Qashqai (Latin script) Speech-to-Text: A New Era for a Turkic Language

Where is Qashqai Spoken?

Qashqai is a Turkic language spoken by the Qashqai people, primarily in the Fars province of southern Iran. The language has an estimated 1.5 million speakers, making it one of the larger minority languages in Iran. Qashqai belongs to the Oghuz branch of the Turkic family, sharing features with Azerbaijani, Turkish, and Turkmen. The language has several dialects, including Kashkuli, Shishbeyli, and Sangkani, each with subtle phonetic differences. While the traditional writing system is based on the Perso-Arabic script, a Latin orthography has been developed for educational and digital purposes. Speechyou supports the Latin script variant, making it accessible to users who prefer this modern standard.

Why Accurate Qashqai Speech-to-Text Matters

In an increasingly digital world, the ability to convert spoken Qashqai into text is crucial for language preservation. Many Qashqai speakers are bilingual in Persian, and the younger generation often uses Persian for social media and communication. By providing accurate speech-to-text for Qashqai, we encourage the use of the language in writing, from text messages to blog posts. This helps maintain linguistic diversity and ensures that future generations can access cultural knowledge recorded in Qashqai. Additionally, transcription enables the creation of subtitles for videos, making Qashqai content accessible to a global audience.

Specific Transcription Challenges

Building a reliable ASR system for Qashqai comes with unique challenges:

• Vowel Harmony: Like other Turkic languages, Qashqai vowels must harmonize within a word. This rule affects suffixation and can confuse models trained on languages without harmony. Speechyou's model is specifically tuned to handle vowel harmony patterns.
• Phonemic Inventory: Qashqai has sounds like the voiceless uvular stop /q/ and the voiced velar fricative /ɣ/, which are rare in many languages. These require specialized acoustic features to recognise accurately.
• Limited Training Data: As a low-resource language, Qashqai has few publicly available speech corpora. Speechyou uses transfer learning from related Turkic languages and active learning from user feedback to improve performance.
• Dialectal Variation: Different dialects of Qashqai have distinct pronunciations. Our model is trained on a mix of dialects, but we recommend using clear speech from a single dialect for best results.

Use Cases for Qashqai Transcription

• Podcast and Video Subtitling: Qashqai content creators can automatically generate SRT subtitles for their videos, expanding their reach to both native speakers and learners.
• Oral History Preservation: Community elders can record stories and interviews, which are then transcribed to create a written archive of Qashqai folklore and history.
• Language Learning: Students of Qashqai can use transcribed audio as reading material, improving their listening and comprehension skills.
• Accessibility: Deaf and hard-of-hearing Qashqai speakers can access audio content through real-time captions or transcribed text.
• Research: Linguists and anthropologists can quickly transcribe field recordings, saving hours of manual work.
• Diaspora Connection: Qashqai families living abroad can transcribe family conversations, keeping the language alive in text form for younger generations.

How Speechyou Helps

Speechyou's Qashqai (Latin script) model is built on a deep neural network trained on thousands of hours of Turkic language data. The model is fine-tuned for Qashqai's unique phonetic features, ensuring high accuracy for clear recordings. The web interface is simple: upload an audio or video file, select Qashqai (Latin script), and receive a transcription within minutes. You can also generate SRT or VTT subtitle files with a single click. The service is available on the Solo plan, which includes unlimited transcription for individual users. For larger projects, team plans are available.

By using Speechyou, the Qashqai community can preserve their language, create content, and connect with the world. Try it today and experience the power of AI for a language that deserves to be heard.