This paper presents HierSpeech, a high-quality end-to-end text-to-speech (TTS) system based on a hierarchical conditional variational autoencoder (VAE) utilizing self-supervised speech representations. Recently, single-stage TTS systems, which directly generate raw speech waveform from text, have been getting interest thanks to their ability in generating high-quality audio within a fully end-to-end training pipeline. However, there is still a room for improvement in the conventional TTS systems. Since it is challenging to infer both the linguistic and acoustic attributes from the text directly, missing the details of attributes, specifically linguistic information, is inevitable, which results in mispronunciation and over-smoothing problem in their synthetic speech. To address the aforementioned problem, we leverage self-supervised speech representations as additional linguistic representations to bridge an information gap between text and speech.

HierSpeech-U can adapt to a novel speaker by utilizing self-supervised speech representations without text transcripts.

This paper presents HierSpeech, a high-quality end-to-end text-to-speech (TTS) system based on a hierarchical conditional variational autoencoder (VAE) utilizing self-supervised speech representations. Recently, single-stage TTS systems, which directly generate raw speech waveform from text, have been getting interest thanks to their ability in generating high-quality audio within a fully end-to-end training pipeline. However, there is still a room for improvement in the conventional TTS systems. Since it is challenging to infer both the linguistic and acoustic attributes from the text directly, missing the details of attributes, specifically linguistic information, is inevitable, which results in mispronunciation and over-smoothing problem in their synthetic speech. To address the aforementioned problem, we leverage self-supervised speech representations as additional linguistic representations to bridge an information gap between text and speech.

Self-supervised speech models (S3Ms) have become an effective backbone for speech applications. Various analyses suggest that S3Ms encode linguistic properties. In this work, we seek a more fine-grained analysis of the word-level linguistic properties encoded in S3Ms. Specifically, we curate a novel dataset of near homophone (phonetically similar) and synonym (semantically similar) word pairs and measure the similarities between S3M word representation pairs. Our study reveals that S3M representations consistently and significantly exhibit more phonetic than semantic similarity. Further, we question whether widely used intent classification datasets such as Fluent Speech Commands and Snips Smartlights are adequate for measuring semantic abilities. Our simple baseline, using only the word identity, surpasses S3M-based models. This corroborates our findings and suggests that high scores on these datasets do not necessarily guarantee the presence of semantic content.

In this paper, we propose a singing voice synthesis model, Karaoker-SSL, that is trained only on text and speech data as a typical multi-speaker acoustic model. It is a low-resource pipeline that does not utilize any singing data end-to-end, since its vocoder is also trained on speech data. Karaoker-SSL is conditioned by self-supervised speech representations in an unsupervised manner. We preprocess these representations by selecting only a subset of their task-correlated dimensions. The conditioning module is indirectly guided to capture style information during training by multi-tasking. This is achieved with a Conformer-based module, which predicts the pitch from the acoustic model's output. Thus, Karaoker-SSL allows singing voice synthesis without reliance on hand-crafted and domain-specific features. There are also no requirements for text alignments or lyrics timestamps. To refine the voice quality, we employ a U-Net discriminator that is conditioned on the target speaker and follows a Diffusion GAN training scheme.

While many speakers of low-resource languages regularly code-switch between their languages and other regional languages or English, datasets of codeswitched speech are too small to train bespoke acoustic models from scratch or do language model rescoring. Here we propose finetuning self-supervised speech representations such as wav2vec 2.0 XLSR to recognize code-switched data. We find that finetuning self-supervised multilingual representations and augmenting them with n-gram language models trained from transcripts reduces absolute word error rates by up to 20% compared to baselines of hybrid models trained from scratch on code-switched data. Our findings suggest that in circumstances with limited training data finetuning self-supervised representations is a better performing and viable solution.

Despite rapid progress in the voice style transfer (VST) field, recent zero-shot VST systems still lack the ability to transfer the voice style of a novel speaker. In this paper, we present HierVST, a hierarchical adaptive end-to-end zero-shot VST model. Without any text transcripts, we only use the speech dataset to train the model by utilizing hierarchical variational inference and self-supervised representation. In addition, we adopt a hierarchical adaptive generator that generates the pitch representation and waveform audio sequentially. Moreover, we utilize unconditional generation to improve the speaker-relative acoustic capacity in the acoustic representation. With a hierarchical adaptive structure, the model can adapt to a novel voice style and convert speech progressively. The experimental results demonstrate that our method outperforms other VST models in zero-shot VST scenarios. Audio samples are available at \url{https://hiervst.github.io/}.

Recent work has explored using self-supervised learning (SSL) speech representations such as wav2vec2.0 as the representation medium in standard two-stage TTS, in place of conventionally used mel-spectrograms. It is however unclear which speech SSL is the better fit for TTS, and whether or not the performance differs between read and spontaneous TTS, the later of which is arguably more challenging. This study aims at addressing these questions by testing several speech SSLs, including different layers of the same SSL, in two-stage TTS on both read and spontaneous corpora, while maintaining constant TTS model architecture and training settings. Results from listening tests show that the 9th layer of 12-layer wav2vec2.0 (ASR finetuned) outperforms other tested SSLs and mel-spectrogram, in both read and spontaneous TTS. Our work sheds light on both how speech SSL can readily improve current TTS systems, and how SSLs compare in the challenging generative task of TTS. Audio examples can be found at https://www.speech.kth.se/tts-demos/ssr_tts

Recent work in the domain of speech enhancement has explored the use of self-supervised speech representations to aid in the training of neural speech enhancement models. However, much of this work focuses on using the deepest or final outputs of self supervised speech representation models, rather than the earlier feature encodings. The use of self supervised representations in such a way is often not fully motivated. In this work it is shown that the distance between the feature encodings of clean and noisy speech correlate strongly with psychoacoustically motivated measures of speech quality and intelligibility, as well as with human Mean Opinion Score (MOS) ratings. Experiments using this distance as a loss function are performed and improved performance over the use of STFT spectrogram distance based loss as well as other common loss functions from speech enhancement literature is demonstrated using objective measures such as perceptual evaluation of speech quality (PESQ) and short-time objective intelligibility (STOI).