Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,26 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: mit
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
# DreamVoice Dataset
|
| 6 |
+
|
| 7 |
+
## Introduction
|
| 8 |
+
|
| 9 |
+
The DreamVoice dataset is designed to support research in text-guided voice conversion and synthetic voice generation. The dataset is structured into three main files:
|
| 10 |
+
|
| 11 |
+
- `dreamvoice_s1_raw_label.csv`: Contains raw keyword votes from multiple experts. Each row represents the collective opinions of experts regarding various aspects of the speaker's voice characteristics.
|
| 12 |
+
- `dreamvoice_s2_keyword.csv`: Provides verified keywords for each speaker, serving as a reliable reference for their voice profile based on expert consensus.
|
| 13 |
+
- `dreamvoice_s3_synthetic_prompt.csv`: Includes synthetic timbre description prompts. These prompts are designed to describe the unique timbre characteristics of each speaker, facilitating the generation of synthetic voices that match specific timbral qualities.
|
| 14 |
+
|
| 15 |
+
## Usage
|
| 16 |
+
|
| 17 |
+
To use the DreamVoice dataset in your research or projects, follow these steps:
|
| 18 |
+
|
| 19 |
+
1. **Download VCTK and LibriTTS-R Datasets**
|
| 20 |
+
The DreamVoice dataset is built to work with the VCTK and LibriTTS-R speech datasets. You can download them from the following links:
|
| 21 |
+
- [VCTK Dataset](https://datashare.ed.ac.uk/handle/10283/2651)
|
| 22 |
+
- [LibriTTS-R Dataset](https://openslr.org/141)
|
| 23 |
+
|
| 24 |
+
2. **Working with the Dataset**
|
| 25 |
+
Once the datasets are downloaded, you can combine them with the DreamVoice dataset based on the column speaker_id.
|
| 26 |
+
|