TTS Timeout Error

[xings19]

I write a func to convert text to audio, but I always meet this error:

def conver_text_to_audio(text, save_dir, lang=None):
    if lang is None:
        lang = compare_en_cn_chars(text)
    sha256_text = sha256(text.encode()).hexdigest()
    audio_output_path = os.path.join(save_dir,'{}.wav'.format(sha256_text))
    if os.path.exists(audio_output_path):
        return
    max_try_num = 3
    count = 0
    while True:
        if count >= max_try_num:
            return
        if lang == 'en':
            voice_name = random.choice(ENGLISH_VOICE_NAME)
        elif lang == 'zh':
            voice_name = random.choice(CHINESE_VOICE_NAME)
        else:
            raise ValueError('Unknown language code')
        speech_config = speechsdk.SpeechConfig(subscription='mykey', region='eastus')
        audio_config = speechsdk.audio.AudioOutputConfig(filename=audio_output_path)
        speech_config.speech_synthesis_voice_name = voice_name
        speech_config.enable_audio_logging()
        speech_synthesizer = speechsdk.SpeechSynthesizer(speech_config=speech_config, audio_config=audio_config)
        speech_synthesis_result = speech_synthesizer.speak_text_async(text).get()
        if speech_synthesis_result.reason == speechsdk.ResultReason.SynthesizingAudioCompleted:
            return
        elif speech_synthesis_result.reason == speechsdk.ResultReason.Canceled:
            cancellation_details = speech_synthesis_result.cancellation_details
            print("Speech synthesis canceled: {}".format(cancellation_details.reason))
            if cancellation_details.reason == speechsdk.CancellationReason.Error:
                if cancellation_details.error_details:
                    print("Error details: {}".format(cancellation_details.error_details))
                    print("Did you set the speech resource key and region values?")
            print("Failed to synthesize audio, retrying...")
        count += 1

Speech synthesis canceled: CancellationReason.Error
Error details: Timeout while synthesizing. Current RTF: 2.59965 (threshold 2), frame interval 3013ms (threshold 3000ms). USP state: ReceivingData. Received audio size: 37088 bytes.
Did you set the speech resource key and region values?
Failed to synthesize audio, retrying...
Speech synthesis canceled: CancellationReason.Error
Error details: Timeout while synthesizing. Current RTF: 2.30968 (threshold 2), frame interval 3008ms (threshold 3000ms). USP state: ReceivingData. Received audio size: 78036 bytes.
Did you set the speech resource key and region values?
Failed to synthesize audio, retrying...
Speech synthesis canceled: CancellationReason.Error
Error details: Timeout while synthesizing. Current RTF: 2.42203 (threshold 2), frame interval 3013ms (threshold 3000ms). USP state: ReceivingData. Received audio size: 39818 bytes.
Did you set the speech resource key and region values?
Failed to synthesize audio, retrying...
Speech synthesis canceled: CancellationReason.Error
Error details: Timeout while synthesizing. Current RTF: 2.42203 (threshold 2), frame interval 3013ms (threshold 3000ms). USP state: ReceivingData. Received audio size: 39818 bytes.
Did you set the speech resource key and region values?

My quota is standard S0

[pankopon]

Hi, the description does not say what was used for

• Speech SDK version
• input text
• voice name

and what the Python version and OS platform were.

I tested the code with

sentence = "The quick brown fox jumped over the lazy watchdog's head. "
conver_text_to_audio(sentence, "C:/Temp", "en")

and

long_sentence = sentence * 150
conver_text_to_audio(long_sentence, "C:/Temp", "en")

using voice_name="en-US-AriaNeural" and the latest Speech SDK 1.42.0 release, on WinPython 3.12.5, with success.

Please verify that your subscription key is indeed working, run e.g.
/~https://github.com/Azure-Samples/cognitive-services-speech-sdk/tree/master/quickstart/python/text-to-speech

Note that enable_audio_logging() is only for audio logging on the service. Local Speech SDK logging is enabled differently: https://learn.microsoft.com/azure/ai-services/speech-service/how-to-use-logging

@yulin-li for follow-up and/or in case the issue is somehow 'eastus' specific.