Latent Timbre Synthesis

Tatar, Kıvanç; Bisig, Daniel; Pasquier, Philippe

doi:10.1007/s00521-020-05424-2

Cited by 13 publications

(8 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…is synthesis method allows composers and sound designers to interpolate and extrapolate between the timbres of multiple sounds in the latent space of audio frames [2]. Banerjee et al believe that the analysis of sound signals in the linear deterministic approach reached a new dimension and developed many well-equipped software to precisely measure and control the basic parameters of sound, such as pitch, intensity, and rhythm [3].…”

Section: Related Workmentioning

confidence: 99%

“…e main innovations of this paper are as follows: (1) feature extraction-this part mainly analyzes the characteristics of the audio signal on the basis of preprocessing and then extracts the characteristic curve of the audio signal to pave the way for the subsequent audio melody matching. (2) Audio feature library construction-this part mainly studies the music in MIDI le format and uses the improved contour algorithm to build the audio feature library, which is used as the data source for melody matching. is plays an important role in the timbre analysis of various musical instruments in the text, and he can distinguish the timbres of various musical instruments.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Music Timbre Extracted from Audio Signal Features

2022

Mobile Information Systems

View full text Add to dashboard Cite

Among the basic elements of music, timbre is one of the most important elements of sound, and it is also the main basis for distinguishing one pronunciation from another. People usually have the ability to “listen and argue” because everyone’s pronunciation is different. However, the existing audio extraction technology has low efficiency and low accuracy. Therefore, this paper aims to discuss the algorithm that can make music timbre feature extraction more accurate and efficient. For audio signal feature extraction, this paper proposed an audio feature based on harmonic components to describe the harmonic structure information in the audio signal spectrum. The algorithm in this paper extracts timbre features from the sound data of Western musical instruments and national musical instruments and analyzes the recognition accuracy. The experimental results showed that the classification accuracy of the four feature extractors is above 92%, among which B has the worst effect, with an accuracy of 92.42%, and D has the best classification effect, with an accuracy of 99.15%, which shows that the feature extraction algorithm designed in this paper combined with the traditional feature extraction algorithm can achieve better results.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Music Timbre Extracted from Audio Signal Features

2022

Mobile Information Systems

View full text Add to dashboard Cite

show abstract

“…For the SVM, tested parameter values for C and gamma were respectively [0.01, 0.1, 1, 10, 100, 1000] and [1, 0.1, 0.01, 0.001, 0.00001, 0.000001, 10]. For KNN, the tested N values were [1,2,3,4,5,6,7,8,9,10] and for the Random forest [1, 10, 100, 500,1000] trees were considered. In the case of the FuzzyRuLer, d was explored from 1 to half the number of features in the dataset and ratio with [0.9, 0.8, 0.7, 0.6, 0.5] values.…”

Section: Cross-validationmentioning

confidence: 99%

“…A deep learning based system that allows for interpolation and extrapolation between the timbre of multiple sounds is presented in [10]. Deep-learning systems are a promising path for sound synthesis applications, although their training times still do not allow for real-time feedback.…”

Section: Introductionmentioning

confidence: 99%

On-The-Fly Syntheziser Programming with Fuzzy Rule Learning

Paz

Nebot

Mugica

et al. 2020

Entropy

View full text Add to dashboard Cite

This manuscript explores fuzzy rule learning for sound synthesizer programming within the performative practice known as live coding. In this practice, sound synthesis algorithms are programmed in real time by means of source code. To facilitate this, one possibility is to automatically create variations out of a few synthesizer presets. However, the need for real-time feedback makes existent synthesizer programmers unfeasible to use. In addition, sometimes presets are created mid-performance and as such no benchmarks exist. Inductive rule learning has shown to be effective for creating real-time variations in such a scenario. However, logical IF-THEN rules do not cover the whole feature space. Here, we present an algorithm that extends IF-THEN rules to hyperrectangles, which are used as the cores of membership functions to create a map of the input space. To generalize the rules, the contradictions are solved by a maximum volume heuristics. The user controls the novelty-consistency balance with respect to the input data using the algorithm parameters. The algorithm was evaluated in live performances and by cross-validation using extrinsic-benchmarks and a dataset collected during user tests. The model’s accuracy achieves state-of-the-art results. This, together with the positive criticism received from live coders that tested our methodology, suggests that this is a promising approach.

show abstract

“…3). Th is type of corpus in a musical agent is known as a hybrid corpus [15]. In line with Respire's aesthetic choice of ambiguity, we aimed for a curation of abstract and ambient sounds for the musical agent's memory.…”

Section: Sound Memorymentioning

confidence: 99%

Respire: Virtual Reality Art with Musical Agent Guided by Respiratory Interaction

Tatar

Prpa

Pasquier

2019

Leonardo Music Journal

Self Cite

View full text Add to dashboard Cite

Respire is an immersive art piece that brings together three components: an immersive virtual reality (VR) environment, embodied interaction (via a breathing sensor) and a musical agent system to generate unique experiences of augmented breathing. The breathing sensor controls the user’s vertical elevation of the point of view under and over the virtual ocean. The frequency and patterns of breathing data guide the arousal of the musical agent, and the waviness of a virtual ocean in the environment. Respire proposes an intimate exploration of breathing through an intelligent mapping of breathing data to the parameters of visual and sonic environments.

show abstract

Latent Timbre Synthesis

Cited by 13 publications

References 25 publications

Music Timbre Extracted from Audio Signal Features

Music Timbre Extracted from Audio Signal Features

On-The-Fly Syntheziser Programming with Fuzzy Rule Learning

Respire: Virtual Reality Art with Musical Agent Guided by Respiratory Interaction

Contact Info

Product

Resources

About