Datasets:
Tasks:
Automatic Speech Recognition
Modalities:
Text
Languages:
Chinese
Size:
10K<n<100K
ArXiv:
License:
Update README.md
Browse files
README.md
CHANGED
@@ -62,7 +62,7 @@ extra_gated_fields:
|
|
62 |
[](https://huggingface.co/datasets/BAAI/SeniorTalk)
|
63 |
[](https://www.arxiv.org/pdf/2503.16578)
|
64 |
[](https://creativecommons.org/licenses/by-nc-sa/4.0/)
|
65 |
-
[. This is a crucial age range often overlooked in speech datasets.
|
76 |
-
* **Speakers:** 202 unique
|
77 |
* **Geographic Diversity:** Speakers from 16 of China's 34 provincial-level administrative divisions, capturing a range of regional accents.
|
78 |
-
* **Gender Balance:** Approximately
|
79 |
* **Recording Conditions:** Recordings were made in quiet environments using a variety of smartphones (both Android and iPhone devices) to ensure real-world applicability.
|
80 |
* **Content:** Natural, conversational speech during age-appropriate activities. The content is unrestricted, promoting spontaneous and natural interactions.
|
81 |
* **Audio Format:** WAV files with a 16kHz sampling rate.
|
|
|
62 |
[](https://huggingface.co/datasets/BAAI/SeniorTalk)
|
63 |
[](https://www.arxiv.org/pdf/2503.16578)
|
64 |
[](https://creativecommons.org/licenses/by-nc-sa/4.0/)
|
65 |
+
[](https://github.com/flageval-baai/SeniorTalk)
|
66 |
|
67 |
## Introduction
|
68 |
|
|
|
73 |
This dataset contains 55.53 hours of high-quality speech data collected from 202 elderly across 16 provinces in China. Key features of the dataset include:
|
74 |
|
75 |
* **Age Range:** 75-85 years old (inclusive). This is a crucial age range often overlooked in speech datasets.
|
76 |
+
* **Speakers:** 202 unique elderly speakers.
|
77 |
* **Geographic Diversity:** Speakers from 16 of China's 34 provincial-level administrative divisions, capturing a range of regional accents.
|
78 |
+
* **Gender Balance:** Approximately 7:13 representation of male and female speakers, largely attributed to the differing average ages of males and females among the elderly.
|
79 |
* **Recording Conditions:** Recordings were made in quiet environments using a variety of smartphones (both Android and iPhone devices) to ensure real-world applicability.
|
80 |
* **Content:** Natural, conversational speech during age-appropriate activities. The content is unrestricted, promoting spontaneous and natural interactions.
|
81 |
* **Audio Format:** WAV files with a 16kHz sampling rate.
|