Datasets:
Update dataset card
Browse files
README.md
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-4.0
|
| 3 |
+
task_categories:
|
| 4 |
+
- audio-to-audio
|
| 5 |
+
- audio-classification
|
| 6 |
+
tags:
|
| 7 |
+
- rir
|
| 8 |
+
- acoustics
|
| 9 |
+
- benchmark
|
| 10 |
+
- evaluation
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# RIRMega-Eval
|
| 14 |
+
|
| 15 |
+
Official benchmark views and evaluation harness artifacts built from:
|
| 16 |
+
- `mandipgoswami/rirmega` :contentReference[oaicite:6]{index=6}
|
| 17 |
+
- `mandipgoswami/rir-mega-speech` :contentReference[oaicite:7]{index=7}
|
| 18 |
+
|
| 19 |
+
This HF repo stores:
|
| 20 |
+
- metadata parquet
|
| 21 |
+
- fixed splits
|
| 22 |
+
- checksums + manifest
|
| 23 |
+
- (optional) a small core audio subset
|
| 24 |
+
|
| 25 |
+
## Tasks
|
| 26 |
+
- `v1_param_estimation`: RIR -> RT60, EDT, DRR, C50, C80, Ts
|
| 27 |
+
- `v1_auralization_consistency`: (dry + RIR) -> convolved comparisons
|
| 28 |
+
|
| 29 |
+
## How to evaluate
|
| 30 |
+
Use the official evaluator:
|
| 31 |
+
- Code: https://github.com/mandip42/rirmega-eval
|
| 32 |
+
- CLI: `python scripts/evaluate.py ...`
|
| 33 |
+
|
| 34 |
+
## Citation
|
| 35 |
+
See `CITATION.cff` in the code repo. Add Zenodo DOI when minted.
|