Multiple language support in speech to text API

[locatoca]

"config": {

        "encoding": "MP3",

        "sampleRateHertz": 16000,

        "languageCode": "en-US",

        "alternativeLanguageCodes": [

            "ml-IN","kn-IN","ta-IN"

        ],

        "model": "latest_long",

         "useEnhanced": true

    }

I have passed multiple language codes, but I am not getting results for alternativeLanguageCodes. My audio content language is ml-IN. This is needed for a requirement where the language of the audio file is unknown beforehand

[Poala_Tenorio]

Here are a few things you can check and consider:

• Check if the selected model (latest_long in your case) supports the alternative language codes provided. Not all models may support all languages, so selecting the appropriate model for multilingual transcription is crucial.
• You're using `useEnhanced: true`, which implies that you're opting for the enhanced model. While the enhanced model generally provides better accuracy, it might prioritize the primary language over alternative languages. You might want to experiment with setting `useEnhanced: false` to see if it affects the transcription results for alternative languages.
• If possible, try using sample audio files in Malayalam (ml-IN) and other languages (such as Tamil and Kannada) separately to verify if the API returns accurate results for each language individually. This can help isolate whether the issue lies with the API configuration or the content itself.
• Make sure that the audio quality is adequate for accurate transcription. Poor audio quality can affect the accuracy of the transcription, especially for languages with less robust support.