اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدZero-Shot Text-to-Speech for Text-Based Insertion in Audio Narration
114
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Given a piece of speech and its transcript text, text-based speech editing aims to generate speech that can be seamlessly inserted into the given speech by editing the transcript. Existing methods adopt a two-stage approach: synthesize the input text using a generic text-to-speech (TTS) engine and then transform the voice to the desired voice using voice conversion (VC). A major problem of this framework is that VC is a challenging problem which usually needs a moderate amount of parallel training data to work satisfactorily. In this paper, we propose a one-stage context-aware framework to generate natural and coherent target speech without any training data of the target speaker. In particular, we manage to perform accurate zero-shot duration prediction for the inserted text. The predicted duration is used to regulate both text embedding and speech embedding. Then, based on the aligned cross-modality input, we directly generate the mel-spectrogram of the edited speech with a transformer-based decoder. Subjective listening tests show that despite the lack of training data for the speaker, our method has achieved satisfactory results. It outperforms a recent zero-shot TTS engine by a large margin.
قيم البحث
اقرأ أيضاً
This paper investigates how to leverage a DurIAN-based average model to enable a new speaker to have both accurate pronunciation and fluent cross-lingual speaking with very limited monolingual data. A weakness of the recently proposed end-to-end text
-to-speech (TTS) systems is that robust alignment is hard to achieve, which hinders it to scale well with very limited data. To cope with this issue, we introduce AdaDurIAN by training an improved DurIAN-based average model and leverage it to few-shot learning with the shared speaker-independent content encoder across different speakers. Several few-shot learning tasks in our experiments show AdaDurIAN can outperform the baseline end-to-end system by a large margin. Subjective evaluations also show that AdaDurIAN yields higher mean opinion score (MOS) of naturalness and more preferences of speaker similarity. In addition, we also apply AdaDurIAN to emotion transfer tasks and demonstrate its promising performance.
This paper introduces a novel adversarial algorithm for attacking the state-of-the-art speech-to-text systems, namely DeepSpeech, Kaldi, and Lingvo. Our approach is based on developing an extension for the conventional distortion condition of the adv
ersarial optimization formulation using the Cram`er integral probability metric. Minimizing over this metric, which measures the discrepancies between original and adversarial samples distributions, contributes to crafting signals very close to the subspace of legitimate speech recordings. This helps to yield more robust adversarial signals against playback over-the-air without employing neither costly expectation over transformation operations nor static room impulse response simulations. Our approach outperforms other targeted and non-targeted algorithms in terms of word error rate and sentence-level-accuracy with competitive performance on the crafted adversarial signals quality. Compared to seven other strong white and black-box adversarial attacks, our proposed approach is considerably more resilient against multiple consecutive playbacks over-the-air, corroborating its higher robustness in noisy environments.
This paper describes a variational auto-encoder based non-autoregressive text-to-speech (VAENAR-TTS) model. The autoregressive TTS (AR-TTS) models based on the sequence-to-sequence architecture can generate high-quality speech, but their sequential d
ecoding process can be time-consuming. Recently, non-autoregressive TTS (NAR-TTS) models have been shown to be more efficient with the parallel decoding process. However, these NAR-TTS models rely on phoneme-level durations to generate a hard alignment between the text and the spectrogram. Obtaining duration labels, either through forced alignment or knowledge distillation, is cumbersome. Furthermore, hard alignment based on phoneme expansion can degrade the naturalness of the synthesized speech. In contrast, the proposed model of VAENAR-TTS is an end-to-end approach that does not require phoneme-level durations. The VAENAR-TTS model does not contain recurrent structures and is completely non-autoregressive in both the training and inference phases. Based on the VAE architecture, the alignment information is encoded in the latent variable, and attention-based soft alignment between the text and the latent variable is used in the decoder to reconstruct the spectrogram. Experiments show that VAENAR-TTS achieves state-of-the-art synthesis quality, while the synthesis speed is comparable with other NAR-TTS models.
In this paper, we address the text-to-audio grounding issue, namely, grounding the segments of the sound event described by a natural language query in the untrimmed audio. This is a newly proposed but challenging audio-language task, since it requir
es to not only precisely localize all the on- and off-sets of the desired segments in the audio, but to perform comprehensive acoustic and linguistic understandings and reason the multimodal interactions between the audio and query. To tackle those problems, the existing method treats the query holistically as a single unit by a global query representation, which fails to highlight the keywords that contain rich semantics. Besides, this method has not fully exploited interactions between the query and audio. Moreover, since the audio and queries are arbitrary and variable in length, many meaningless parts of them are not filtered out in this method, which hinders the grounding of the desired segments. To this end, we propose a novel Query Graph with Cross-gating Attention (QGCA) model, which models the comprehensive relations between the words in query through a novel query graph. Besides, to capture the fine-grained interactions between audio and query, a cross-modal attention module that assigns higher weights to the keywords is introduced to generate the snippet-specific query representations. Finally, we also design a cross-gating module to emphasize the crucial parts as well as weaken the irrelevant ones in the audio and query. We extensively evaluate the proposed QGCA model on the public Audiogrounding dataset with significant improvements over several state-of-the-art methods. Moreover, further ablation study shows the consistent effectiveness of different modules in the proposed QGCA model.
This paper proposes VARA-TTS, a non-autoregressive (non-AR) text-to-speech (TTS) model using a very deep Variational Autoencoder (VDVAE) with Residual Attention mechanism, which refines the textual-to-acoustic alignment layer-wisely. Hierarchical lat
ent variables with different temporal resolutions from the VDVAE are used as queries for residual attention module. By leveraging the coarse global alignment from previous attention layer as an extra input, the following attention layer can produce a refined version of alignment. This amortizes the burden of learning the textual-to-acoustic alignment among multiple attention layers and outperforms the use of only a single attention layer in robustness. An utterance-level speaking speed factor is computed by a jointly-trained speaking speed predictor, which takes the mean-pooled latent variables of the coarsest layer as input, to determine number of acoustic frames at inference. Experimental results show that VARA-TTS achieves slightly inferior speech quality to an AR counterpart Tacotron 2 but an order-of-magnitude speed-up at inference; and outperforms an analogous non-AR model, BVAE-TTS, in terms of speech quality.
الأسئلة المقترحة
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
