Singing-voice synthesis using demi-syllable unit selection

Gu, Hung-Yan; He, Jiakang

doi:10.1109/icmlc.2016.7872965

Cited by 5 publications

(8 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Several approaches of rapping and singing voice synthesis have been proposed over the years. Early methods were based on unit concatenation [1,2,3,4] and statistical parametric synthesis [5,6]. There was also an attempt focused on speechto-rap voice conversion, based on a phase vocoder and beat tracking [8].…”

Section: Related Workmentioning

confidence: 99%

“…With the recent development of neural text-to-speech (TTS), the task of singing voice synthesis (SVS) is gaining popularity, since it has become feasible to produce natural and expressive speech more effectively. Before the development of deep neural network based synthesis, SVS systems were mainly based on unit selection technology [1,2,3,4] or parametric TTS [5,6]. During the last few years, with the establishment of neural TTS systems such as Tacotron [7], it has become possible to investigate approaches like neural rapping and singing.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Rapping-Singing Voice Synthesis based on Phoneme-level Prosody Control

Markopoulos,

Ellinas,

Vioni

et al. 2021

Preprint

View full text Add to dashboard Cite

In this paper, a text-to-rapping/singing system is introduced, which can be adapted to any speaker's voice. It utilizes a Tacotron-based multispeaker acoustic model trained on readonly speech data and which provides prosody control at the phoneme level. Dataset augmentation and additional prosody manipulation based on traditional DSP algorithms are also investigated. The neural TTS model is fine-tuned to an unseen speaker's limited recordings, allowing rapping/singing synthesis with the target's speaker voice. The detailed pipeline of the system is described, which includes the extraction of the target pitch and duration values from an a capella song and their conversion into target speaker's valid range of notes before synthesis. An additional stage of prosodic manipulation of the output via WSOLA is also investigated for better matching the target duration values. The synthesized utterances can be mixed with an instrumental accompaniment track to produce a complete song. The proposed system is evaluated via subjective listening tests as well as in comparison to an available alternate system which also aims to produce synthetic singing voice from read-only training data. Results show that the proposed approach can produce high quality rapping/singing voice with increased naturalness.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Rapping-Singing Voice Synthesis based on Phoneme-level Prosody Control

Markopoulos,

Ellinas,

Vioni

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Bernyanyi merupakan aktivitas seni yang dapat beruwujud sebagai sarana untuk kegiatan belajar-mengajar, hiburan, dan ritual pada agama tertentu. Mengembangkan teknologi yang dapat menunjang seni seperti pada penelitian [1], [2], [3] dan [4] kemudian bisa dikatakan sebagai hal yang dapat berguna pada lingkup yang luas dari aspek kehidupan manusia. Sebagaimana dilakukan pada penelitian [4], teknologi TTS (text-to-speech) merupakan salah satu contoh teknologi yang berpotensi untuk menunjang seni bernyanyi.…”

unclassified

“…Penelitian [4] dan [6] merupakan contoh dari penelitian yang berupaya melakukan sintensis berdasarkan aturan dari bahasa tertentu. Penelitian [4] melakukan sintesis suara bernyanyi dalam bahasa Mandarin. Metode yang digunakan pada penelitian ini adalah unitselection synthesis dengan mengambil ukuran sintesis per demi-syllable (setengah silabel) untuk menghindari jumlah kombinasi fonem yang terlalu besar, dan tidak menggunakan unit berukuran diphone untuk menghindari pencampuran dari unit-unit yang berdekatan (yang cukup banyak ditemui pada bahasa Mandarin).…”

unclassified

See 1 more Smart Citation

Sintesis Suara Bernyanyi Dengan Teknologi Text-To-Speech untuk Notasi Musik Angka dan Lirik Lagu Berbahasa Indonesia

Jonathan

Suyanto²

2020

Indonesian J. Electron. Instrum. Syst.

View full text Add to dashboard Cite

Singing is a work of art that can not be separated from human life. It then makes a research about develop the art of singing by technology will brings a useful impact for such a wide aspect of human life. This research is trying to synthesize singing voice with TTS (text-to-speech) technology, as it capability to produce sound with certain pronunciation at certain frequency of sound. Inputs that used in the system are texts of song in TXT format that contain the information of numbered musical notation and lyrics in Indonesian. These inputs will converted to a phonetic transcription, for then synthesize of song voice can done based on the transcription. In general, the system made successfully synthesize song voices with some feature that based on the convention of numbered musical notation. Based on 30 people of respondents, the song voice synthesized has 81.71% of accuracy with 6.24% of deviation standard. The syntax of song text also reputed as a user-friendly convention with only up to 3 times re-compilation done to synthesize 8 bar of song text by each of respondents without any error.

show abstract

Singing Voice Synthesis Using Deep Autoregressive Neural Networks for Acoustic Modeling

Ling

et al. 2019

Interspeech 2019

View full text Add to dashboard Cite

This paper presents a method of using autoregressive neural networks for the acoustic modeling of singing voice synthesis (SVS). Singing voice differs from speech and it contains more local dynamic movements of acoustic features, e.g., vibratos. Therefore, our method adopts deep autoregressive (DAR) models to predict the F0 and spectral features of singing voice in order to better describe the dependencies among the acoustic features of consecutive frames. For F0 modeling, discretized F0 values are used and the influences of the history length in DAR are analyzed by experiments. An F0 post-processing strategy is also designed to alleviate the inconsistency between the predicted F0 contours and the F0 values determined by music notes. Furthermore, we extend the DAR model to deal with continuous spectral features, and a prenet module with self-attention layers is introduced to process historical frames. Experiments on a Chinese singing voice corpus demonstrate that our method using DARs can produce F0 contours with vibratos effectively, and can achieve better objective and subjective performance than the conventional method using recurrent neural networks (RNNs).

show abstract

Singing-voice synthesis using demi-syllable unit selection

Cited by 5 publications

References 6 publications

Rapping-Singing Voice Synthesis based on Phoneme-level Prosody Control

Rapping-Singing Voice Synthesis based on Phoneme-level Prosody Control

Sintesis Suara Bernyanyi Dengan Teknologi Text-To-Speech untuk Notasi Musik Angka dan Lirik Lagu Berbahasa Indonesia

Singing Voice Synthesis Using Deep Autoregressive Neural Networks for Acoustic Modeling

Contact Info

Product

Resources

About