Asturian Speech to Text: Preserving a Language Through Technology

Asturian (asturianu) is a Romance language spoken primarily in the Asturias region of northern Spain. With around 100,000 native speakers and many more who understand it, Asturian is a vital part of the region's cultural identity. In recent years, there has been a growing movement to promote the language in education, media, and public life. However, like many minority languages, Asturian faces challenges in the digital world—limited online content, few software tools, and scarce automatic speech recognition (ASR) support.

Why Accurate Speech-to-Text Matters for Asturian

Transcribing Asturian audio manually is time-consuming and expensive, especially for long recordings like interviews, lectures, or podcasts. Automatic speech-to-text can dramatically speed up this process, making it easier to create subtitles, searchable archives, and written records. For Asturian, accurate ASR is not just a convenience; it is a tool for language preservation. By converting spoken Asturian into text, we can create digital resources that help new learners, document oral traditions, and make the language visible in the digital sphere.

Transcription Challenges Specific to Asturian

Asturian has several features that make ASR challenging:

• Dialectal variation: The three main dialects (Western, Central, Eastern) differ in pronunciation, vocabulary, and even verb conjugations. A model trained only on Central Asturian may mishear Western forms.
• Orthographic instability: Although the Academy of the Asturian Language (ALLA) promotes a standard orthography, many speakers write phonetically or use older conventions. This means the ASR must handle multiple possible spellings for the same word.
• Limited training data: Compared to Spanish or Catalan, there are far fewer transcribed hours of Asturian speech available for training. This scarcity can lead to lower accuracy unless the model uses transfer learning from related languages.
• Phonological nuances: Asturian preserves certain sounds that have disappeared in Spanish, such as the voiceless retroflex affricate /t͡s/ (written "ts") and the palatal lateral /ʎ/ (written "ll"). These must be correctly recognized to avoid errors.

Use Cases for Asturian Speech-to-Text

Podcasts and Radio

Asturian-language podcasts and radio programs can benefit from automatic transcription to provide show notes, searchable episodes, and subtitles for deaf listeners. Speechyou can transcribe an entire hour-long podcast in minutes, outputting SRT or VTT files ready for use.

Oral History Preservation

Many elderly speakers of Asturian have stories, songs, and traditional knowledge that has never been written down. By transcribing these recordings, researchers can create a permanent written record and make the content accessible to younger generations who may not speak the language fluently.

Education

Asturian is taught in some schools in Asturias, and teachers can use speech-to-text to transcribe their lessons for students who need written reinforcement. Additionally, learners can practice their listening skills by reading along with transcriptions of native speakers.

Accessibility

Deaf and hard-of-hearing individuals who use Asturian as their first language need captions for local events, TV programs, and online videos. Speechyou enables content creators to add Asturian subtitles easily, promoting inclusion.

Media Production

Independent filmmakers and YouTubers producing content in Asturian can generate subtitles automatically, saving time and money. With support for both SRT and VTT formats, the subtitles can be embedded directly into videos.

How Speechyou Helps

Speechyou is one of the few ASR platforms that actively supports Asturian. Our model is trained on a diverse corpus of Asturian speech, including recordings from all three dialect areas and both formal and informal registers. We use a hybrid approach combining deep learning with language-specific rules to handle orthographic variation and rare words. The result is a transcription service that achieves over 95% word accuracy on clear, standard Asturian speech.

Our platform supports:

• Audio and video upload in common formats (MP3, WAV, MP4, etc.)
• Automatic language detection for Asturian (or manual selection)
• Subtitle generation in SRT and VTT formats with accurate timestamps
• Export to plain text for further editing or analysis
• Unlimited transcription under the Solo plan, making it affordable for individual users and small organizations.

Conclusion

Asturian is a language with a rich history and a bright future, but it needs digital tools to thrive in the 21st century. Speech-to-text technology is a key enabler, allowing speakers to create written records of their spoken language, share content with subtitles, and preserve oral traditions for posterity. With Speechyou, transcribing Asturian audio has never been easier—try it today and see how AI can help keep Asturian alive and accessible.