| 1. |
|
|
| 2. |
|
|
| 3. |
- Ahlbäck, Sven, 1960-
(author)
-
MSTRE-Net: Multistreaming Acoustic Modeling for Automatic Lyrics Transcription
- 2021
-
Conference paper (peer-reviewed)abstract
- This paper makes several contributions to automatic lyrics transcription (ALT) research. Our main contribution is a novel variant of the Multistreaming Time-Delay Neural Network (MTDNN) architecture, called MSTRE-Net, which processes the temporal information using multiple streams in parallel with varying resolutions keeping the network more compact, and thus with a faster inference and an improved recognition rate than having identical TDNN streams. In addition, two novel preprocessing steps prior to training the acoustic model are proposed. First, we suggest using recordings from both monophonic and polyphonic domains during training the acoustic model. Second, we tag monophonic and polyphonic recordings with distinct labels for discriminating non-vocal silence and music instances during alignment. Moreover, we present a new test set with a considerably larger size and a higher musical variability compared to the existing datasets used in ALT literature, while maintaining the gender balance of the singers. Our best performing model sets the state-of-the-art in lyrics transcription by a large margin. For reproducibility, we publicly share the identifiers to retrieve the data used in this paper.
|
|
| 4. |
- Ahlbäck, Sven, 1960-
(author)
-
Pitch-Informed Instrument Assignment using a Deep Convolutional Network with Multiple Kernel Shapes
- 2021
-
Conference paper (peer-reviewed)abstract
- This paper proposes a deep convolutional neural network for performing note-level instrument assignment. Given a polyphonic multi-instrumental music signal along with its ground truth or predicted notes, the objective is to assign an instrumental source for each note. This problem is addressed as a pitch-informed classification task where each note is analysed individually. We also propose to utilise several kernel shapes in the convolutional layers in order to facilitate learning of timbre-discriminative feature maps. Experiments on the MusicNet dataset using 7 instrument classes show that our approach is able to achieve an average F-score of 0.904 when the original multi-pitch annotations are used as the pitch information for the system, and that it also excels if the note information is provided using third-party multi-pitch estimation algorithms. We also include ablation studies investigating the effects of the use of multiple kernel shapes and comparing different input representations for the audio and the note-related information.
|
|
| 5. |
- Ahlbäck, Sven, 1960-
(author)
-
What does AI learn from music? : on the problem of ground truth in music and the power of concepts
- 2020
-
Conference paper (peer-reviewed)abstract
- Many AI applications on music relates to music theoretical models, in the sense of musical concepts and ideas of how music works, that are not always made explicit, explicitly discussed or questioned. Already the choice of representations of music and sound as well as the selection of data for training models can influence models and creative output greatly. Here we will give some examples of problems arising when modelling basic musical features, from the experience of developing automatic music notation.
|
|
| 6. |
- Anterot, Joakim, 1965-
(author)
-
Seminarium slagverk inriktning orkester : PASIC 2020
- 2020
-
Conference paper (film/video) (peer-reviewed)abstract
- I det här seminariet ser vi närmare på teknik och frasering kring i första hand tamburin och triangel. Vi tar upp frågeställningar och tankar kring hur vi sätter allt i ett sammanhang och förhåller oss till olika musikaliska situationer. En vanlig fråga är hur en skall spela på en audition för att spela bäst. Vi lyfter några av dessa frågor här. Seminariet var en del av konventet PASIC 2020 som detta år hålls helt digitalt. I normala fall hålls PASIC årligen i USA och Indianapolis och är det största konventet i världen inom gebitet slagverk.
|
|
| 7. |
- Appelgren, Alva, et al.
(author)
-
Tuning in on motivation : Differences between non-musicians, amateurs, and professional musicians
- 2019
-
In: Psychology of Music. - : SAGE Publications. - 0305-7356 .- 1741-3087. ; 47:6, s. 864-873
-
Journal article (peer-reviewed)abstract
- The drive to learn and engage in music varies among individuals. Global motivation to do something can be intrinsic, for example, the joy and satisfaction in an activity. But motivation behind our action can also be extrinsic, such as the desire for fame, status or increased financial resources. The type of motivation probably influences to what degree individuals engage in musical activities. In this study, we examined the associations between the level of musical engagement and self-rated global motivation, factoring in age and sex, in a sample of 5,435 individuals. Musical engagement ranged from no music activity to amateurs and professional musicians. We found that intrinsic motivation increases with level of music activity and that motivation differs depending on sex, with females scoring higher on intrinsic motivation than males. Such differences may be considered in adjusting the forms of support offered to young musicians in music education. The phenomenon of motivation is complex, and we have highlighted areas that require further investigation, but this study has elucidated some differences in motivation types found in men and women, and between non-musicians, amateurs and professional musicians.
|
|
| 8. |
-
Arcs
- 2017
-
Artistic work (peer-reviewed)abstract
- Arcs (2017) 48’ – elektroakustiskt verk för fyra kanaler.Musikinstallation som är inköpt av Statens konstråd för placering på Svenska institutet i Paris och spelas där permanent från mitten av oktober 2018.Arcs (2017) elektroakustiskt verk för fyra kanaler 48'
|
|
| 9. |
|
|
| 10. |
- Atienza, Ricardo, et al.
(author)
-
Playing the design : Creating soundscapes through playful interaction
- 2023
-
In: SMC 2023 - Proceedings of the Sound and Music Computing Conference 2023. - : Sound and Music Computing Network. ; , s. 362-369
-
Conference paper (peer-reviewed)abstract
- This study takes inspiration from provocative design methods to gain knowledge on sound preferences regarding future vehicles’ designed sounds. A particular population subset was a triggering component of this study: people with hearing impairments. To that aim, we have developed a public installation in which to test a hypothetical futuristic city square. It includes three electrical vehicles whose sound can be designed by the visitor. The interface allows the user to interact and play with a number of provided sonic textures within a real-time web application, thus “playing” the design. This opens a design space of three distinct sounds that are mixed into an overall soundscape presented in a multichannel immersive environment. The paper describes the design processes involved.
|
|
