tempobeatdownbeat / tutorial Goto Github PK

Hey, Here I am opening this issue to request to include the following works in the related works table if they fit into the category!

#1:
M. Heydari and Z. Duan, “Don’t Look Back: An online beat tracking method using RNN and enhanced particle filtering,” in Proc. IEEE Int. Conf. Acoust. Speech Signal Process. (ICASSP), 2021.
Don't look back uses pre-trained RNN's from Madmom and a particle filtering inference for online beat tracking
https://arxiv.org/abs/2011.02619

#2:
M. Heydari, F. Cwitkowitz, and Z. Duan, “BeatNet: CRNN and particle filtering for online joint beat downbeat and meter tracking,” in Proc. of the 22nd Intl. Conf. on Music Information Retrieval (ISMIR), 2021.
The BeatNet uses CRNN and particle filtering for online music joint beat, downbeat, and meter tracking.
https://arxiv.org/abs/2108.03576
https://github.com/mjhydri/BeatNet

#3:
M. Heydari, M. McCallum, A. Ehmann, and Z. Duan, "A Novel 1D State Space for Efficient Music Rhythmic Analysis", In Proc. IEEE Int. Conf. Acoust. Speech Signal Process. (ICASSP), 2022. #(Recently submitted)
This paper proposes a novel 1D state space and a jump-reward inference model for online music joint beat, downbeat, tempo, and meter tracking.
https://arxiv.org/abs/2111.00704
https://github.com/mjhydri/1D-StateSpace

Many thanks!

Typo:

From the technical perspective ... multi-disciplinary context of the phenonema ...

Typo:

Here we see that almost all of the peaks that would correspond to hi-hat events have been surpressed.

In the welcome message we should mention the license as well

In the Hidden Markov Model section, the objective of inference is written as:

Since y represents the hidden states, shouldn't the correct objective be argmax_y rather than argmax_x?