[2102.06306] DEEPF0: End-To-End Fundamental Frequency Estimation for Music and Speech Signals

We propose a novel pitch estimation technique called DeepF0, which leverages the available annotated data to directly learns from the raw audio in a data-driven manner. F0 estimation is important in various speech processing and music information retrieval applications. Existing deep learning models for pitch estimations have relatively limited learning capabilities due to their shallow receptive field. The proposed model addresses this issue by extending the receptive field of a network by introducing the dilated convolutional blocks into the network. The dilation factor increases the network receptive field exponentially without increasing the parameters of the model exponentially. To make the training process more efficient and faster, DeepF0 is augmented with residual blocks with residual connections. Our empirical evaluation demonstrates that the proposed model outperforms the baselines in terms of raw pitch accuracy and raw chroma accuracy even using 77.4% fewer network parameter

1 mentions: @m_morise
Keywords: end-to-end
Date: 2021/02/19 09:51

Referring Tweets

@m_morise t.co/Z88HIAW1XY 先日見かけたDeepF0の論文が公開されていた.のだけど,評価のやり方がアンフェアなような...

Related Entries

Read more GitHub - mmorise/kiritan_singing
0 users, 2 mentions 2019/11/18 14:20
Read more [1905.09263] FastSpeech: Fast, Robust and Controllable Text to Speechcontact arXivarXiv Twitter
0 users, 1 mentions 2020/03/05 12:54
Read more 5. The Unbearable Lightness of Artificial Intelligence and Big Data - YouTube
0 users, 7 mentions 2020/09/18 14:22
Read more [2009.10588] Anomalous diffusion dynamics of learning in deep neural networksopen searchopen navigat...
0 users, 1 mentions 2020/09/23 21:52
Read more 2ちゃんねるを対象とした悪口表現の抽出
4 users, 2 mentions 2020/10/15 09:51