Vocal-Accompaniment Compatibility Estimation Using Self-Supervised and Joint-Embedding Techniques

Nakatsuka, T.; Watanabe, K.; Koyama, Yuki; Hamasaki, Masahiro; Goto, Masataka; Morishima, Shigeo

doi:10.1109/access.2021.3096819

Cited by 3 publications

(1 citation statement)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed evaluation method maximizes the accuracy rate in an evaluation process that enhances the effectiveness of the system. Nakatsuka et al [18] introduced a compatibility estimation method for vocal and accompaniment in audio tracks. Joint embedding techniques and self-supervised techniques are used for the estimation process.…”

Section: Related Workmentioning

confidence: 99%

Application of Information Teaching in Vocal Music Assisted Practice in Mobile Internet Era

Liu¹,

Wang²

2022

Preprint

View full text Add to dashboard Cite

Vocal music practices using the internet and mobile computing rely on the distinguishable pitch, note, and musical library information for ease of understanding. For addressing this problem an Assisted Data Processing Model for Tutor Function (ADPM-TF) is projected in this article. The proposed model learns the user requirements and their improvements based on training hours. The learned inputs are validated for the mishandled tutor sessions for identifying errors. The tutor assistance based on independent and joint (fused) data is recommended by the neural network learning based on previous session user performance/ rating. This process is streamlined depending on the user preference and switching ability over various sessions. The performance is validated using the metrics assistance ratio, error, processing time, and fusion rate.

show abstract

Section: Related Workmentioning

confidence: 99%

Application of Information Teaching in Vocal Music Assisted Practice in Mobile Internet Era

Liu¹,

Wang²

2022

Preprint

View full text Add to dashboard Cite

show abstract

MDX-Mixer: Music Demixing by Leveraging Source Signals Separated by Existing Demixing Models

NAKANO,

GOTO

2024

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

This paper presents MDX-Mixer, which improves music demixing (MDX) performance by leveraging source signals separated by multiple existing MDX models. Deep-learning-based MDX models have improved their separation performances year by year for four kinds of sound sources: "vocals," "drums," "bass," and "other". Our research question is whether mixing (i.e., weighted sum) the signals separated by state-ofthe-art MDX models can obtain either the best of everything or higher separation performance. Previously, in singing voice separation and MDX, there have been studies in which separated signals of the same sound source are mixed with each other using time-invariant or time-varying positive mixing weights. In contrast to those, this study is novel in that it allows for negative weights as well and performs time-varying mixing using all of the separated source signals and the music acoustic signal before separation. The time-varying weights are estimated by modeling the music acoustic signals and their separated signals by dividing them into short segments. In this paper we propose two new systems: one that estimates time-invariant weights using 1×1 convolution, and one that estimates time-varying weights by applying the MLP-Mixer layer proposed in the computer vision field to each segment. The latter model is called MDX-Mixer. Their performances were evaluated based on the source-to-distortion ratio (SDR) using the well-known MUSDB18-HQ dataset. The results show that the MDX-Mixer achieved higher SDR than the separated signals given by three state-of-theart MDX models.

show abstract

RETRACTED: Vocal music teaching method using fuzzy logic approach for musical performance evaluation

Dong

2023

IFS

View full text Add to dashboard Cite

Vocal music teaching through E-learning platforms and in-person classrooms requires intense attention and practice. The vocals and strings with/without instrument support are required for improving the voice, pitch presentation, and learning improvement. With digitalization, the assessments are performed using a computer and process-aided technologies for musical performance evaluation. This article introduces a Leveled-Fuzzy Logic Approach (LFLA) for evaluating the musical performance of the vocal teaching method. The evaluation improves the teaching mode by matching the actual learning guidelines. The vocal music teaching guidelines vary for different musical sessions and types. Such different aspects are analyzed using leveled fuzzy; the session flow is analyzed for maximum inference with the actual learning process. The number of learning levels forms the crisp input and the input is analyzed until the fuzzification extracts precise high matching for the guided and real-time teaching. The crisp inputs are induced for combinational inference during the fuzzification process using lagging features. The lagging features such as time, sessions, and evaluation per session are considered for fuzzification. The proposed approach is verified using the metrics assessment rate, lag detection, assessment time, and errors.

show abstract

Vocal-Accompaniment Compatibility Estimation Using Self-Supervised and Joint-Embedding Techniques

Cited by 3 publications

References 33 publications

Application of Information Teaching in Vocal Music Assisted Practice in Mobile Internet Era

Application of Information Teaching in Vocal Music Assisted Practice in Mobile Internet Era

MDX-Mixer: Music Demixing by Leveraging Source Signals Separated by Existing Demixing Models

RETRACTED: Vocal music teaching method using fuzzy logic approach for musical performance evaluation

Contact Info

Product

Resources

About