| | --- |
| | license: cc-by-nc-4.0 |
| | task_categories: |
| | - audio-classification |
| | language: |
| | - en |
| | - da |
| | tags: |
| | - music |
| | pretty_name: Instrument_Classification |
| | size_categories: |
| | - n<1K |
| | --- |
| | |
| | # Dataset Card for Dataset Name |
| |
|
| | This dataset is a collection of audioclips used for training machinelearning models on instrument based sound classification. |
| |
|
| |
|
| | ## Dataset Details |
| |
|
| | ### Dataset Description |
| |
|
| |
|
| |
|
| | - **Curated by:** [AlfredVThor] |
| | - **Funded by [AlfredVThor]:** [More Information Needed] |
| | - **Shared by [AlfredVThor]:** [More Information Needed] |
| | - **Language(s) (NLP):** [More Information Needed] |
| | - **License:** [More Information Needed] |
| |
|
| | ## Uses |
| | This dataset can be used for training sound classification models on basic instruments |
| |
|
| | ### Direct Use |
| |
|
| | The Direct use of this dataset is to train a model on Sound Classification data based on instruments, using it to help those with Hearing Impairment identifying instruments. |
| |
|
| |
|
| | ### Out-of-Scope Use |
| |
|
| | This dataset will not be usefull for anything other than sound based models |
| |
|
| |
|
| | ## Dataset Structure |
| |
|
| | Clap |
| | 89% / 11% (5m 10s / 38s) |
| | Guitar |
| | 75% / 25% (42s / 14s) |
| | Kling |
| | 80% / 20% (2m 0s / 30s) |
| | Maracas |
| | 75% / 25% (3m 31s / 1m 11s) |
| | Noise |
| | 90% / 10% (8m 54s / 1m 0s) |
| | Snap |
| | 83% / 17% (4m 21s / 52s) |
| | Whistle |
| | 85% / 15% (1m 51s / 19s) |
| |
|
| | ## Dataset Creation |
| |
|
| | ### Curation Rationale |
| |
|
| | This dataset it aimed at training models to help people with impaired hearing identify instruments based on sound in a public space. |
| |
|
| | ### Source Data |
| |
|
| | .Wav audio files |
| |
|
| | #### Data Collection and Processing |
| |
|
| | Data was collected with a mobile microphone to best simulate the usecase for users. Data was cleaned as to minimally include high volume background noise. |
| |
|
| | #### Who are the source data producers? |
| |
|
| | The data was collected by students at Medialogy Masters at AAU copenhagen |
| |
|
| | #### Annotation process |
| |
|
| | The process included data cleaning, removing data that was unclear or duplicated. |
| |
|
| | #### Who are the annotators? |
| |
|
| | Student at Medialogy Masters AAU Copenhagen. |
| |
|
| | #### Personal and Sensitive Information |
| |
|
| | This dataset does not include private or sensitive data. |
| |
|
| | ## Bias, Risks, and Limitations |
| |
|
| | Cannot be used for anything non-soundbased, Additional data can be recomended for higher scale models. |
| |
|
| |
|
| | ### Recommendations |
| |
|
| | Additional data can be recomended for higher scale models. |
| |
|
| | **BibTeX:** |
| |
|
| | @misc{key, |
| | author = {AlfredVThor}, |
| | title = {Instrument_Classification}, |
| | howpublished = {\url{https://huggingface.co/datasets/Athorl22/MLME_Instrument_Classification/edit/main/README.md}}, |
| | year = {2025}, |
| | note = {[Accessed 05-12-2025]}, |
| | } |
| | |
| | **APA:** |
| |
|
| | Hugging Face – The AI community building the future. (n.d.). https://huggingface.co/datasets/Athorl22/MLME_Instrument_Classification/edit/main/README.md |
| |
|
| | ## Dataset Card Contact |
| | Github: AlfredVThor |
| |
|
