Latest addition to Speechmatics engine saves time on manually reviewing files and is applicable to a wide variety of use cases
Cambridge, August 31st 2022: Speechmatics, the leading autonomous speech recognition technology scaleup has now added Language Identification (Language ID) to its industry leading speech-to-text engine. This latest addition allows customers to automatically identify the predominant language spoken in any media file. Customers will save time and effort on manually reviewing files, safe in the knowledge that they will be provided with an accurate transcription of any media file.
Language ID drives efficiency by removing the manual step of selecting which language pack should be used, when the language is not explicitly stated on the file. Often-requested, it not only helps users identify unknown languages, but also adds useful metadata about the language of the spoken audio. Media and broadcast organizations have extensive archives of audio, the content of which is often unknown. Instead of manually listening to hours of speech – and relying on human interpretation to label it – Speechmatics Language ID confirms the language pre-transcription. For contact centers, being able to identify the predominant language spoken (especially when callers switch languages) is a huge benefit to those conducting call analysis.
Speechmatics has built the most accurate and inclusive speech-to-text engine available. Historically, training data had to be manually tagged, classified or ‘labelled’. This has resulted in engines that have been trained on narrow datasets, which fail to represent the diversity of voices that use them. In contrast, Speechmatics’ speech-to-text engine is trained through exposure to hundreds of thousands of individual voices using millions of hours of unlabelled, more representative voice data. Speechmatics has applied this technique to identifying predominant spoken languages on a diverse set of audio data.
Commenting on this rollout of Language ID, CEO Katy Wigdahl said, “Up until now, identifying languages without human intervention has been costly and time consuming for users of speech-to-text. However, with our new Language ID, this will be a thing of the past and allow customers to swiftly identify and transcribe media files – with less hassle and more efficiency. We can’t wait for our customers to use this Language ID and see it deliver accurate and valuable results.’’
This latest update can be used with pre-recorded media files, works with up to 12 languages and adds a confidence score to show the certainty of the predominant language. Supported languages are English, German, Spanish, French, Hindi, Italian, Japanese, Korean, Mandarin, Dutch, Portuguese, Russian.