Custom Neural Voice Access Approved but Still Getting "Access Required" Error in Speech Studio - Unable to Train Model

It seems that you are experiencing issues with accessing the Custom Neural Voice feature in Speech Studio despite receiving approval. Here are some insights that may help address your questions:

1. Resource Access: Custom Neural Voice approval does not automatically apply to all resources in the subscription. Each resource may need to be manually enabled for Custom Neural Voice access. Ensure that the specific Speech resource you are using has been enabled for this feature.
2. Roles Required: The role of "Cognitive Services Speech Contributor" is indeed necessary for data plane access to Speech services. While Owner and Contributor roles provide broad access, they may not grant the specific permissions needed for certain functionalities in Speech services. It is advisable to assign the appropriate role to ensure you have the necessary permissions.
3. Activation Steps: After receiving the approval email, there may be additional steps required to activate the access. This could include ensuring that the resource is correctly configured or that any necessary settings are applied.

Given that you have already tried creating a new Speech resource and waited for permission propagation, I recommend checking the specific configurations of your Speech resource and ensuring that the correct roles are assigned. If the issue persists, contacting Microsoft support for further assistance may be beneficial.

References:

• Limited Access
• Custom voice lite

Hello Daksh Goyal,

Welcome to Microsoft Q&A and Thank you for reaching out.

I understand that you're running into some frustrating access issues even after getting your Custom Neural Voice approval. Let’s break down your situation and tackle those questions of yours:

Here's what you can check:

Resource Access: Custom Neural Voice approval does not automatically apply to every resource in your subscription. Each resource needs to be specifically enabled for this feature. Since you've created a new Speech resource after your approval but are facing the same issue, this points to a potential access issue rather than a problem with your resource configuration.

Roles and Permissions: The roles you mentioned (Owner, Contributor) generally provide sufficient permissions, but it is often a good idea to verify if the "Cognitive Services Speech Contributor" role is necessary, as it might grant additional permissions you need to work with the Speech services effectively.

Activation Steps: After receiving the approval email, you typically don't have a separate activation step. However, you must ensure your project is correctly set up and configured to use Custom Neural Voice settings specifically.

Wait Time for Propagation: You've done well by waiting 48+ hours; usually, access should propagate within this timeframe. If you’re still experiencing issues, it may be worth reaching out to support.

Answers to Your Questions:

Access Scope: No, approval doesn’t apply universally to all resources in your subscription; each needs manual setup.

Role Requirement: While Owner/Contributor roles are powerful, it may be helpful to check if the "Cognitive Services Speech Contributor" role can resolve your issue.

Activation Step: Generally, there’s no additional step after approval, but configuration is key. 

I Hope this helps. Do let me know if you have any further queries.

If this answers your query, please do click Accept Answer and Yes for was this answer helpful.

Thank you!