Noctus(@noctus91):Can an STT model make sense of a 135 year old recording? I tested @MistralAI new Voxtral transcription model on a noisy French phonograph recording of Gustave Eiffel from 1891. It matched 100 of 107 words in INA’s published excerpts (6.5% WER), while transcribing 75 additional words omitted from the reference. It handled Eiffel’s dictation well but struggled with proper names and some guest dialogue. A fascinating test of modern speech recognition on audio from 1891.

Noctus

@noctus91

Benchmarking for fun | Ambassador @MistralAI | Open to interesting projects & collaborations 📩

加入 August 2025

361 正在關注 1.4K 粉絲

Noctus@noctus91

2026.10.20 16:41

Can an STT model make sense of a 135 year old recording? I tested @MistralAI new Voxtral transcription model on a noisy French phonograph recording of Gustave Eiffel from 1891. It matched 100 of 107 words in INA’s published excerpts (6.5% WER), while transcribing 75 additional words omitted from the reference. It handled Eiffel’s dictation well but struggled with proper names and some guest dialogue. A fascinating test of modern speech recognition on audio from 1891.