---
dataset_info:
  features:
  - name: index
    dtype: int64
  - name: audio
    dtype:
      audio:
        sampling_rate: 16000
        decode: false
  - name: subset
    dtype: string
  - name: speaker
    dtype: string
  - name: label
    dtype: string
  - name: original_name
    dtype: string
  splits:
  - name: train
    num_bytes: 24343526.0
    num_examples: 887
  download_size: 22452939
  dataset_size: 24343526.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: cc-by-4.0
tags:
- audio
- animal-vocalization
- birdsong
- zebra-finch
- perceptual-similarity
- benchmark
- zero-shot
- vocsim
- avian-perceptual-judgment
- audio-perceptual-judgment
size_categories:
- n<1K
pretty_name: VocSim - Avian Perception Alignment
---

# Dataset Card for VocSim - Avian Perception Alignment

## Dataset Description

This dataset is used in the **VocSim benchmark** paper, specifically designed to evaluate how well neural audio embeddings align with biological perceptual judgments of similarity. It utilizes data from **Zandberg et al. (2024)**, which includes recordings of zebra finch (*Taeniopygia guttata*) song syllables and results from behavioral experiments (probe and triplet tasks) measuring the birds' perception of syllable similarity.

The dataset allows researchers to:
1.  Extract features/embeddings from the song syllables using various computational models.
2.  Compute pairwise distances between these embeddings.
3.  Compare the resulting computational similarity matrices against the avian perceptual judgments recorded in the accompanying `probes.csv` and `triplets.csv` files.

This facilitates the development and benchmarking of audio representations that better capture biologically relevant acoustic features.

**Included Files:**
*   Hugging Face `Dataset` object containing audio file paths and metadata.
*   `probes.csv`: Contains results from perceptual probe trials (sound_id, left, right, decision, etc.). Filtered to include only rows where all mentioned audio files exist.
*   `triplets.csv`: Contains results from perceptual triplet trials (Anchor, Positive, Negative, diff, etc.). Filtered to include only rows where all mentioned audio files exist.
*   `missing_audio_files.txt` (optional): Lists identifiers from the original CSVs for which no corresponding audio file was found.

## Dataset Structure

### Data Instances

A typical example in the dataset looks like this:

```python
{
  'audio': {'path': '/path/to/datasets/avian_perception/wavs/ZF_M_123_syllable_A.wav', 'array': array([-0.00024414, -0.00048828, ...,  0.00024414], dtype=float32), 'sampling_rate': 16000},
  'subset': 'avian_perception',
  'index': 42,
  'speaker': 'ZF_M_123',
  'label': 'ZF_M_123', # Label is set to speaker ID for this dataset
  'original_name': 'ZF_M_123_syllable_A.wav' # Identifier as used in CSVs
}
```
## Citation Information

If you use this dataset in your work, please cite both the VocSim benchmark paper and the original source data paper:

```bib
@unpublished{vocsim2025,
  title={VocSim: Zero-Shot Audio Similarity Benchmark for Neural Embeddings},
  author={Anonymous},
  year={2025},
  note={Submitted manuscript}
}

@article{zandberg2024bird,
  author = {Zandberg, Lies and Morfi, Veronica and George, Julia M. and Clayton, David F. and Stowell, Dan and Lachlan, Robert F.},
  title = {Bird song comparison using deep learning trained from avian perceptual judgments},
  journal = {PLoS Computational Biology},
  volume = {20},
  number = {8},
  year = {2024},
  month = {aug},
  pages = {e1012329},
  doi = {10.1371/journal.pcbi.1012329},
  publisher = {Public Library of Science}
}
```