Speech synthesis failed

Question

Speech synthesis failed

Mounica T 20 Microsoft Employee

Speech synthesis failed - Error code: 1007. Error details: Speak node can only be the root. USP state

SSML - "\u003cspeak version="1.0" xml:lang="en-US" xmlns:mstts="http://www.w3.org/2001/mstts" xmlns="http://www.w3.org/2001/10/synthesis"\u003e\u003cvoice name="en-US-AvaMultilingualNeural" xmlns=""\u003e\u003cprosody rate="20%" pitch="0%"\u003e\u003cspeak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:lang="en-US"\u003e\u003cvoice name="en-US-AvaMultilingualNeural"\u003e\r\nThis is the third question that is asked.\r\n\u003c/voice\u003e\u003c/speak\u003e You can say english, or hindi.\u003c/prosody\u003e\u003c/voice\u003e\u003c/speak\u003e"

Mounica T 20 Reputation points Microsoft Employee

2024-09-09T06:06:37.2733333+00:00

When I unescape and convert special characters, the SSML looks good to me
navba-MSFT 27,620 Reputation points Microsoft Employee Moderator

2024-09-09T09:08:22.7066667+00:00
@Mounica T Welcome to Microsoft Q&A Forum, Thank you for posting your query here!

.

Could you please let me know if you are using Speech REST APi ? or Speech SDK ?

Could you please test the Speech Synthesis from Speech Studio and check if that helps ?

Please elaborate your use case. Where / When are you encountering this issue ?

Did you try with a simple SSML file like below and check if that helps ?

<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:lang="en-US"> <voice name="en-US-AvaMultilingualNeural"> Welcome to <prosody pitch="high">Enjoy using text to speech.</prosody> </voice> </speak>

.

Awaiting your reply.
Mounica T 20 Reputation points Microsoft Employee

2024-09-09T09:16:36+00:00

Hi, I tried using Studio, it works. We are using the SDK
navba-MSFT 27,620 Reputation points Microsoft Employee Moderator

2024-09-09T09:18:53.07+00:00

@Mounica T Please share your SSML here, so that I can test at my end using speech SDK.

Also let me know the programming language you are using.

Awaiting your reply.
Mounica T 20 Reputation points Microsoft Employee

2024-09-09T09:21:39.57+00:00

It's linked in the question itself
navba-MSFT 27,620 Reputation points Microsoft Employee Moderator

2024-09-09T09:40:47.81+00:00

@Mounica T Could you please share it again in the correct format as the one mentioned in the question doesn't seem to be formatted right.

navba-MSFT 27,620 Microsoft Employee Moderator

@Mounica T I tried with the C# Speech SDK with the below SSML and it worked fine.

Could you please test the same and let me know if that helps ?

using System;
using System.IO;
using System.Net.NetworkInformation;
using System.Threading.Tasks;
using System.Xml.Linq;
using Microsoft.CognitiveServices.Speech;
using Microsoft.CognitiveServices.Speech.Audio;


namespace TTMSSMLDemo
{
    class Program
    {
        // This example requires environment variables named "SPEECH_KEY" and "SPEECH_REGION"
        static string speechKey = "XXXXXXXXXX";
        static string speechRegion = "XXXXXX";

 async static Task Main(string[] args)
 {

     var speechConfig = SpeechConfig.FromSubscription(speechKey, speechRegion);

     var speechSynthesizer = new SpeechSynthesizer(speechConfig, null);

     var ssml = @"
     <speak version='1.0' xml:lang='en-US' xmlns:mstts='http://www.w3.org/2001/mstts' xmlns='http://www.w3.org/2001/10/synthesis'>
       <voice name='en-US-AvaMultilingualNeural'>
         <prosody rate='20%' pitch='0%'>
           This is the third question that is asked. 
          </prosody>
       </voice>
      <voice name='en-US-AvaMultilingualNeural'>
             You can say English, or Hindi.
      </voice>
     </speak>";

     var result = await speechSynthesizer.SpeakSsmlAsync(ssml);
     var stream = AudioDataStream.FromResult(result);
     await stream.SaveToWaveFileAsync("file.wav");
 }
}
}

Mounica T 20 Reputation points Microsoft Employee

2024-09-09T11:45:23.53+00:00

Do you think having 2 speak tags is the problem?

But this seems to be a valid ssml too

<speak version="1.0" xml:lang="en-US" xmlns:mstts="http://www.w3.org/2001/mstts" xmlns="http://www.w3.org/2001/10/synthesis">

<voice name="en-US-AvaMultilingualNeural" xmlns="">

<prosody rate="20%" pitch="0%">

<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:l₹ng="en-US">

<voice name="en-US-AvaMultilingualNeural">This is the third question that is asked.</voice> </speak> You can say english, or hindi.

</prosody>

</voice>

</speak>
navba-MSFT 27,620 Reputation points Microsoft Employee Moderator

2024-09-10T04:49:01.43+00:00

@Mounica T Please let me know if you had a chance to look at my above comment and have any further follow-up questions.

Answer accepted by question author

0 additional answers

Your answer

Mounica T 20 Reputation points Microsoft Employee

2024-09-09T06:06:37.2733333+00:00

When I unescape and convert special characters, the SSML looks good to me
navba-MSFT 27,620 Reputation points Microsoft Employee Moderator

2024-09-09T09:08:22.7066667+00:00

@Mounica T Welcome to Microsoft Q&A Forum, Thank you for posting your query here!

.

Could you please let me know if you are using Speech REST APi ? or Speech SDK ?

Could you please test the Speech Synthesis from Speech Studio and check if that helps ?

Please elaborate your use case. Where / When are you encountering this issue ?

Did you try with a simple SSML file like below and check if that helps ?

<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:lang="en-US"> <voice name="en-US-AvaMultilingualNeural"> Welcome to <prosody pitch="high">Enjoy using text to speech.</prosody> </voice> </speak>

.

Awaiting your reply.
Mounica T 20 Reputation points Microsoft Employee

2024-09-09T09:16:36+00:00

Hi, I tried using Studio, it works. We are using the SDK
navba-MSFT 27,620 Reputation points Microsoft Employee Moderator

2024-09-09T09:18:53.07+00:00

@Mounica T Please share your SSML here, so that I can test at my end using speech SDK.

Also let me know the programming language you are using.

Awaiting your reply.
Mounica T 20 Reputation points Microsoft Employee

2024-09-09T09:21:39.57+00:00

It's linked in the question itself
navba-MSFT 27,620 Reputation points Microsoft Employee Moderator

2024-09-09T09:40:47.81+00:00

@Mounica T Could you please share it again in the correct format as the one mentioned in the question doesn't seem to be formatted right.
navba-MSFT 27,620 Reputation points Microsoft Employee Moderator

2024-09-09T10:20:48.3833333+00:00

@Mounica T I tried with the C# Speech SDK with the below SSML and it worked fine.

Could you please test the same and let me know if that helps ?

using System; using System.IO; using System.Net.NetworkInformation; using System.Threading.Tasks; using System.Xml.Linq; using Microsoft.CognitiveServices.Speech; using Microsoft.CognitiveServices.Speech.Audio; namespace TTMSSMLDemo { class Program { // This example requires environment variables named "SPEECH_KEY" and "SPEECH_REGION" static string speechKey = "XXXXXXXXXX"; static string speechRegion = "XXXXXX"; async static Task Main(string[] args) { var speechConfig = SpeechConfig.FromSubscription(speechKey, speechRegion); var speechSynthesizer = new SpeechSynthesizer(speechConfig, null); var ssml = @" <speak version='1.0' xml:lang='en-US' xmlns:mstts='http://www.w3.org/2001/mstts' xmlns='http://www.w3.org/2001/10/synthesis'> <voice name='en-US-AvaMultilingualNeural'> <prosody rate='20%' pitch='0%'> This is the third question that is asked. </prosody> </voice> <voice name='en-US-AvaMultilingualNeural'> You can say English, or Hindi. </voice> </speak>"; var result = await speechSynthesizer.SpeakSsmlAsync(ssml); var stream = AudioDataStream.FromResult(result); await stream.SaveToWaveFileAsync("file.wav"); } } }
Mounica T 20 Reputation points Microsoft Employee

2024-09-09T11:45:23.53+00:00

Do you think having 2 speak tags is the problem?

But this seems to be a valid ssml too

<speak version="1.0" xml:lang="en-US" xmlns:mstts="http://www.w3.org/2001/mstts" xmlns="http://www.w3.org/2001/10/synthesis">

<voice name="en-US-AvaMultilingualNeural" xmlns="">

<prosody rate="20%" pitch="0%">

<speak version="1.0" xmlns="http://www.w3.org/2001/10/synthesis" xml:l₹ng="en-US">

<voice name="en-US-AvaMultilingualNeural">This is the third question that is asked.</voice> </speak> You can say english, or hindi.

</prosody>

</voice>

</speak>
navba-MSFT 27,620 Reputation points Microsoft Employee Moderator

2024-09-10T04:49:01.43+00:00

@Mounica T Please let me know if you had a chance to look at my above comment and have any further follow-up questions.

Answer 1

@Mounica T The error you’re encountering, “Speak node can only be the root,” is due to the incorrect nesting of <speak> elements. In SSML (Speech Synthesis Markup Language), the <speak> element must be the root element and cannot be nested within other elements.

Here’s a corrected version of your SSML:

 <speak version='1.0' xml:lang='en-US' xmlns:mstts='http://www.w3.org/2001/mstts' xmlns='http://www.w3.org/2001/10/synthesis'>
       <voice name='en-US-AvaMultilingualNeural'>
         <prosody rate='20%' pitch='0%'>
           This is the third question that is asked. 
          </prosody>
       </voice>
      <voice name='en-US-AvaMultilingualNeural'>
             You can say English, or Hindi.
      </voice>
     </speak>

Hope this answers.

Share via

Speech synthesis failed

0 additional answers

Your answer