Case Study: Leading Audio Hardware Company achieves robust multilingual ASR and high-quality training data with TransPerfect (DataForce)

DataForce supports a global audio hardware leader with high quality data for fine-tuning their ASR engine

Leading Audio Hardware Company needed multilingual, demographically diverse training and test data to fine-tune its automatic speech recognition (ASR) engine across English, Hindi, German, French, and Italian, accounting for gender, age, accents and real-world background noise. They engaged TransPerfect’s DataForce to source high-quality voice and ambient recordings that met strict technical requirements.

TransPerfect’s DataForce used its global community (1,000,000 members) and the proprietary DataForce Contribute mobile app to collect voice commands and environmental noise in scenarios such as parking, driving and windows open/closed, ensured signal-to-noise and sampling-rate standards, and convolved recordings to simulate real environments. The result: the client developed and tested an efficient ASR engine that reliably recognizes voice commands across multiple languages and noisy scenarios, with data quality and scale delivered by TransPerfect.