Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
---
|
| 2 |
-
license:
|
| 3 |
task_categories:
|
| 4 |
- feature-extraction
|
| 5 |
language:
|
|
@@ -34,9 +34,7 @@ No audio is included. This is a structured acoustic feature dataset for linguist
|
|
| 34 |
|
| 35 |
Enterprise & Commercial Licensing
|
| 36 |
|
| 37 |
-
This 7-language prosody dataset is a free,
|
| 38 |
-
|
| 39 |
-
|
| 40 |
|
| 41 |
---
|
| 42 |
|
|
@@ -53,6 +51,27 @@ The earlier `korean_cv24` and `hindi_cv24` files in the two-language pack were g
|
|
| 53 |
All three are fixed in this release. The full T2 gate now requires simultaneously: SNR ≥ 25 dB **and** C50 ≥ 20 dB **and** speech_ratio ≥ 0.30. Tier labels have been retroactively corrected across all ledgers.
|
| 54 |
|
| 55 |
If you downloaded the previous pack, please replace both files.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 56 |
|
| 57 |
---
|
| 58 |
|
|
@@ -64,7 +83,7 @@ If you downloaded the previous pack, please replace both files.
|
|
| 64 |
- **Version:** 1.1
|
| 65 |
- **Languages:** Korean, Hindi, Hebrew, Manx, Tzeltal, Maguindanao, Lasi
|
| 66 |
- **Source corpora:** Mozilla Common Voice CV24 (CC0-1.0) and Spontaneous Speech SPS2 (CC0-1.0)
|
| 67 |
-
- **License:**
|
| 68 |
- **Total clips:** 58,830
|
| 69 |
- **Anonymization standard:** `moonscape_k5_v1`
|
| 70 |
|
|
@@ -241,7 +260,7 @@ Use `gender_source == 'meta'` to filter to self-reported labels for any demograp
|
|
| 241 |
title = {Alexandria Voice Corpus --- Multilingual Macro-Prosody Telemetry v1.1},
|
| 242 |
author = {MoonScape Software},
|
| 243 |
year = {2026},
|
| 244 |
-
license = {
|
| 245 |
note = {Derived from Mozilla Common Voice CV24 and Spontaneous Speech SPS2 (CC0).
|
| 246 |
Acoustic features extracted via Parselmouth/Praat. v1.1 corrects
|
| 247 |
quality-gate bugs BUG-029, BUG-030, and high-SNR C50 bypass present
|
|
@@ -280,4 +299,4 @@ Use `gender_source == 'meta'` to filter to self-reported labels for any demograp
|
|
| 280 |
## Contact
|
| 281 |
|
| 282 |
Chris - Founder and Lead Systems Design
|
| 283 |
-
moonscapesoftware@gmail.com
|
|
|
|
| 1 |
---
|
| 2 |
+
license: openrail
|
| 3 |
task_categories:
|
| 4 |
- feature-extraction
|
| 5 |
language:
|
|
|
|
| 34 |
|
| 35 |
Enterprise & Commercial Licensing
|
| 36 |
|
| 37 |
+
This 7-language prosody dataset is a free, open sample of Moonscape Software’s feature-extraction pipeline released under Open RAIL license. Moonscape Software also maintains a proprietary, commercially-cleared 1.2 million row Cross-Linguistic Macro Prosody Index. For enterprise licensing, bulk data access, or commercial applications, please contact: moonscapesoftware@gmail.com
|
|
|
|
|
|
|
| 38 |
|
| 39 |
---
|
| 40 |
|
|
|
|
| 51 |
All three are fixed in this release. The full T2 gate now requires simultaneously: SNR ≥ 25 dB **and** C50 ≥ 20 dB **and** speech_ratio ≥ 0.30. Tier labels have been retroactively corrected across all ledgers.
|
| 52 |
|
| 53 |
If you downloaded the previous pack, please replace both files.
|
| 54 |
+
---
|
| 55 |
+
License updated to Open RAIL.
|
| 56 |
+
|
| 57 |
+
## License and Acceptable Use Policy (OpenRAIL)
|
| 58 |
+
|
| 59 |
+
This dataset is released under an **OpenRAIL (Responsible AI License)** framework. By downloading, accessing, or using this dataset, you explicitly agree to the following Permitted and Forbidden uses.
|
| 60 |
+
|
| 61 |
+
**Permitted Uses:**
|
| 62 |
+
You are encouraged to use this de-identified macro-prosody telemetry data for:
|
| 63 |
+
* Linguistic and cross-linguistic prosody research.
|
| 64 |
+
* Acoustic feature analysis and statistical modeling.
|
| 65 |
+
* Training and evaluating deepfake detection, anti-spoofing, or synthetic audio identification models.
|
| 66 |
+
|
| 67 |
+
**Strictly Forbidden Uses:**
|
| 68 |
+
Because this dataset contains extracted acoustic features and voice physics telemetry, the following use cases are strictly prohibited to protect biometric privacy:
|
| 69 |
+
1. **Speaker Re-Identification:** You may not attempt to reverse-engineer, unmask, re-link, or re-identify any individual speaker from this data, nor may you attempt to reconstruct the original source audio.
|
| 70 |
+
2. **Biometric Surveillance:** You may not use this data to develop, train, fine-tune, or enhance any biometric identification system, real-time tracking software, or law enforcement surveillance tool.
|
| 71 |
+
3. **Speech Synthesis / Voice Cloning:** You may not use this telemetry data to train, condition, or fine-tune generative Text-to-Speech (TTS), Voice Conversion (VC), or other synthetic voice generation models.
|
| 72 |
+
|
| 73 |
+
*Failure to comply with these behavioral restrictions immediately terminates your license to use this dataset.*
|
| 74 |
+
|
| 75 |
|
| 76 |
---
|
| 77 |
|
|
|
|
| 83 |
- **Version:** 1.1
|
| 84 |
- **Languages:** Korean, Hindi, Hebrew, Manx, Tzeltal, Maguindanao, Lasi
|
| 85 |
- **Source corpora:** Mozilla Common Voice CV24 (CC0-1.0) and Spontaneous Speech SPS2 (CC0-1.0)
|
| 86 |
+
- **License:** OpenRAIL
|
| 87 |
- **Total clips:** 58,830
|
| 88 |
- **Anonymization standard:** `moonscape_k5_v1`
|
| 89 |
|
|
|
|
| 260 |
title = {Alexandria Voice Corpus --- Multilingual Macro-Prosody Telemetry v1.1},
|
| 261 |
author = {MoonScape Software},
|
| 262 |
year = {2026},
|
| 263 |
+
license = {OpenRAIL},
|
| 264 |
note = {Derived from Mozilla Common Voice CV24 and Spontaneous Speech SPS2 (CC0).
|
| 265 |
Acoustic features extracted via Parselmouth/Praat. v1.1 corrects
|
| 266 |
quality-gate bugs BUG-029, BUG-030, and high-SNR C50 bypass present
|
|
|
|
| 299 |
## Contact
|
| 300 |
|
| 301 |
Chris - Founder and Lead Systems Design
|
| 302 |
+
moonscapesoftware@gmail.com
|