└── README.md


/README.md:
--------------------------------------------------------------------------------
 1 | # **Handy ASR noise dataset**
 2 | 
 3 | A handy dataset for noise augmentations for ASR / TTS:
 4 | - ~20k noise files;
 5 | - ~200 distinct categories;
 6 | 
 7 | ![](https://pics.spark-in.me/upload/c65cd3ef082bc035da000300e5b83eab.png)
 8 | 
 9 | Contact [us](mailto:open_stt@googlegroups.com)!
10 | Open issues, collaborate, submit a PR, contribute, share your datasets!
11 | 
12 | ## **Contribution ideas**
13 | 
14 | Add much more data from BBC Sound Effects dataset.
15 | 
16 | # **Download links**
17 | 
18 | Meta data [file](https://asr-noise.fra1.digitaloceanspaces.com/noises_df.feather) / 2.0M / `73cb528656a484b20e02d6c5fd05f14c`
19 | Noise archive [file](https://asr-noise.fra1.digitaloceanspaces.com/asr_noises.tar.gz) / 4.7G / `5e069c867a0da891f57616905129b6c3`
20 | 
21 | **Open feather file:**
22 | ```
23 | import pandas as pd
24 | 
25 | df = pd.read_feather(file_path)
26 | ```
27 | 
28 | # **Data preparation**
29 | 
30 | The dataset is compiled using open domain sources.
31 | All labels resembling loud human speech were removed (but background noise, i.e. street chatter, was not removed).
32 | All of the items are 0 - 60 seconds long.
33 | 
34 | ![](https://pics.spark-in.me/upload/f935c54efed15bd40f1262d29b5dbbad.png)
35 | 
36 | All files are normalized as follows:
37 | - Converted to mono, if necessary;
38 | - Converted to 16 kHz sampling rate, if necessary;
39 | - Stored as 16-bit integers;
40 | 
41 | # **Contacts**
42 | 
43 | Please contact us [here](mailto:open_stt@googlegroups.com) or just create a GitHub issue!
44 | 
45 | # **License**
46 | cc-by
47 | 
48 | # **References / citations / licenses**
49 | 
50 | **Links / license**
51 | - [rnnoise](https://people.xiph.org/~jm/demo/rnnoise/) / [CC0](https://creativecommons.org/publicdomain/zero/1.0/);
52 | - [acoustic events](https://data.vision.ee.ethz.ch/cvl/ae_dataset/) / `if you end up using the dataset, we ask you to cite the following paper`;
53 | - [urban sounds](urbansounddataset.weebly.com/urbansound8k.html) / [cc-by-nc](http://creativecommons.org/licenses/by-nc/3.0/);
54 | - [esc-50](https://github.com/karoldvl/ESC-50) / [license](https://github.com/karoldvl/ESC-50/blob/master/LICENSE) (cc-by-nc);
55 | - [freiburg-106](http://www.csc.kth.se/~jastork/pages/datasets.html) / ?;
56 | - [sound-events](https://www.sciencedirect.com/science/article/abs/pii/S0167865515002925) / ?;
57 | - [BBC Sound Effects](http://bbcsfx.acropolis.org.uk/) (a small part) / [license](https://github.com/bbcarchdev/Remarc/blob/master/doc/2016.09.27_RemArc_Content%20licence_Terms%20of%20Use_final.pdf);
58 | - [nar dataset](https://team.inria.fr/perception/nard/) / `the data are freely accessible for scientific research purposes and for non-commercial applications`
59 | 
60 | **Paper citations:**
61 | -  Naoya Takahashi, Michael Gygli, Beat Pfister and Luc Van Gool,"Deep Convolutional Neural Networks and Data Augmentation for Acoustic Event Recognition", Proc. Interspeech 2016, San Fransisco;
62 | - J. Salamon, C. Jacoby and J. P. Bello, "A Dataset and Taxonomy for Urban Sound Research", 22nd ACM International Conference on Multimedia, Orlando USA, Nov. 2014;
63 | 
64 | 
65 | # **Donations**
66 | 
67 | [Donate](https://buymeacoff.ee/8oneCIN) (each coffee pays for several full downloads) / use our DO referral [link](https://sohabr.net/habr/post/357748/) to help.
68 | 
69 | 


--------------------------------------------------------------------------------