Papers › MediaSpeech: Multilanguage ASR Benchmark and Dataset

MediaSpeech: Multilanguage ASR Benchmark and Dataset

30 Mar 2021arXiv:2103.16193archive 2025-07-28

Rostislav Kolobov, Olga Okhapkina, Olga Omelchishina, Andrey Platunov, Roman Bedyakin, Vyacheslav Moshkin, Dmitry Menshikov, Nikolay Mikhaylovskiy

The performance of automated speech recognition (ASR) systems is well known to differ for varied application domains. At the same time, vendors and research groups typically report ASR quality results either for limited use simplistic domains (audiobooks, TED talks), or proprietary datasets. To fill this gap, we provide an open-source 10-hour ASR system evaluation dataset NTR MediaSpeech for 4 languages: Spanish, French, Turkish and Arabic. The dataset was collected from the official youtube channels of media in the respective languages, and manually transcribed. We estimate that the WER of the dataset is under 5%. We have benchmarked many ASR systems available both commercially and freely, and provide the benchmark results. We also open-source baseline QuartzNet models for each language.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

NTRLab/MediaSpeech officialmentioned on GitHub report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Speech Recognitionspeech-recognition

Datasets

Introduced by this paper, per the archive.

MediaSpeech

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Speech Recognition MediaSpeech Quartznet WER for Arabic 0.1300 #1 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Quartznet WER for French 0.1915 #1 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Quartznet WER for Spanish 0.1826 #1 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Quartznet WER for Turkish 0.1422 #1 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Wit WER for Arabic 0.2333 #2 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Wit WER for French 0.1759 #2 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Wit WER for Spanish 0.0879 #2 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Wit WER for Turkish 0.0768 #2 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Azure WER for Arabic 0.3016 #3 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Azure WER for French 0.1683 #3 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Azure WER for Spanish 0.1296 #3 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Azure WER for Turkish 0.2296 #3 of 8 Archive leaderboard report
Speech Recognition MediaSpeech VOSK WER for Arabic 0.3085 #4 of 8 Archive leaderboard report
Speech Recognition MediaSpeech VOSK WER for French 0.2111 #4 of 8 Archive leaderboard report
Speech Recognition MediaSpeech VOSK WER for Spanish 0.1970 #4 of 8 Archive leaderboard report
Speech Recognition MediaSpeech VOSK WER for Turkish 0.3050 #4 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Google WER for Arabic 0.4464 #5 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Google WER for French 0.2385 #5 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Google WER for Spanish 0.2176 #5 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Google WER for Turkish 0.2707 #5 of 8 Archive leaderboard report
Speech Recognition MediaSpeech wav2vec WER for Arabic 0.9596 #6 of 8 Archive leaderboard report
Speech Recognition MediaSpeech wav2vec WER for French 0.3113 #6 of 8 Archive leaderboard report
Speech Recognition MediaSpeech wav2vec WER for Spanish 0.2469 #6 of 8 Archive leaderboard report
Speech Recognition MediaSpeech wav2vec WER for Turkish 0.5812 #6 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Deepspeech WER for French 0.4741 #7 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Deepspeech WER for Spanish 0.4236 #7 of 8 Archive leaderboard report
Speech Recognition MediaSpeech Silero WER for Spanish 0.3070 #8 of 8 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections