Datasets:

---

DBD-research-group

/

BirdSet

like 24

Follow

DBD-research-group 35

Hello!

I have found some inconsistencies between what the README says about data licensing, and what the individual recordings say.

For the XCL train split I'm seeing this distribution:

License                                           Clips
-------------------------------------------------------
nan                                              122531
//creativecommons.org/licenses/by-nc-sa/4.0/       8439
//creativecommons.org/publicdomain/zero/1.0/         74
//creativecommons.org/licenses/by-sa/4.0/            47
//creativecommons.org/licenses/by-nc-sa/3.0/         13
//creativecommons.org/licenses/by/4.0/               10
//creativecommons.org/licenses/by-nc/4.0/             7
TOTAL                                            131121

For all cumulated test_5s splits I'm seeing this distribution:

License                                                            Clips
------------------------------------------------------------------------
Creative Commons Attribution 4.0 International Public License     101685
Creative Commons Zero v1.0 Universal                                4414
None                                                                  62
TOTAL                                                             106161

1. Can you clarify where the nan and missing licenses (= None) are coming from?
2. The README says:

   Each recording in the training datasets has a unique recordist and the corresponding license from XC. We omit all recordings from XC that are CC-ND.

   The lack of a license (i.e. nan or None) is a stronger restriction than CC-ND. Can you confirm that all training recordings can be used in a non-commercial way?
3. The README says:

   We exclude all recordings with a SA licenses. Every recording is NC.

   I'm still seeing recordings with CC-BY-SA and CC-BY-NC-SA. Is this an inconsistency, and those recordings were supposed to be filtered out in the first place?
4. The README says:

   Each test dataset is licensed under CC BY 4.0.

   Please clarify the missing licenses (= None) from the test_5s split.

Thanks,
Ștefan