Update README.md

README.md

@@ -7,7 +7,9 @@ sdk: static
 pinned: false
 ---
 
-The End-to-end Speech Challenge (ESC) is a benchmark for assessing a single ASR system on a collection of eight different speech recognition datasets. The ESC datasets are sourced from different domains and cover a range of audio and text distributions (speaking styles, background noise, transcription requirements). ESC consists of:
+The End-to-end Speech Challenge (ESC) is a benchmark for assessing a single ASR system on a collection of eight different speech recognition datasets. The ESC datasets are sourced from different domains and cover a range of audio and text distributions (speaking styles, background noise, transcription requirements). These distributions are a-priori unknown: systems must use the same training and evaluation algorithms across datasets and may not use any dataset-specific pre- or post-processing. Consequently, ESC encourages generalisable ASR systems that can be applied in a multi-domain setting.
+
+ESC consists of:
 
 - A [Hugging Face dataset](https://huggingface.co/datasets/esc-bench/esc-datasets) to download and use pre-prepared ESC audio-text data