# A Speech Test Set of Practice Business Presentations with Additional Relevant Texts

## TL;DR

This paper introduces a specialized speech test set comprising student business presentations with associated texts and multimedia, aimed at evaluating and improving ASR systems in domain-specific contexts.

## Contribution

It provides a new, annotated corpus of business presentation recordings with transcripts and multimedia, and benchmarks existing ASR systems on this dataset.

## Key findings

- Baseline ASR systems show significant errors on the corpus.
- The corpus highlights challenges in recognizing domain-specific vocabulary.
- The dataset supports future improvements in domain-adapted speech recognition.

## Abstract

We present a test corpus of audio recordings and transcriptions of presentations of students' enterprises together with their slides and web-pages. The corpus is intended for evaluation of automatic speech recognition (ASR) systems, especially in conditions where the prior availability of in-domain vocabulary and named entities is benefitable. The corpus consists of 39 presentations in English, each up to 90 seconds long. The speakers are high school students from European countries with English as their second language. We benchmark three baseline ASR systems on the corpus and show their imperfection.