![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F4133984%2F2a64826fc14f940930941e84a9daaea4%2Fspec_A.png?generation=1593521490537392&alt=media\")
As I am new to this topic also, I hope to share any insights that I learn along the way. Advanced practitioners may not find my insight useful, but your feedback is always appreciated!
\n\nHere is a resource to learn how to convert an audio file to a spectogram - https://www.kaggle.com/msripooja/steps-to-convert-audio-clip-to-spectrogram \n.
", "messages": [ { "id": "907993", "postDate": "06/30/2020 09:55:11", "content": "As I am new to this topic also, I hope to share any insights that I learn along the way. Advanced practitioners may not find my insight useful, but your feedback is always appreciated!
\n\nHere is a resource to learn how to convert an audio file to a spectogram - https://www.kaggle.com/msripooja/steps-to-convert-audio-clip-to-spectrogram \n.
", "rawMarkdown": "As I am new to this topic also, I hope to share any insights that I learn along the way. Advanced practitioners may not find my insight useful, but your feedback is always appreciated! \n\nHere is a resource to learn how to convert an audio file to a spectogram - [https://www.kaggle.com/msripooja/steps-to-convert-audio-clip-to-spectrogram \n](https://www.kaggle.com/msripooja/steps-to-convert-audio-clip-to-spectrogram).", "votes": null }, { "id": "908067", "postDate": "06/30/2020 10:57:04", "content": "Nice notebook, I was really confused in this part.
", "rawMarkdown": "Nice notebook, I was really confused in this part.", "votes": null }, { "id": "908193", "postDate": "06/30/2020 12:42:54", "content": "Nice notebook. Just curious, what is the motivation behind taking log transform of a spectrogram?
", "rawMarkdown": "Nice notebook. Just curious, what is the motivation behind taking log transform of a spectrogram?", "votes": null }, { "id": "908215", "postDate": "06/30/2020 12:54:26", "content": "
The important information that is required is towards the bottom part of the spectrogram, but when we take the log...
\n\n![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F4133984%2Fe87be0c5903b6366c10122ca6a4bb3ff%2Fspec_B.png?generation=1593521604683113&alt=media\")
We now have the important information that we need... Now this is what my intuition tells me, it would be worthwhile double checking as I am also (I will edit this to confirm).
", "rawMarkdown": "\n\nThe important information that is required is towards the bottom part of the spectrogram, but when we take the log... \n\n\n\nWe now have the important information that we need... Now this is what my intuition tells me, it would be worthwhile double checking as I am also (I will edit this to confirm).", "votes": null }, { "id": "908244", "postDate": "06/30/2020 13:13:20", "content": "Oops missed out this part. Thanks!
", "rawMarkdown": "Oops missed out this part. Thanks!", "votes": null }, { "id": "908256", "postDate": "06/30/2020 13:23:26", "content": "Yes, it is correct, we are more interested in what happens in the lower part because a human can hear from 20Hz to 20000Hz but the most important and sensitive frequencies are in the range of 1000-5000Hz, and in this specific case we also know that most of the birds sing in the range of 1000 and 8000Hz so we would like to stretch the frequency scale in the lower part.\nAnother reason is that you can see the higher dBs are in the lower part (a human can usually start listening from 0 dBs), and from the cbar we deduce that most of the information is contained in lower area.
", "rawMarkdown": "Yes, it is correct, we are more interested in what happens in the lower part because a human can hear from 20Hz to 20000Hz but the most important and sensitive frequencies are in the range of 1000-5000Hz, and in this specific case we also know that most of the birds sing in the range of 1000 and 8000Hz so we would like to stretch the frequency scale in the lower part.\nAnother reason is that you can see the higher dBs are in the lower part (a human can usually start listening from 0 dBs), and from the cbar we deduce that most of the information is contained in lower area.", "votes": null }, { "id": "909546", "postDate": "06/30/2020 16:42:43", "content": "Thank you for clearing that up!
", "rawMarkdown": "Thank you for clearing that up!", "votes": null }, { "id": "910600", "postDate": "07/01/2020 08:33:24", "content": "good stater 🤓
", "rawMarkdown": "good stater 🤓", "votes": null } ], "comments": [ { "id": 908067, "author_name": "yash612", "author_url": "", "post_date": "06/30/2020 10:57:04", "content": "Nice notebook, I was really confused in this part.
", "votes": null, "replies": [] }, { "id": 908193, "author_name": "alanchn31", "author_url": "", "post_date": "06/30/2020 12:42:54", "content": "Nice notebook. Just curious, what is the motivation behind taking log transform of a spectrogram?
", "votes": null, "replies": [ { "id": 908215, "author_name": "kurtispykes", "author_url": "", "post_date": "06/30/2020 12:54:26", "content": "
The important information that is required is towards the bottom part of the spectrogram, but when we take the log...
\n\n
We now have the important information that we need... Now this is what my intuition tells me, it would be worthwhile double checking as I am also (I will edit this to confirm).
", "votes": null, "replies": [] }, { "id": 908244, "author_name": "alanchn31", "author_url": "", "post_date": "06/30/2020 13:13:20", "content": "Oops missed out this part. Thanks!
", "votes": null, "replies": [] }, { "id": 908256, "author_name": "pranavkasela", "author_url": "", "post_date": "06/30/2020 13:23:26", "content": "Yes, it is correct, we are more interested in what happens in the lower part because a human can hear from 20Hz to 20000Hz but the most important and sensitive frequencies are in the range of 1000-5000Hz, and in this specific case we also know that most of the birds sing in the range of 1000 and 8000Hz so we would like to stretch the frequency scale in the lower part.\nAnother reason is that you can see the higher dBs are in the lower part (a human can usually start listening from 0 dBs), and from the cbar we deduce that most of the information is contained in lower area.
", "votes": null, "replies": [] }, { "id": 909546, "author_name": "kurtispykes", "author_url": "", "post_date": "06/30/2020 16:42:43", "content": "Thank you for clearing that up!
", "votes": null, "replies": [] } ] }, { "id": 910600, "author_name": "deepakdeepu8978", "author_url": "", "post_date": "07/01/2020 08:33:24", "content": "good stater 🤓
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "907993": "As I am new to this topic also, I hope to share any insights that I learn along the way. Advanced practitioners may not find my insight useful, but your feedback is always appreciated! \n\nHere is a resource to learn how to convert an audio file to a spectogram - [https://www.kaggle.com/msripooja/steps-to-convert-audio-clip-to-spectrogram \n](https://www.kaggle.com/msripooja/steps-to-convert-audio-clip-to-spectrogram).", "908067": "Nice notebook, I was really confused in this part.", "908193": "Nice notebook. Just curious, what is the motivation behind taking log transform of a spectrogram?", "908215": "\n\nThe important information that is required is towards the bottom part of the spectrogram, but when we take the log... \n\n\n\nWe now have the important information that we need... Now this is what my intuition tells me, it would be worthwhile double checking as I am also (I will edit this to confirm).", "908244": "Oops missed out this part. Thanks!", "908256": "Yes, it is correct, we are more interested in what happens in the lower part because a human can hear from 20Hz to 20000Hz but the most important and sensitive frequencies are in the range of 1000-5000Hz, and in this specific case we also know that most of the birds sing in the range of 1000 and 8000Hz so we would like to stretch the frequency scale in the lower part.\nAnother reason is that you can see the higher dBs are in the lower part (a human can usually start listening from 0 dBs), and from the cbar we deduce that most of the information is contained in lower area.", "909546": "Thank you for clearing that up!", "910600": "good stater 🤓" }, "source": "meta" }