{
  "id": 570577,
  "title": "BirdCLEF+ 2025:Papers on audio augmentation.",
  "url": "/competitions/birdclef-2025/discussion/570577",
  "author_name": "",
  "post_date": "2025-03-28T23:08:30.439707Z",
  "votes": 28,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hello everyone,</p>\n<p>I would like to share some research papers on audio data augmentation. I will continue to update this post as I find new or relevant papers. If you have any suggestions or know of other valuable resources, please let me know.</p>\n<h1><a href=\"https://www.researchgate.net/publication/381758861_Improving_Learning-Based_Birdsong_Classification_by_Utilizing_Combined_Audio_Augmentation_Strategies\" target=\"_blank\"><strong>Improving Learning-Based Birdsong Classification by Utilizing Combined Audio Augmentation Strategies</strong></a></h1>\n<p>code:</p>\n<h1><a href=\"https://arxiv.org/abs/1904.08779\" target=\"_blank\">SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition</a></h1>\n<p>code:&nbsp;<a href=\"https://github.com/DemisEom/SpecAugment\" target=\"_blank\">https://github.com/DemisEom/SpecAugment</a></p>\n<h1><a href=\"https://arxiv.org/abs/2107.03649\" target=\"_blank\">Heavily Augmented Sound Event Detection utilizing Weak Predictions</a></h1>\n<p>code:</p>\n<h1><a href=\"https://arxiv.org/abs/2110.03282\" target=\"_blank\">FilterAugment: An Acoustic Environmental Data Augmentation Method</a></h1>\n<p>code:<a href=\"https://github.com/frednam93/FilterAugSED\" target=\"_blank\">https://github.com/frednam93/FilterAugSED</a></p>\n<h1><a href=\"https://arxiv.org/abs/2103.16858\" target=\"_blank\">SpecAugment++: A Hidden Space Data Augmentation Method for Acoustic Scene Classification</a></h1>\n<p>code:<a href=\"https://github.com/WangHelin1997/SpecAugment-plus\" target=\"_blank\">https://github.com/WangHelin1997/SpecAugment-plus</a></p>\n<h1><a href=\"https://arxiv.org/abs/2108.03020\" target=\"_blank\">SpecMix : A Mixed Sample Data Augmentation method for Training withTime-Frequency Domain Features</a></h1>\n<p>code:<a href=\"https://github.com/anas-rz/specmix-pytorch\" target=\"_blank\">https://github.com/anas-rz/specmix-pytorch</a></p>\n<h1><a href=\"https://research.samsung.com/blog/RandMasking-Augment-A-Simple-and-Randomized-Data-Augmentation-for-Acoustic-Scene-Classification\" target=\"_blank\">RandMasking Augment: A Simple and Randomized Data Augmentation for Acoustic Scene Classification</a></h1>\n<p>code:</p>\n<h1><a href=\"https://arxiv.org/pdf/2111.04433\" target=\"_blank\">RAWBOOST: A RAW DATA BOOSTING AND AUGMENTATION METHOD APPLIED TO AUTOMATIC SPEAKER VERIFICATION ANTI-SPOOFING</a></h1>\n<p>code:<a href=\"https://github.com/TakHemlata/RawBoost-antispoofing\" target=\"_blank\">https://github.com/TakHemlata/RawBoost-antispoofing</a></p>\n<h1><a href=\"https://arxiv.org/pdf/2110.00046\" target=\"_blank\">SPLICEOUT: A Simple and Efficient Audio Augmentation Method</a></h1>\n<p>code:</p>\n<h1><a href=\"https://arxiv.org/pdf/2410.18322\" target=\"_blank\">MIXSTYLE BASED DOMAIN GENERALIZATION FOR SOUND EVENT DETECTION WITH HETEROGENEOUS TRAINING DATA</a></h1>\n<p>code:</p>",
  "messages": [
    {
      "id": "3162200",
      "postDate": "03/28/2025 23:08:30",
      "content": "<p>Hello everyone,</p>\n<p>I would like to share some research papers on audio data augmentation. I will continue to update this post as I find new or relevant papers. If you have any suggestions or know of other valuable resources, please let me know.</p>\n<h1><a href=\"https://www.researchgate.net/publication/381758861_Improving_Learning-Based_Birdsong_Classification_by_Utilizing_Combined_Audio_Augmentation_Strategies\" target=\"_blank\"><strong>Improving Learning-Based Birdsong Classification by Utilizing Combined Audio Augmentation Strategies</strong></a></h1>\n<p>code:</p>\n<h1><a href=\"https://arxiv.org/abs/1904.08779\" target=\"_blank\">SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition</a></h1>\n<p>code:&nbsp;<a href=\"https://github.com/DemisEom/SpecAugment\" target=\"_blank\">https://github.com/DemisEom/SpecAugment</a></p>\n<h1><a href=\"https://arxiv.org/abs/2107.03649\" target=\"_blank\">Heavily Augmented Sound Event Detection utilizing Weak Predictions</a></h1>\n<p>code:</p>\n<h1><a href=\"https://arxiv.org/abs/2110.03282\" target=\"_blank\">FilterAugment: An Acoustic Environmental Data Augmentation Method</a></h1>\n<p>code:<a href=\"https://github.com/frednam93/FilterAugSED\" target=\"_blank\">https://github.com/frednam93/FilterAugSED</a></p>\n<h1><a href=\"https://arxiv.org/abs/2103.16858\" target=\"_blank\">SpecAugment++: A Hidden Space Data Augmentation Method for Acoustic Scene Classification</a></h1>\n<p>code:<a href=\"https://github.com/WangHelin1997/SpecAugment-plus\" target=\"_blank\">https://github.com/WangHelin1997/SpecAugment-plus</a></p>\n<h1><a href=\"https://arxiv.org/abs/2108.03020\" target=\"_blank\">SpecMix : A Mixed Sample Data Augmentation method for Training withTime-Frequency Domain Features</a></h1>\n<p>code:<a href=\"https://github.com/anas-rz/specmix-pytorch\" target=\"_blank\">https://github.com/anas-rz/specmix-pytorch</a></p>\n<h1><a href=\"https://research.samsung.com/blog/RandMasking-Augment-A-Simple-and-Randomized-Data-Augmentation-for-Acoustic-Scene-Classification\" target=\"_blank\">RandMasking Augment: A Simple and Randomized Data Augmentation for Acoustic Scene Classification</a></h1>\n<p>code:</p>\n<h1><a href=\"https://arxiv.org/pdf/2111.04433\" target=\"_blank\">RAWBOOST: A RAW DATA BOOSTING AND AUGMENTATION METHOD APPLIED TO AUTOMATIC SPEAKER VERIFICATION ANTI-SPOOFING</a></h1>\n<p>code:<a href=\"https://github.com/TakHemlata/RawBoost-antispoofing\" target=\"_blank\">https://github.com/TakHemlata/RawBoost-antispoofing</a></p>\n<h1><a href=\"https://arxiv.org/pdf/2110.00046\" target=\"_blank\">SPLICEOUT: A Simple and Efficient Audio Augmentation Method</a></h1>\n<p>code:</p>\n<h1><a href=\"https://arxiv.org/pdf/2410.18322\" target=\"_blank\">MIXSTYLE BASED DOMAIN GENERALIZATION FOR SOUND EVENT DETECTION WITH HETEROGENEOUS TRAINING DATA</a></h1>\n<p>code:</p>",
      "rawMarkdown": "Hello everyone,\n\nI would like to share some research papers on audio data augmentation. I will continue to update this post as I find new or relevant papers. If you have any suggestions or know of other valuable resources, please let me know.\n\n# [**Improving Learning-Based Birdsong Classification by Utilizing Combined Audio Augmentation Strategies**](https://www.researchgate.net/publication/381758861_Improving_Learning-Based_Birdsong_Classification_by_Utilizing_Combined_Audio_Augmentation_Strategies)\ncode:\n\n# [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://arxiv.org/abs/1904.08779)\ncode: [https://github.com/DemisEom/SpecAugment](https://github.com/DemisEom/SpecAugment)\n\n# [Heavily Augmented Sound Event Detection utilizing Weak Predictions](https://arxiv.org/abs/2107.03649)\ncode:\n\n# [FilterAugment: An Acoustic Environmental Data Augmentation Method](https://arxiv.org/abs/2110.03282)\ncode:[https://github.com/frednam93/FilterAugSED](https://github.com/frednam93/FilterAugSED)\n\n# [SpecAugment++: A Hidden Space Data Augmentation Method for Acoustic Scene Classification](https://arxiv.org/abs/2103.16858)\ncode:[https://github.com/WangHelin1997/SpecAugment-plus](https://github.com/WangHelin1997/SpecAugment-plus)\n\n# [SpecMix : A Mixed Sample Data Augmentation method for Training withTime-Frequency Domain Features](https://arxiv.org/abs/2108.03020)\ncode:[https://github.com/anas-rz/specmix-pytorch](https://github.com/anas-rz/specmix-pytorch)\n\n# [RandMasking Augment: A Simple and Randomized Data Augmentation for Acoustic Scene Classification](https://research.samsung.com/blog/RandMasking-Augment-A-Simple-and-Randomized-Data-Augmentation-for-Acoustic-Scene-Classification)\ncode:\n\n# [RAWBOOST: A RAW DATA BOOSTING AND AUGMENTATION METHOD APPLIED TO AUTOMATIC SPEAKER VERIFICATION ANTI-SPOOFING](https://arxiv.org/pdf/2111.04433)\ncode:[https://github.com/TakHemlata/RawBoost-antispoofing](https://github.com/TakHemlata/RawBoost-antispoofing)\n\n# [SPLICEOUT: A Simple and Efficient Audio Augmentation Method](https://arxiv.org/pdf/2110.00046)\ncode:\n\n# [MIXSTYLE BASED DOMAIN GENERALIZATION FOR SOUND EVENT DETECTION WITH HETEROGENEOUS TRAINING DATA](https://arxiv.org/pdf/2410.18322)\ncode:",
      "votes": null
    },
    {
      "id": "3206467",
      "postDate": "05/21/2025 11:05:17",
      "content": "<p>SpecAugment and FilterAugment both improved the LB score.</p>",
      "rawMarkdown": "SpecAugment and FilterAugment both improved the LB score.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3206467,
      "author_name": "myso1987",
      "author_url": "",
      "post_date": "05/21/2025 11:05:17",
      "content": "<p>SpecAugment and FilterAugment both improved the LB score.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3162200": "Hello everyone,\n\nI would like to share some research papers on audio data augmentation. I will continue to update this post as I find new or relevant papers. If you have any suggestions or know of other valuable resources, please let me know.\n\n# [**Improving Learning-Based Birdsong Classification by Utilizing Combined Audio Augmentation Strategies**](https://www.researchgate.net/publication/381758861_Improving_Learning-Based_Birdsong_Classification_by_Utilizing_Combined_Audio_Augmentation_Strategies)\ncode:\n\n# [SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition](https://arxiv.org/abs/1904.08779)\ncode: [https://github.com/DemisEom/SpecAugment](https://github.com/DemisEom/SpecAugment)\n\n# [Heavily Augmented Sound Event Detection utilizing Weak Predictions](https://arxiv.org/abs/2107.03649)\ncode:\n\n# [FilterAugment: An Acoustic Environmental Data Augmentation Method](https://arxiv.org/abs/2110.03282)\ncode:[https://github.com/frednam93/FilterAugSED](https://github.com/frednam93/FilterAugSED)\n\n# [SpecAugment++: A Hidden Space Data Augmentation Method for Acoustic Scene Classification](https://arxiv.org/abs/2103.16858)\ncode:[https://github.com/WangHelin1997/SpecAugment-plus](https://github.com/WangHelin1997/SpecAugment-plus)\n\n# [SpecMix : A Mixed Sample Data Augmentation method for Training withTime-Frequency Domain Features](https://arxiv.org/abs/2108.03020)\ncode:[https://github.com/anas-rz/specmix-pytorch](https://github.com/anas-rz/specmix-pytorch)\n\n# [RandMasking Augment: A Simple and Randomized Data Augmentation for Acoustic Scene Classification](https://research.samsung.com/blog/RandMasking-Augment-A-Simple-and-Randomized-Data-Augmentation-for-Acoustic-Scene-Classification)\ncode:\n\n# [RAWBOOST: A RAW DATA BOOSTING AND AUGMENTATION METHOD APPLIED TO AUTOMATIC SPEAKER VERIFICATION ANTI-SPOOFING](https://arxiv.org/pdf/2111.04433)\ncode:[https://github.com/TakHemlata/RawBoost-antispoofing](https://github.com/TakHemlata/RawBoost-antispoofing)\n\n# [SPLICEOUT: A Simple and Efficient Audio Augmentation Method](https://arxiv.org/pdf/2110.00046)\ncode:\n\n# [MIXSTYLE BASED DOMAIN GENERALIZATION FOR SOUND EVENT DETECTION WITH HETEROGENEOUS TRAINING DATA](https://arxiv.org/pdf/2410.18322)\ncode:",
    "3206467": "SpecAugment and FilterAugment both improved the LB score."
  },
  "source": "meta"
}