# AudioCodec Benchmark (Emotion recognition)
Auther: Haibin Wu

## Resynthesis data (IEMOCAP)
- The original data is here: https://huggingface.co/datasets/AudioDecBenchmark/Voxceleb1_test_original
- Resynthesis the data by different codecs: https://github.com/voidful/AudioDecBenchmark/blob/main/dataset_creator.py

## Follow superb to run the testing
- Follow the script document: https://github.com/s3prl/s3prl/blob/f0ddfc4072336ea71ed032ed3580eaab742d98d0/s3prl/downstream/docs/superb.md?plain=1#L595
- use wavLM large as the upstream model