dc.contributor.author | Tamulionis, Mantas | |
dc.contributor.author | Sledevič, Tomyslav | |
dc.contributor.author | Serackis, Artūras | |
dc.date.accessioned | 2023-09-18T16:40:53Z | |
dc.date.available | 2023-09-18T16:40:53Z | |
dc.date.issued | 2023 | |
dc.identifier.other | (WOS_ID)000987245700001 | |
dc.identifier.uri | https://etalpykla.vilniustech.lt/handle/123456789/115926 | |
dc.description.abstract | This paper discusses an algorithm that attempts to automatically calculate the effect of room reverberation by training a mathematical model based on a recurrent neural network on anechoic and reverberant sound samples. Modelling the room impulse response (RIR) recorded at a 44.1 kHz sampling rate using a system identification-based approach in the time domain, even with deep learning models, is prohibitively complex and it is almost impossible to automatically learn the parameters of the model for a reverberation time longer than 1 s. Therefore, this paper presents a method to model a reverberated audio signal in the frequency domain. To reduce complexity, the spectrum is analyzed on a logarithmic scale, based on the subjective characteristics of human hearing, by calculating 10 octaves in the range 20–20,000 Hz and dividing each octave by 1/3 or 1/12 of the bandwidth. This maintains equal resolution at high, mid, and low frequencies. The study examines three different recurrent network structures: LSTM, BiLSTM, and GRU, comparing the different sizes of the two hidden layers. The experimental study was carried out to compare the modelling when each octave of the spectrum is divided into a different number of bands, as well as to assess the feasibility of using a single model to predict the spectrum of a reverberated audio in adjacent frequency bands. The paper also presents and describes in detail a new RIR dataset that, although synthetic, is calibrated with recorded impulses. | eng |
dc.format | PDF | |
dc.format.extent | p. 1-12 | |
dc.format.medium | tekstas / txt | |
dc.language.iso | eng | |
dc.relation.isreferencedby | Science Citation Index Expanded (Web of Science) | |
dc.relation.isreferencedby | Scopus | |
dc.source.uri | https://www.mdpi.com/2076-3417/13/9/5604 | |
dc.title | Investigation of machine learning model flexibility for automatic application of reverberation effect on audio signal | |
dc.type | Straipsnis Web of Science DB / Article in Web of Science DB | |
dcterms.accessRights | This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). | |
dcterms.license | Creative Commons – Attribution – 4.0 International | |
dcterms.references | 28 | |
dc.type.pubtype | S1 - Straipsnis Web of Science DB / Web of Science DB article | |
dc.contributor.institution | Vilniaus Gedimino technikos universitetas | |
dc.contributor.faculty | Elektronikos fakultetas / Faculty of Electronics | |
dc.subject.researchfield | T 001 - Elektros ir elektronikos inžinerija / Electrical and electronic engineering | |
dc.subject.researchfield | T 007 - Informatikos inžinerija / Informatics engineering | |
dc.subject.vgtuprioritizedfields | IK05 - Virtuali ir pridėtinė realybė / Virtual and augmented reality | |
dc.subject.ltspecializations | L106 - Transportas, logistika ir informacinės ir ryšių technologijos (IRT) / Transport, logistic and information and communication technologies | |
dc.subject.en | room reverberation | |
dc.subject.en | room impulse response | |
dc.subject.en | recurrent neural networks | |
dc.subject.en | audio signal spectrum | |
dc.subject.en | filter bank | |
dcterms.sourcetitle | Applied sciences: New Advances in Audio Signal Processing | |
dc.description.issue | iss. 9 | |
dc.description.volume | vol. 13 | |
dc.publisher.name | MDPI | |
dc.publisher.city | Basel | |
dc.identifier.doi | 000987245700001 | |
dc.identifier.doi | 2-s2.0-85159261976 | |
dc.identifier.doi | 85159261976 | |
dc.identifier.doi | 1 | |
dc.identifier.doi | 147463146 | |
dc.identifier.doi | 10.3390/app13095604 | |
dc.identifier.elaba | 167553630 | |