Questions About CNV(Custom Neural Voice) Quotas and Limits.

Question

Hi
I have few questions while reading CNV(Custom Neural Voice) quotas and limits. (link below)
https://learn.microsoft.com/en-us/azure/cognitive-services/speech-service/speech-services-quotas-and-limits#speech-to-text-quotas-and-limits-per-speech-resource

1) I thought that I can make several CNV projects on one single Speech Resource. Is that right?

2) According to the document, "Max number of custom endpoints per Speech resource" is 50. Then it means I can make 50 endpoints on one single trained CNV model. Is that right?

2-1) If I can make 50 endpoints on one single trained CNV model, Is the endpoint hosting price $4 x 50(endpoints) per hour?

2-2) Let's assume that I make 50 endpoints on one single trained CNV model. According to the document, "Concurrent Request limit for Custom Neural voice" is 10. Is that mean I can request 500 requests concurrently?

3) "Concurrent Request limit for Custom Neural voice" is 10 and "adjustable" is "yes". How much can Concurrent Requests be adjusted?

4) What is Transaction?
I thought that transaction is generated per request. In case of endpoints is 50 and concurrent requests per each endpoint are 10, then total requests are 500 requests. But "Max number of Transactions per Second (TPS) per Speech resource" is 200.

What is a transaction and Can you explain the relationship between a transaction and a request?

Answer 1

Hi, yes, you can create multiple projects on one resource. There is a Maximum of 50 endpoints per CVN resource. Endpoint hosting can be calculated as (models x hours x $4.03). The number of concurrent requests for CVN endpoints is limited to 10 (you can request/confirm how much increase through support request). A transaction is a successful API call request.

---

--- *Kindly Accept Answer if the information helps. Thanks.*