Get default rate of words per minute for text to speech synthetizer for configured SpeechConfig

Question

Get default rate of words per minute for text to speech synthetizer for configured SpeechConfig

Anna Dřevíkovská 1

Hi,

is it possible to get default speech rate in words per minute or something similar from the Java Speech SDK when using SpeechConfig and SpeechSynthesizer when using SpeakTextAsync for text-to-speech services?

We'd like to display for user a rough estimate of how much time it's going to take to finish the text-to-speech audio file.

Thank you for your answer.

Kind Regards,
Anna Drevikovska

1 answer

Your answer

Answer 1

romungi-MSFT 48,911 Microsoft Employee Moderator

@Anna Dřevíkovská I think you should be able to get the duration of the audio file by using the SpeechSynthesisResult.getAudioLength() method. This will be accessible after the synthesis is complete. This should also give you the exact duration of the audio file generated. I hope this helps!!

If an answer is helpful, please click on or upvote which might help other community members reading this thread.

Anna Dřevíkovská 1 Reputation point

2022-10-07T07:53:38.467+00:00

Hi @romungi-MSFT , thanks for the answer. I know that we could get the audio duration after the synthesis are completed, what I've meant, if there is any way to get an estimate of how long the synthesis will take to complete. Is this somehow related to the number of words in the text and rate of the speech and of course probably some amount of computation time included? Is there anyway to get the default rate of spoken words per minute for chosen language or voice dynamically?
romungi-MSFT 48,911 Reputation points Microsoft Employee Moderator

2022-10-07T13:24:52.253+00:00
@Anna Dřevíkovská Yes, as you mentioned the various factors that are used like the language, rate of speech, speaking style, initial silence and the voice used causes the length of the audio to differ. But I do see that the voice list API does provide an estimate. The response of the the GET request to the API should provide you details like below:

Use the header Ocp-Apim-Subscription-Key along with its value with the GET request to below endpoint as per your region.

https://<your_region>.tts.speech.microsoft.com/cognitiveservices/voices/list [ { "Name": "Microsoft Server Speech Text to Speech Voice (af-ZA, AdriNeural)", "DisplayName": "Adri", "LocalName": "Adri", "ShortName": "af-ZA-AdriNeural", "Gender": "Female", "Locale": "af-ZA", "LocaleName": "Afrikaans (South Africa)", "SampleRateHertz": "24000", "VoiceType": "Neural", "Status": "GA", "WordsPerMinute": "147" }, ...

WordsPerMinute from above response should be a rough estimate and you can confirm that from your actual synthesis.
Anna Dřevíkovská 1 Reputation point

2022-10-11T08:47:02.227+00:00
Hi @romungi-MSFT , thank you for your answer, this is exactly what I've needed. :) Actually one more question.

This is the text we're sending to synthesis when testing the services:

Our mission at GoodTrust is to help protect the digital life and legacy of everyone.

With language set to: en-US, voice set to: en-US-JennyNeural, and output format set to non-streaming riff-16khz-16bit-mono-pcm currently the computation time from SDK and REST API both took around 6-7 seconds to get the audio file result,, is this expected?

When using this site: https://studio.d-id.com we get the TTS result almost immediately. From what we know from our communication with D-ID they're also using Azure for TTS services, is there something in our setup that is causing this difference in the computation time? I'm currently following the Azure docs implementation examples.

We're using westeurope region and pay-as-you-go financing tier.

Thank you for your answer.

Kind Regards,
Anna
romungi-MSFT 48,911 Reputation points Microsoft Employee Moderator

2022-10-21T05:38:28.35+00:00

I think it shouldn't really take 6-7 seconds. You can probably look at the metrics or enable logging on the SDK to check for more information on the delay.

Share via

Get default rate of words per minute for text to speech synthetizer for configured SpeechConfig

1 answer

Your answer