating environments. This creates a strong demand for efficient real-time measurements of quality and intelligibly. Wideband Audio Waveform Evaluation Networks (WAWEnets) are convolutional neural networks that operate directly on wideband audio waveforms in order to produce evaluations of those waveforms. In the present work these evaluations give qualities of telecommunications speech (e.g., noisiness, intelligibility, overall speech quality). WAWEnets are no-reference networks because they do not require ''reference'' (original or undistorted) versions of the waveforms they evaluate. Our initial WAWEnet publication introduced four WAWEnets and each emulated the output of an established full-reference speech quality or intelligibility estimation algorithm. We have updated the WAWEnet architecture to be more efficient and effective. Here we present a single WAWEnet that closely tracks seven different quality and intelligibility values with per-segment correlations in the range of 0.92 to 0.96. We create a second network that additionally tracks four subjective speech quality dimensions. We offer a third network that focuses on just subjective quality scores and achieves a per-segment correlation of 0.97. The performance of our WAWEnet architecture compares favorably to models with orders-of-magnitude more parameters and computational complexity. This work has leveraged 334 hours of speech in 13 languages, over two million full-reference target values and over 93,000 subjective mean opinion scores. We also interpret the operation of WAWEnets and identify the key to their operation using the language of signal processing: ReLUs strategically move spectral information from non-DC components into the DC component. The DC values of 96 output signals define a vector in a 96-D latent space and this vector is then mapped to a quality or intelligibility value for the input waveform.

show abstract

Section: A Existing Machine Learning Approachesmentioning

confidence: 99%

Section: A Existing Machine Learning Approachesmentioning

confidence: 99%

Wideband Audio Waveform Evaluation Networks: Efficient, Accurate Estimation of Speech Qualities

Catellier,

Voran

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…As machine learning (ML) has become more powerful and accessible, numerous research groups have sought to apply ML to develop NR tools [17]- [50]. Some of these NR tools produce estimates of subjective test scores that report speech or sound quality mean opinion score (MOS) [17]- [19], [24]- [27], [30], [35], [38], [40], [41], [46], [47], naturalness [28], [34], [36], listening effort [23], noise intrusiveness [47], and speech intelligibility [20], [32]. The non-intrusive speech quality assessment model called NISQA [50] produces estimates of subjective speech quality as well as four constituent dimensions: noisiness, coloration, discontinuity, and loudness.…”

Section: A Existing Machine Learning Approachesmentioning

confidence: 99%

“…Sufficient datasets are rare and expensive to generate through laboratory testing, so crowd-sourced tests are becoming common. Joint training or transfer learning can leverage objective FR quality values [25], [30], [47] or impairment categories [46] alongside MOS values to maximize the benefit of those MOS values. Semi-supervised learning [40] is also an effective way to compensate for scarce MOS values.…”

Section: A Existing Machine Learning Approachesmentioning

confidence: 99%

“…But that same condition might be rated 3.0 when it appears in a test with higher quality conditions. Per-subject corrections are explored in [47]. A method to learn per-test correction factors is given in [45] and this approach may be viewed as the ML version of the linear algebra solution previously given in [51].…”

Section: A Existing Machine Learning Approachesmentioning

confidence: 99%

See 1 more Smart Citation

Wideband Audio Waveform Evaluation Networks: Efficient, Accurate Estimation of Speech Qualities

Catellier¹,

Voran²

2022

Preprint

View full text Add to dashboard Cite

Wideband Audio Waveform Evaluation Networks (WAWEnets) are convolutional neural networks that operate directly on wideband audio waveforms in order to produce evaluations of those waveforms. In the present work these evaluations give qualities of telecommunications speech (e.g., noisiness, intelligibility, overall speech quality). WAWEnets are no-reference networks because they do not require “reference” (original or undistorted) versions of the waveforms they evaluate. Our initial WAWEnet publication introduced four WAWEnets and each emulated the output of an established full-reference speech quality or intelligibility estimation algorithm. We have updated the WAWEnet architecture to be more efficient and effective. Here we present a single WAWEnet that closely tracks seven different quality and intelligibility values. We create a second network that additionally tracks four subjective speech quality dimensions. We offer a third network that focuses on just subjective quality scores and achieves very high levels of agreement. This work has leveraged 334 hours of speech in 13 languages, over two million full-reference target values and over 93,000 subjective mean opinion scores. We also interpret the operation of WAWEnets and identify the key to their operation using the language of signal processing: ReLUs strategically move spectral information from non-DC components into the DC component. The DC values of 96 output signals define a vector in a 96-D latent space and this vector is then mapped to a quality or intelligibility value for the input waveform.

show abstract

Multi-Channel Mosra: Mean Opinion Score and Room Acoustics Estimation Using Simulated Data and A Teacher Model

Coldenhoff,

Harper,

Kendrick

et al. 2024

ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Non-Intrusive Speech Quality Assessment with Transfer Learning and Subject-Specific Scaling

Cited by 5 publications

References 0 publications

Wideband Audio Waveform Evaluation Networks: Efficient, Accurate Estimation of Speech Qualities

Wideband Audio Waveform Evaluation Networks: Efficient, Accurate Estimation of Speech Qualities

Wideband Audio Waveform Evaluation Networks: Efficient, Accurate Estimation of Speech Qualities

Multi-Channel Mosra: Mean Opinion Score and Room Acoustics Estimation Using Simulated Data and A Teacher Model

Contact Info

Product

Resources

About