---
dataset_info:
  features:
  - name: id
    dtype: int64
  - name: audio
    dtype: audio
  - name: sentence
    dtype: string
  - name: lang
    dtype: string
  - name: style
    dtype: string
  - name: duration
    dtype: float64
  splits:
  - name: j7KV53NgP8U4LRS2k2Gs
    num_bytes: 426403347.067
    num_examples: 2717
  - name: EXAVITQu4vr4xnSDxMaL
    num_bytes: 768677383.349
    num_examples: 4899
  - name: AZnzlk1XvdvUeBnXmlld
    num_bytes: 1472171505.584
    num_examples: 9384
  - name: tnSpp4vdxKPjI9w0GnoV
    num_bytes: 1304806012.667
    num_examples: 8317
  - name: jBpfuIE2acCO8z3wKNLl
    num_bytes: 1472171505.584
    num_examples: 9384
  - name: tQ4MEZFJOzsahSEEZtHK
    num_bytes: 1472171505.584
    num_examples: 9384
  - name: cgSgspJ2msm6clMCkdW9
    num_bytes: 1472171505.584
    num_examples: 9384
  - name: ErXwobaYiN019PkySvjV
    num_bytes: 1471587074.702
    num_examples: 9383
  download_size: 6608582771
  dataset_size: 9860159840.120998
configs:
- config_name: default
  data_files:
  - split: j7KV53NgP8U4LRS2k2Gs
    path: data/j7KV53NgP8U4LRS2k2Gs-*
  - split: EXAVITQu4vr4xnSDxMaL
    path: data/EXAVITQu4vr4xnSDxMaL-*
  - split: AZnzlk1XvdvUeBnXmlld
    path: data/AZnzlk1XvdvUeBnXmlld-*
  - split: tnSpp4vdxKPjI9w0GnoV
    path: data/tnSpp4vdxKPjI9w0GnoV-*
  - split: jBpfuIE2acCO8z3wKNLl
    path: data/jBpfuIE2acCO8z3wKNLl-*
  - split: tQ4MEZFJOzsahSEEZtHK
    path: data/tQ4MEZFJOzsahSEEZtHK-*
  - split: cgSgspJ2msm6clMCkdW9
    path: data/cgSgspJ2msm6clMCkdW9-*
  - split: ErXwobaYiN019PkySvjV
    path: data/ErXwobaYiN019PkySvjV-*
license: cc-by-4.0
task_categories:
- text-to-speech
- automatic-speech-recognition
- audio-classification
language:
- ba
size_categories:
- 10K<n<100K
---

# Bashkort TTS Dataset

The **largest open dataset** for speech synthesis in the **Bashkir language** — featuring **multi-speaker recordings** and **speaking styles**.

## 📊 Dataset Overview

* **Total audio files:** 62,852
* **Speakers:** 7 female, 1 male
* **Speaking styles:** `friendly`, `question`, `neutral`
* **Languages:** Bashkir
* **Format:** MP3 audio + transcription text

## 🎙 How It Was Collected

1. **Initial recording:** A female voice actor recorded \~15 hours of speech in Bashkir.
2. **Voice cloning:** Using **ElevenLabs voice cloning**, 8 additional synthetic voices were generated — including a male voice.
3. **Multi-style dataset:** All voices were aligned with transcripts, with **three distinct speaking styles**:

* `friendly` — warm, conversational tone
* `question` — interrogative intonation
* `neutral` — balanced, plain delivery