{ "id": 176368, "title": "Nocall samples extracted from the example test audios", "url": "/competitions/birdsong-recognition/discussion/176368", "author_name": "Theo Viel", "post_date": "2020-08-21T13:15:16.811000", "votes": 44, "comment_count": 9, "views": 0, "content": "

Hi everybody,

\n

I have put together a punch of audio samples of the class nocall
\nAs a lot of people have been sharing stuff already I felt like contributing a bit as well.

\n

I generated them by manually selecting ares of >3 seconds from the example test audios.

\n

Samples shorter than 5 seconds were mirrored to be more than 5 seconds.

\n

That being said, they are available here :

\n\n

Format is .wav, sr=32kHz

\n

Hope they can be useful to anyone, they can be used as a small validation set for instance :)

", "messages": [ { "id": 980290, "postDate": "2020-08-21T13:15:16.810Z", "content": "

Hi everybody,

\n

I have put together a punch of audio samples of the class nocall
\nAs a lot of people have been sharing stuff already I felt like contributing a bit as well.

\n

I generated them by manually selecting ares of >3 seconds from the example test audios.

\n

Samples shorter than 5 seconds were mirrored to be more than 5 seconds.

\n

That being said, they are available here :

\n\n

Format is .wav, sr=32kHz

\n

Hope they can be useful to anyone, they can be used as a small validation set for instance :)

", "rawMarkdown": "Hi everybody,\n\nI have put together a punch of audio samples of the class `nocall`\nAs a lot of people have been sharing stuff already I felt like contributing a bit as well.\n\nI generated them by manually selecting ares of >3 seconds from the [example test audios](https://www.kaggle.com/c/birdsong-recognition/data?select=example_test_audio).\n\nSamples shorter than 5 seconds were mirrored to be more than 5 seconds.\n\nThat being said, they are available here : \n- https://www.kaggle.com/theoviel/nocall-samples\n\nFormat is `.wav`, `sr=32kHz`\n\nHope they can be useful to anyone, they can be used as a small validation set for instance :)", "votes": 44 }, { "id": 980297, "postDate": "2020-08-21T13:30:29.603Z", "content": "

This snippet should be fine to read the audios :

\n
import soundfile\nimport IPython.display as ipd\n\npath = \"path_to_audio\"\ny, sr = soundfile.read(path)\nipd.Audio(y, rate=sr)\n
\n

Let me know if there's any issue with the data

", "rawMarkdown": "This snippet should be fine to read the audios : \n``` \n\nimport soundfile\nimport IPython.display as ipd\n\npath = \"path_to_audio\"\ny, sr = soundfile.read(path)\nipd.Audio(y, rate=sr)\n```\n\nLet me know if there's any issue with the data", "votes": 6 }, { "id": 982777, "postDate": "2020-08-23T16:54:19.423Z", "content": "

so in this case no-call recordings are always perfectly \"clean\"? only if there is NO hint of a bird it's classified as no-call?

", "rawMarkdown": "so in this case no-call recordings are always perfectly \"clean\"? only if there is NO hint of a bird it's classified as no-call?", "votes": 3, "replies": [ { "id": 982805, "postDate": "2020-08-23T17:14:15.437Z", "content": "

There's no bird I could dinstiguish in these audios. But keep in mind that sometines bird sounds are really faint so I may have missed some.

", "rawMarkdown": "There's no bird I could dinstiguish in these audios. But keep in mind that sometines bird sounds are really faint so I may have missed some.", "votes": 3 }, { "id": 982834, "postDate": "2020-08-23T17:40:14.717Z", "rawMarkdown": "", "isDeleted": true } ] }, { "id": 984046, "postDate": "2020-08-24T19:34:55.207Z", "content": "

So many real world applications for this. Great code submittal below. Looking forward to trying it out.

", "rawMarkdown": "So many real world applications for this. Great code submittal below. Looking forward to trying it out." }, { "id": 982685, "postDate": "2020-08-23T15:20:52.170Z", "content": "

Looks Interesting.

", "rawMarkdown": "Looks Interesting." }, { "id": 980858, "postDate": "2020-08-21T23:57:10.373Z", "content": "

Wow, many thanks!
\nThis can lead to many good ideass

", "rawMarkdown": "Wow, many thanks!\nThis can lead to many good ideass" }, { "id": 1011789, "postDate": "2020-09-15T17:19:10.310Z", "content": "

thanks

", "rawMarkdown": "thanks\n" }, { "id": 980585, "postDate": "2020-08-21T17:48:29.580Z", "content": "

Thanks for that. It's very helpfully

", "rawMarkdown": "Thanks for that. It's very helpfully" } ], "comments": [ { "id": 980297, "author_name": "Theo Viel", "author_url": "", "post_date": "2020-08-21T13:30:29.603000", "content": "

This snippet should be fine to read the audios :

\n
import soundfile\nimport IPython.display as ipd\n\npath = \"path_to_audio\"\ny, sr = soundfile.read(path)\nipd.Audio(y, rate=sr)\n
\n

Let me know if there's any issue with the data

", "votes": 6, "replies": [] }, { "id": 982777, "author_name": "Proletheus", "author_url": "", "post_date": "2020-08-23T16:54:19.423000", "content": "

so in this case no-call recordings are always perfectly \"clean\"? only if there is NO hint of a bird it's classified as no-call?

", "votes": 3, "replies": [ { "id": 982805, "author_name": "Theo Viel", "author_url": "", "post_date": "2020-08-23T17:14:15.437000", "content": "

There's no bird I could dinstiguish in these audios. But keep in mind that sometines bird sounds are really faint so I may have missed some.

", "votes": 3, "replies": [] }, { "id": 982834, "author_name": "", "author_url": "", "post_date": "2020-08-23T17:40:14.717000", "content": "", "votes": 0, "replies": [] } ] }, { "id": 984046, "author_name": "Baron Smith", "author_url": "", "post_date": "2020-08-24T19:34:55.207000", "content": "

So many real world applications for this. Great code submittal below. Looking forward to trying it out.

", "votes": 0, "replies": [] }, { "id": 982685, "author_name": "Deep Contractor", "author_url": "", "post_date": "2020-08-23T15:20:52.170000", "content": "

Looks Interesting.

", "votes": 0, "replies": [] }, { "id": 980858, "author_name": "Fernandes", "author_url": "", "post_date": "2020-08-21T23:57:10.373000", "content": "

Wow, many thanks!
\nThis can lead to many good ideass

", "votes": 0, "replies": [] }, { "id": 1011789, "author_name": "RIYADH HAITHAM", "author_url": "", "post_date": "2020-09-15T17:19:10.310000", "content": "

thanks

", "votes": 0, "replies": [] }, { "id": 980585, "author_name": "Manh Lab", "author_url": "", "post_date": "2020-08-21T17:48:29.580000", "content": "

Thanks for that. It's very helpfully

", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "980290": "Hi everybody,\n\nI have put together a punch of audio samples of the class `nocall`\nAs a lot of people have been sharing stuff already I felt like contributing a bit as well.\n\nI generated them by manually selecting ares of >3 seconds from the [example test audios](https://www.kaggle.com/c/birdsong-recognition/data?select=example_test_audio).\n\nSamples shorter than 5 seconds were mirrored to be more than 5 seconds.\n\nThat being said, they are available here : \n- https://www.kaggle.com/theoviel/nocall-samples\n\nFormat is `.wav`, `sr=32kHz`\n\nHope they can be useful to anyone, they can be used as a small validation set for instance :)", "980297": "This snippet should be fine to read the audios : \n``` \n\nimport soundfile\nimport IPython.display as ipd\n\npath = \"path_to_audio\"\ny, sr = soundfile.read(path)\nipd.Audio(y, rate=sr)\n```\n\nLet me know if there's any issue with the data", "982777": "so in this case no-call recordings are always perfectly \"clean\"? only if there is NO hint of a bird it's classified as no-call?", "984046": "So many real world applications for this. Great code submittal below. Looking forward to trying it out.", "982685": "Looks Interesting.", "980858": "Wow, many thanks!\nThis can lead to many good ideass", "1011789": "thanks\n", "980585": "Thanks for that. It's very helpfully" } }