Exception [SPXERR_RUNTIME_ERROR 0x1b] When Using SpeechRecognizer with AudioProcessingOptions for Echo Cancellation

Question

I am trying to implement Acoustic Echo Cancellation (AEC) using Microsoft's Speech SDK in C#. I am using AudioProcessingOptions.Create(AudioProcessingConstants.AUDIO_INPUT_PROCESSING_ENABLE_DEFAULT) to enable audio processing. However, I am encountering the following exception when initializing the SpeechRecognizer:-

Exception: Exception with an error code: 0x1b (SPXERR_RUNTIME_ERROR)

Code Snippet:

using (var audioProcessingOptions = AudioProcessingOptions.Create(AudioProcessingConstants.AUDIO_INPUT_PROCESSING_ENABLE_DEFAULT))
{
    using (var audioConfig = AudioConfig.FromDefaultMicrophoneInput(audioProcessingOptions))
    {
        try
        {
            SpeechConfig config1 = SpeechConfig.FromSubscription(HaiStaticData.strSTTVoiceKey, HaiStaticData.strSTTVoiceRegion);
            
            using (SpeechRecognizer recognizer = new SpeechRecognizer(config1, audioConfig)) // Exception occurs here
            {
                // Recognition logic
            }
        }
        catch (Exception ex)
        {
            Console.WriteLine(ex.ToString());
        }
    }
}

I suspect the issue might be related to:

1. Incorrect use of AudioProcessingOptions for enabling AEC.
2. Missing or incompatible microphone settings for AEC.
3. SDK version compatibility issues.

Questions:

• Is AudioProcessingConstants.AUDIO_INPUT_PROCESSING_ENABLE_DEFAULT the correct flag for enabling AEC in Microsoft Speech SDK?
• Are there any specific microphone or hardware requirements for AEC to work properly?
• Could this error be caused by an unsupported SDK version? If so, which version fully supports AEC?

I am using .NETFramework 4.7.2 and Microsoft.CognitiveServices.Speech SDK 1.42.0.

Any guidance or troubleshooting steps would be appreciated.

Answer 1

Hi,

Thank you for responding to my question!

I have two concerns:

1. I am unable to find AudioProcessingConstants.AUDIO_INPUT_PROCESSING_ENABLE_ECHO_CANCELLATION.

       Reference: [Microsoft Documentation](https://learn.microsoft.com/en-us/dotnet/api/microsoft.cognitiveservices.speech.audio.audioprocessingconstants?view=azure-dotnet)
   

2. If AUDIO_INPUT_PROCESSING_ENABLE_DEFAULT is not the correct option for AEC, it should still support basic audio processing. However, I am encountering an exception.

I would greatly appreciate clarification on both points. However, my primary requirement is to implement AEC, so my first concern is more critical.

Thanks!