upl base

.gitignore

@@ -0,0 +1,5 @@
+rename.sh
+test.py
+*.wav
+*.txt
+*.jpg

README.md

@@ -1,3 +1,131 @@
 ---
 license: mit
+task_categories:
+- audio-classification
+language:
+- zh
+- en
+tags:
+- music
+- art
+pretty_name: Timbre and Range Dataset
+size_categories:
+- 1K<n<10K
+viewer: false
 ---
+
+# Dataset Card for Timbre and Range Dataset
+## Dataset Summary
+The timbre dataset contains acapella singing audio of 9 singers, as well as cut single-note audio, totaling 775 clips (.wav format)
+The vocal range dataset includes several up and down chromatic scales audio clips of several vocals, as well as the cut single-note audio clips (.wav format).
+
+### Supported Tasks and Leaderboards
+Audio classification
+
+### Languages
+Chinese, English
+
+## Dataset Structure
+<https://www.modelscope.cn/datasets/ccmusic-database/timbre_range/dataPeview>
+### Data Instances
+.zip(.wav, .jpg), .csv
+
+### Data Fields
+```txt
+timbre: song1-32
+range: vox1_19-22/26-29/32/33/36-38/41-47/51-55/59-64/69-71/79-81
+```
+
+### Data Splits
+Train, Validation, Test
+
+## Usage
+### Timbre subset
+```python
+from datasets import load_dataset
+
+ds = load_dataset("ccmusic-database/timbre_range", name="timbre")
+for item in ds["train"]:
+    print(item)
+
+for item in ds["validation"]:
+    print(item)
+
+for item in ds["test"]:
+    print(item)
+```
+
+### Range subset
+```python
+from datasets import load_dataset
+# default subset
+ds = load_dataset("ccmusic-database/timbre_range", name="range")
+for item in ds["train"]:
+    print(item)
+
+for item in ds["validation"]:
+    print(item)
+
+for item in ds["test"]:
+    print(item)
+```
+
+## Maintenance
+```bash
+git clone [email protected]:datasets/ccmusic-database/timbre_range
+cd timbre_range
+```
+
+## Dataset Creation
+### Curation Rationale
+Promoting the development of music AI industry
+
+### Source Data
+#### Initial Data Collection and Normalization
+Zijin Li, Zhaorui Liu, Monan Zhou
+
+#### Who are the source language producers?
+Composers of the songs in dataset
+
+### Annotations
+#### Annotation process
+CCMUSIC students collected acapella singing audios of 9 singers, as well as cut single-note audio, totaling 775 clips
+
+#### Who are the annotators?
+Students from CCMUSIC
+
+### Personal and Sensitive Information
+Due to copyright issues with the original music, only acapella singing audios are provided in the dataset
+
+## Considerations for Using the Data
+### Social Impact of Dataset
+Promoting the development of AI in the music industry
+
+### Discussion of Biases
+Most are Chinese songs
+
+### Other Known Limitations
+Samples are not balanced enough
+
+## Mirror
+<https://www.modelscope.cn/datasets/ccmusic-database/timbre_range>
+
+## Additional Information
+### Dataset Curators
+Zijin Li
+
+### Evaluation
+[1] [Yiliang, J. et al. (2019) ‘Data Augmentation based Convolutional Neural Network for Auscultation’, Journal of Fudan University(Natural Science), pp. 328–334. doi:10.15943/j.cnki.fdxb-jns.2019.03.004.](https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CJFD&dbname=CJFDLAST2019&filename=FDXB201903004&uniplatform=NZKPT&v=VAszHDtjPUYMi3JYVrdSGx4fcqlEtgCeKwRGTacCj98CGEQg5CUFHxakrvuaMzm3)
+
+### Citation Information
+```bibtex
+@article{2019Data,
+  title={Data Augmentation based Convolutional Neural Network for Auscultation},
+  author={ Yiliang, Jiang  and  Xulong, Zhang  and  Jin, Deng  and  Wenqiang, Zhang  and  Wei, L. I. },
+  journal={Journal of Fudan University(Natural Science)},
+  year={2019},
+}
+```
+
+### Contributions
+Provide a dataset for music timbre and range

timbre_range.py

@@ -0,0 +1,107 @@
+import os
+import random
+import datasets
+import pandas as pd
+from datasets.tasks import AudioClassification
+
+
+_NAMES = {
+    "timbre": ["Base", "Split", "Short"],
+    "range": ["Narrow", "Moderate", "Wide"],
+}
+
+_URL = f"https://www.modelscope.cn/datasets/ccmusic-database/{os.path.basename(__file__)[:-3]}/resolve/master/data"
+
+
+class timbre_range(datasets.GeneratorBasedBuilder):
+    def _info(self):
+        if self.config.name == "default":
+            self.config.name = "range"
+
+        return datasets.DatasetInfo(
+            features=datasets.Features(
+                {
+                    "audio": datasets.Audio(sampling_rate=44_100),
+                    "mel": datasets.Image(),
+                    "label": datasets.features.ClassLabel(names=_NAMES["timbre"]),
+                    "score1": datasets.Value("float64"),
+                    "score2": datasets.Value("float64"),
+                    "avg_score": datasets.Value("float64"),
+                }
+                if self.config.name == "timbre"
+                else {
+                    "audio": datasets.Audio(sampling_rate=44_100),
+                    "mel": datasets.Image(),
+                    "label": datasets.features.ClassLabel(names=_NAMES["range"]),
+                }
+            ),
+            supervised_keys=("audio", "label"),
+            license="CC-BY-NC-ND",
+            version="1.2.0",
+            task_templates=[
+                AudioClassification(
+                    task="audio-classification",
+                    audio_column="audio",
+                    label_column="label",
+                )
+            ],
+        )
+
+    def _split_generators(self, dl_manager):
+        dataset = []
+        data_files = dl_manager.download_and_extract(f"{_URL}/{self.config.name}.zip")
+        for fpath in dl_manager.iter_files([data_files]):
+            if os.path.basename(fpath).endswith(".wav"):
+                dataset.append(fpath)
+
+        random.shuffle(dataset)
+        count = len(dataset)
+        p80 = int(0.8 * count)
+        p90 = int(0.9 * count)
+
+        csv_file = dl_manager.download(f"{_URL}/{self.config.name}.csv")
+        labels = pd.read_csv(csv_file, index_col="id")
+
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={"files": dataset[:p80], "labels": labels},
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.VALIDATION,
+                gen_kwargs={"files": dataset[p80:p90], "labels": labels},
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
+                gen_kwargs={"files": dataset[p90:], "labels": labels},
+            ),
+        ]
+
+    def _generate_examples(self, files, labels):
+        if self.config.name == "timbre":
+            for i, fpath in enumerate(files):
+                id: str = os.path.basename(fpath)[:-4]
+                if "-" in id:
+                    id = id.split("-")[0]
+
+                yield i, {
+                    "audio": fpath,
+                    "mel": fpath.replace("/audios/", "/images/")
+                    .replace("\\audios\\", "\\images\\")
+                    .replace(".wav", ".jpg"),
+                    "label": os.path.basename(os.path.dirname(fpath)).capitalize(),
+                    "score1": float(labels.loc[id]["score1"]),
+                    "score2": float(labels.loc[id]["score2"]),
+                    "avg_score": float(labels.loc[id]["avg_score"]),
+                }
+
+        else:
+            for i, fpath in enumerate(files):
+                id = os.path.basename(fpath)[:-4]
+                yield i, {
+                    "audio": fpath,
+                    "mel": fpath.replace("/audios/", "/images/")
+                    .replace("\\audios\\", "\\images\\")
+                    .replace(".wav", ".jpg"),
+                    "label": _NAMES["range"][int(labels.loc[id]["range"])],
+                }