![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2Fab9cea8b25971be9a41af85304974edc%2FSelection_035.png?generation=1595638712474392&alt=media\")
https://medium.com/@IckeIlknur/central-pattern-generators-to-synthesize-birdsongs-f0d09d6936c0
\n\ni didn't know that you can actually model it mathematically ...
\n\nWith the maths, maybe the below becomes relevant?
\n\nhttps://vsitzmann.github.io/siren/\nImplicit Neural Representations with Periodic Activation Functions
\n\nthere is another one using wavenet:\nhttps://github.com/shiba24/birdsong-generation-project
", "messages": [ { "id": "942004", "postDate": "07/23/2020 14:39:26", "content": "https://medium.com/@IckeIlknur/central-pattern-generators-to-synthesize-birdsongs-f0d09d6936c0
\n\ni didn't know that you can actually model it mathematically ...
\n\nWith the maths, maybe the below becomes relevant?
\n\nhttps://vsitzmann.github.io/siren/\nImplicit Neural Representations with Periodic Activation Functions
\n\nthere is another one using wavenet:\nhttps://github.com/shiba24/birdsong-generation-project
", "rawMarkdown": "https://medium.com/@IckeIlknur/central-pattern-generators-to-synthesize-birdsongs-f0d09d6936c0\n\ni didn't know that you can actually model it mathematically ...\n\nWith the maths, maybe the below becomes relevant?\n\nhttps://vsitzmann.github.io/siren/\nImplicit Neural Representations with Periodic Activation Functions\n\nthere is another one using wavenet:\nhttps://github.com/shiba24/birdsong-generation-project", "votes": null }, { "id": "943627", "postDate": "07/24/2020 13:30:59", "content": "In the Representing Audio Signals section of the SIREN paper it appears that SIREN should be able to model any audio signal with the loss function they provide. I'm still trying to understand this paper but it appears SIREN can model any audio signal by something like additive synthesis.
", "rawMarkdown": "In the Representing Audio Signals section of the SIREN [paper](https://arxiv.org/pdf/2006.09661.pdf ) it appears that SIREN should be able to model any audio signal with the loss function they provide. I'm still trying to understand this paper but it appears SIREN can model any audio signal by something like additive synthesis.", "votes": null }, { "id": "943689", "postDate": "07/24/2020 14:14:47", "content": "I am also trying out this competition and the article really gives me the background of the sound call. Thanks !!!
", "rawMarkdown": "**I am also trying out this competition and the article really gives me the background of the sound call. Thanks !!!**", "votes": null }, { "id": "944253", "postDate": "07/25/2020 00:58:34", "content": "DEEP FAKE BIRDSONG, 2020! \nhttps://www.kellyheatonstudio.com/deep-fake-birdsong
\n\n
i find an interesting work
\n\nhttps://github.com/timsainb/avgn\nhttps://github.com/timsainb/avgn_paper\nhttps://www.youtube.com/watch?v=0T2UU8-NCbM
\n\nthis is based on tensorflow. it can take in segmented and unsegmented audio. it would then perform auto-encoder to find latent bird song syllabus automatically
\n\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2Fc940221b2861ba292943898169d8d902%2FSelection_065.png?generation=1596243820052665&alt=media\")
in particular, this code implement syllables segmentation by dynamic thresholding:\nhttps://github.com/timsainb/AVGN/blob/master/avgn/segment_song/wav_to_syllables.py\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F113660%2F54e635fc7dc906fd2b484c3cb3f26a9c%2FSelection_066.png?generation=1596245642692279&alt=media\")
That is so cool! Thanks for posting!
", "rawMarkdown": "That is so cool! Thanks for posting!", "votes": null }, { "id": "1002707", "postDate": "09/08/2020 11:09:08", "content": "Indeed very interesting, thank you!
", "rawMarkdown": "Indeed very interesting, thank you!", "votes": null } ], "comments": [ { "id": 1002707, "author_name": "dmytroilin", "author_url": "", "post_date": "09/08/2020 11:09:08", "content": "Indeed very interesting, thank you!
", "votes": null, "replies": [] }, { "id": 943627, "author_name": "jackvial", "author_url": "", "post_date": "07/24/2020 13:30:59", "content": "In the Representing Audio Signals section of the SIREN paper it appears that SIREN should be able to model any audio signal with the loss function they provide. I'm still trying to understand this paper but it appears SIREN can model any audio signal by something like additive synthesis.
", "votes": null, "replies": [] }, { "id": 943689, "author_name": "jaseemck", "author_url": "", "post_date": "07/24/2020 14:14:47", "content": "I am also trying out this competition and the article really gives me the background of the sound call. Thanks !!!
", "votes": null, "replies": [] }, { "id": 944253, "author_name": "hengck23", "author_url": "", "post_date": "07/25/2020 00:58:34", "content": "DEEP FAKE BIRDSONG, 2020! \nhttps://www.kellyheatonstudio.com/deep-fake-birdsong
\n\n
That is so cool! Thanks for posting!
", "votes": null, "replies": [] } ] }, { "id": 953658, "author_name": "hengck23", "author_url": "", "post_date": "08/01/2020 00:59:58", "content": "i find an interesting work
\n\nhttps://github.com/timsainb/avgn\nhttps://github.com/timsainb/avgn_paper\nhttps://www.youtube.com/watch?v=0T2UU8-NCbM
\n\nthis is based on tensorflow. it can take in segmented and unsegmented audio. it would then perform auto-encoder to find latent bird song syllabus automatically
\n\n
in particular, this code implement syllables segmentation by dynamic thresholding:\nhttps://github.com/timsainb/AVGN/blob/master/avgn/segment_song/wav_to_syllables.py\n