Search

In this paper, we present SCOREQ, a novel approach for speech quality prediction. SCOREQ is a triplet loss function for contrastive regression that addresses the domain generalisation shortcoming exhibited by state of the art no-reference speech quality metrics. In the paper we: (i) illustrate the problem of L2 loss training failing at capturing the continuous nature of the mean opinion score (MOS) labels; (ii) demonstrate the lack of generalisation through a benchmarking evaluation across several speech domains; (iii) outline our approach and explore the impact of the architectural design decisions through incremental evaluation; (iv) evaluate the final model against state of the art models for a wide variety of data and domains. The results show that the lack of generalisation observed in state of the art speech quality metrics is addressed by SCOREQ. We conclude that using a triplet loss function for contrastive regression improves generalisation for speech quality prediction models but also has potential utility across a wide range of applications using regression-based predictive models., Comment: Accepted NeurIPS 2024

This paper explores the integration of model-based and data-driven approaches within the realm of neural speech and audio coding systems. It highlights the challenges posed by the subjective evaluation processes of speech and audio codecs and discusses the limitations of purely data-driven approaches, which often require inefficiently large architectures to match the performance of model-based methods. The study presents hybrid systems as a viable solution, offering significant improvements to the performance of conventional codecs through meticulously chosen design enhancements. Specifically, it introduces a neural network-based signal enhancer designed to post-process existing codecs' output, along with the autoencoder-based end-to-end models and LPCNet--hybrid systems that combine linear predictive coding (LPC) with neural networks. Furthermore, the paper delves into predictive models operating within custom feature spaces (TF-Codec) or predefined transform domains (MDCTNet) and examines the use of psychoacoustically calibrated loss functions to train end-to-end neural audio codecs. Through these investigations, the paper demonstrates the potential of hybrid systems to advance the field of speech and audio coding by bridging the gap between traditional model-based approaches and modern data-driven techniques., Comment: Accepted for publication in IEEE Signal Processing Magazine

This paper presents NOMAD (Non-Matching Audio Distance), a differentiable perceptual similarity metric that measures the distance of a degraded signal against non-matching references. The proposed method is based on learning deep feature embeddings via a triplet loss guided by the Neurogram Similarity Index Measure (NSIM) to capture degradation intensity. During inference, the similarity score between any two audio samples is computed through Euclidean distance of their embeddings. NOMAD is fully unsupervised and can be used in general perceptual audio tasks for audio analysis e.g. quality assessment and generative tasks such as speech enhancement and speech synthesis. The proposed method is evaluated with 3 tasks. Ranking degradation intensity, predicting speech quality, and as a loss function for speech enhancement. Results indicate NOMAD outperforms other non-matching reference approaches in both ranking degradation intensity and quality assessment, exhibiting competitive performance with full-reference audio metrics. NOMAD demonstrates a promising technique that mimics human capabilities in assessing audio quality with non-matching references to learn perceptual embeddings without the need for human-generated labels., Comment: Accepted for ICASSP 2024

We introduce LMCodec, a causal neural speech codec that provides high quality audio at very low bitrates. The backbone of the system is a causal convolutional codec that encodes audio into a hierarchy of coarse-to-fine tokens using residual vector quantization. LMCodec trains a Transformer language model to predict the fine tokens from the coarse ones in a generative fashion, allowing for the transmission of fewer codes. A second Transformer predicts the uncertainty of the next codes given the past transmitted codes, and is used to perform conditional entropy coding. A MUSHRA subjective test was conducted and shows that the quality is comparable to reference codecs at higher bitrates. Example audio is available at https://mjenrungrot.github.io/chrome-media-audio-papers/publications/lmcodec., Comment: 5 pages, accepted to ICASSP 2023, project page: https://mjenrungrot.github.io/chrome-media-audio-papers/publications/lmcodec

Non-reference speech quality models are important for a growing number of applications. The VoiceMOS 2022 challenge provided a dataset of synthetic voice conversion and text-to-speech samples with subjective labels. This study looks at the amount of variance that can be explained in subjective ratings of speech quality from metadata and the distribution imbalances of the dataset. Speech quality models were constructed using wav2vec 2.0 with additional metadata features that included rater groups and system identifiers and obtained competitive metrics including a Spearman rank correlation coefficient (SRCC) of 0.934 and MSE of 0.088 at the system-level, and 0.877 and 0.198 at the utterance-level. Using data and metadata that the test restricted or blinded further improved the metrics. A metadata analysis showed that the system-level metrics do not represent the model's system-level prediction as a result of the wide variation in the number of utterances used for each system on the validation and test datasets. We conclude that, in general, conditions should have enough utterances in the test set to bound the sample mean error, and be relatively balanced in utterance count between systems, otherwise the utterance-level metrics may be more reliable and interpretable., Comment: Preprint; accepted for Interspeech 2022

Speech coding facilitates the transmission of speech over low-bandwidth networks with minimal distortion. Neural-network based speech codecs have recently demonstrated significant improvements in quality over traditional approaches. While this new generation of codecs is capable of synthesizing high-fidelity speech, their use of recurrent or convolutional layers often restricts their effective receptive fields, which prevents them from compressing speech efficiently. We propose to further reduce the bitrate of neural speech codecs through the use of pretrained Transformers, capable of exploiting long-range dependencies in the input signal due to their inductive bias. As such, we use a pretrained Transformer in tandem with a convolutional encoder, which is trained end-to-end with a quantizer and a generative adversarial net decoder. Our numerical experiments show that supplementing the convolutional encoder of a neural speech codec with Transformer speech embeddings yields a speech codec with a bitrate of $600\,\mathrm{bps}$ that outperforms the original neural speech codec in synthesized speech quality when trained at the same bitrate. Subjective human evaluations suggest that the quality of the resulting codec is comparable or better than that of conventional codecs operating at three to four times the rate., Comment: Proceedings of INTERSPEECH 2022

Speech synthesis quality prediction has made remarkable progress with the development of supervised and self-supervised learning (SSL) MOS predictors but some aspects related to the data are still unclear and require further study. In this paper, we evaluate several MOS predictors based on wav2vec 2.0 and the NISQA speech quality prediction model to explore the role of the training data, the influence of the system type, and the role of cross-domain features in SSL models. Our evaluation is based on the VoiceMOS challenge dataset. Results show that SSL-based models show the highest correlation and lowest mean squared error compared to supervised models. The key point of this study is that benchmarking the statistical performance of MOS predictors alone is not sufficient to rank models since potential issues hidden in the data could bias the evaluated performances., Comment: Accepted ISSC 2023

We present SoundStream, a novel neural audio codec that can efficiently compress speech, music and general audio at bitrates normally targeted by speech-tailored codecs. SoundStream relies on a model architecture composed by a fully convolutional encoder/decoder network and a residual vector quantizer, which are trained jointly end-to-end. Training leverages recent advances in text-to-speech and speech enhancement, which combine adversarial and reconstruction losses to allow the generation of high-quality audio content from quantized embeddings. By training with structured dropout applied to quantizer layers, a single model can operate across variable bitrates from 3kbps to 18kbps, with a negligible quality loss when compared with models trained at fixed bitrates. In addition, the model is amenable to a low latency implementation, which supports streamable inference and runs in real time on a smartphone CPU. In subjective evaluations using audio at 24kHz sampling rate, SoundStream at 3kbps outperforms Opus at 12kbps and approaches EVS at 9.6kbps. Moreover, we are able to perform joint compression and enhancement either at the encoder or at the decoder side with no additional latency, which we demonstrate through background noise suppression for speech.

Recent advances in neural-network based generative modeling of speech has shown great potential for speech coding. However, the performance of such models drops when the input is not clean speech, e.g., in the presence of background noise, preventing its use in practical applications. In this paper we examine the reason and discuss methods to overcome this issue. Placing a denoising preprocessing stage when extracting features and target clean speech during training is shown to be the best performing strategy., Comment: 5 pages, 3 figures, presented at the Asilomar Conference on Signals, Systems, and Computers 2020

Good speech quality has been achieved using waveform matching and parametric reconstruction coders. Recently developed very low bit rate generative codecs can reconstruct high quality wideband speech with bit streams less than 3 kb/s. These codecs use a DNN with parametric input to synthesise high quality speech outputs. Existing objective speech quality models (e.g., POLQA, ViSQOL) do not accurately predict the quality of coded speech from these generative models underestimating quality due to signal differences not highlighted in subjective listening tests. We present WARP-Q, a full-reference objective speech quality metric that uses dynamic time warping cost for MFCC speech representations. It is robust to small perceptual signal changes. Evaluation using waveform matching, parametric and generative neural vocoder based codecs as well as channel and environmental noise shows that WARP-Q has better correlation and codec quality ranking for novel codecs compared to traditional metrics in addition to versatility for general quality assessment scenarios., Comment: Accepted for presentation at IEEE ICASSP 2021. Source code and data can be found on https://github.com/wjassim/WARP-Q.git

The recent emergence of machine-learning based generative models for speech suggests a significant reduction in bit rate for speech codecs is possible. However, the performance of generative models deteriorates significantly with the distortions present in real-world input signals. We argue that this deterioration is due to the sensitivity of the maximum likelihood criterion to outliers and the ineffectiveness of modeling a sum of independent signals with a single autoregressive model. We introduce predictive-variance regularization to reduce the sensitivity to outliers, resulting in a significant increase in performance. We show that noise reduction to remove unwanted signals can significantly increase performance. We provide extensive subjective performance evaluations that show that our system based on generative modeling provides state-of-the-art coding performance at 3 kb/s for real-world speech signals at reasonable computational complexity.

Estimation of perceptual quality in audio and speech is possible using a variety of methods. The combined v3 release of ViSQOL and ViSQOLAudio (for speech and audio, respectively,) provides improvements upon previous versions, in terms of both design and usage. As an open source C++ library or binary with permissive licensing, ViSQOL can now be deployed beyond the research context into production usage. The feedback from internal production teams at Google has helped to improve this new release, and serves to show cases where it is most applicable, as well as to highlight limitations. The new model is benchmarked against real-world data for evaluation purposes. The trends and direction of future work is discussed., Comment: 2020 Twelfth International Conference on Quality of Multimedia Experience (QoMEX)

This study compares the performances of different algorithms for coding speech at low bit rates. In addition to widely deployed traditional vocoders, a selection of recently developed generative-model-based coders at different bit rates are contrasted. Performance analysis of the coded speech is evaluated for different quality aspects: accuracy of pitch periods estimation, the word error rates for automatic speech recognition, and the influence of speaker gender and coding delays. A number of performance metrics of speech samples taken from a publicly available database were compared with subjective scores. Results from subjective quality assessment do not correlate well with existing full reference speech quality metrics. The results provide valuable insights into aspects of the speech signal that will be used to develop a novel metric to accurately predict speech quality from generative-model-based coders., Comment: 6 pages, 11 figures, conference

We propose to implement speech enhancement by the regeneration of clean speech from a salient representation extracted from the noisy signal. The network that extracts salient features is trained using a set of weight-sharing clones of the extractor network. The clones receive mel-frequency spectra of different noisy versions of the same speech signal as input. By encouraging the outputs of the clones to be similar for these different input signals, we train a feature extractor network that is robust to noise. At inference, the salient features form the input to a WaveNet network that generates a natural and clean speech signal with the same attributes as the ground-truth clean signal. As the signal becomes noisier, our system produces natural sounding errors that stay on the speech manifold, in place of traditional artifacts found in other systems. Our experiments confirm that our generative enhancement system provides state-of-the-art enhancement performance within the generative class of enhancers according to a MUSHRA-like test. The clones based system matches or outperforms the other systems at each input signal-to-noise (SNR) range with statistical significance., Comment: Accepted WASPAA 2019

We define salient features as features that are shared by signals that are defined as being equivalent by a system designer. The definition allows the designer to contribute qualitative information. We aim to find salient features that are useful as conditioning for generative networks. We extract salient features by jointly training a set of clones of an encoder network. Each network clone receives as input a different signal from a set of equivalent signals. The objective function encourages the network clones to map their input into a set of features that is identical across the clones. It additionally encourages feature independence and, optionally, reconstruction of a desired target signal by a decoder. As an application, we train a system that extracts a time-sequence of feature vectors of speech and uses it as a conditioning of a WaveNet generative system, facilitating both coding and enhancement., Comment: Interspeech 2019

The voice mode of the Opus audio coder can compress wideband speech at bit rates ranging from 6 kb/s to 40 kb/s. However, Opus is at its core a waveform matching coder, and as the rate drops below 10 kb/s, quality degrades quickly. As the rate reduces even further, parametric coders tend to perform better than waveform coders. In this paper we propose a backward-compatible way of improving low bit rate Opus quality by re-synthesizing speech from the decoded parameters. We compare two different neural generative models, WaveNet and LPCNet. WaveNet is a powerful, high-complexity, and high-latency architecture that is not feasible for a practical system, yet provides a best known achievable quality with generative models. LPCNet is a low-complexity, low-latency RNN-based generative model, and practically implementable on mobile phones. We apply these systems with parameters from Opus coded at 6 kb/s as conditioning features for the generative models. A listening test shows that for the same 6 kb/s Opus bit stream, synthesized speech using LPCNet clearly outperforms the output of the standard Opus decoder. This opens up ways to improve the decoding quality of existing speech and audio waveform coders without breaking compatibility., Comment: Proc. Interspeech 2020, 5 pages

Neural speech synthesis algorithms are a promising new approach for coding speech at very low bitrate. They have so far demonstrated quality that far exceeds traditional vocoders, at the cost of very high complexity. In this work, we present a low-bitrate neural vocoder based on the LPCNet model. The use of linear prediction and sparse recurrent networks makes it possible to achieve real-time operation on general-purpose hardware. We demonstrate that LPCNet operating at 1.6 kb/s achieves significantly higher quality than MELP and that uncompressed LPCNet can exceed the quality of a waveform codec operating at low bitrate. This opens the way for new codec designs based on neural synthesis models., Comment: Accepted for Interspeech 2019, 5 pages

We explore a variety of neural networks configurations for one- and two-channel spectrogram-mask-based speech enhancement. Our best model improves on previous state-of-the-art performance on the CHiME2 speech enhancement task by 0.4 decibels in signal-to-distortion ratio (SDR). We examine trade-offs such as non-causal look-ahead, computation, and parameter count versus enhancement performance and find that zero-look-ahead models can achieve, on average, within 0.03 dB SDR of our best bidirectional model. Further, we find that 200 milliseconds of look-ahead is sufficient to achieve equivalent performance to our best bidirectional model.

Neural speech synthesis models have recently demonstrated the ability to synthesize high quality speech for text-to-speech and compression applications. These new models often require powerful GPUs to achieve real-time operation, so being able to reduce their complexity would open the way for many new applications. We propose LPCNet, a WaveRNN variant that combines linear prediction with recurrent neural networks to significantly improve the efficiency of speech synthesis. We demonstrate that LPCNet can achieve significantly higher quality than WaveRNN for the same network size and that high quality LPCNet speech synthesis is achievable with a complexity under 3 GFLOPS. This makes it easier to deploy neural synthesis applications on lower-power devices, such as embedded systems and mobile phones., Comment: ICASSP 2019, 5 pages

Traditional parametric coding of speech facilitates low rate but provides poor reconstruction quality because of the inadequacy of the model used. We describe how a WaveNet generative speech model can be used to generate high quality speech from the bit stream of a standard parametric coder operating at 2.4 kb/s. We compare this parametric coder with a waveform coder based on the same generative model and show that approximating the signal waveform incurs a large rate penalty. Our experiments confirm the high performance of the WaveNet based coder and show that the speech produced by the system is able to additionally perform implicit bandwidth extension and does not significantly impair recognition of the original speaker for the human listener, even when that speaker has not been used during the training of the generative model., Comment: 5 pages, 2 figures