You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: articles/ai-services/speech-service/migrate-to-batch-synthesis.md
+36-29Lines changed: 36 additions & 29 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -14,30 +14,36 @@ ms.custom: devx-track-csharp
14
14
15
15
# Migrate code from Long Audio API to Batch synthesis API
16
16
17
-
The [Batch synthesis API](batch-synthesis.md)(Preview) provides asynchronous synthesis of long-form text to speech. This article describes the benefits of upgrading from Long Audio API to Batch synthesis API, and details about how to do so.
17
+
The [Batch synthesis API](batch-synthesis.md) provides asynchronous synthesis of long-form text to speech. This article describes the benefits of upgrading from Long Audio API to Batch synthesis API, and details about how to do so.
18
18
19
19
> [!IMPORTANT]
20
-
> [Batch synthesis API](batch-synthesis.md) is currently in public preview. Once it's generally available, the Long Audio API will be deprecated.
20
+
> [Batch synthesis API](batch-synthesis.md) is generally available. the Long Audio API will be retired on April 1st, 2027.
21
21
22
-
## Base path
22
+
## Base path and version
23
23
24
-
You must update the base path in your code from `/texttospeech/v3.0/longaudiosynthesis` to `/texttospeech/3.1-preview1/batchsynthesis`. For example, to list synthesis jobs for your Speech resource in the `eastus` region, use `https://eastus.customvoice.api.speech.microsoft.com/api/texttospeech/3.1-preview1/batchsynthesis` instead of `https://eastus.customvoice.api.speech.microsoft.com/api/texttospeech/v3.0/longaudiosynthesis`.
24
+
Update the endpoint from `https://YourSpeechRegion.customvoice.api.speech.microsoft.com` to `https://YourSpeechRegion.api.cognitive.microsoft.com` or you can use custom domain instead: `https://{customDomainName}.cognitiveservices.azure.com/`.
25
+
26
+
Update the base path in your code from `/texttospeech/v3.0/longaudiosynthesis` to `/texttospeech/batchsyntheses`.
27
+
28
+
Update the version from base path to query string `/texttospeech/v3.0/longaudiosynthesis` to `?api-version=2024-04-01`.
29
+
30
+
For example, to list synthesis jobs for your Speech resource in the `eastus` region, use `https://eastus.api.cognitive.microsoft.com/texttospeech/batchsyntheses?api-version=2024-04-01` instead of `https://eastus.customvoice.api.speech.microsoft.com/api/texttospeech/v3.0/longaudiosynthesis`.
25
31
26
32
## Regions and endpoints
27
33
28
-
Batch synthesis API is available in all[Speech regions](regions.md).
34
+
Batch synthesis API is available in more[Speech regions](regions.md).
29
35
30
36
The Long Audio API is limited to the following regions:
31
37
32
-
| Region | Endpoint |
33
-
|--------|----------|
34
-
| Australia East |`https://australiaeast.customvoice.api.speech.microsoft.com`|
35
-
| East US |`https://eastus.customvoice.api.speech.microsoft.com`|
36
-
| India Central |`https://centralindia.customvoice.api.speech.microsoft.com`|
| Australia East |`https://australiaeast.customvoice.api.speech.microsoft.com`|
41
+
| East US |`https://eastus.customvoice.api.speech.microsoft.com`|
42
+
| India Central |`https://centralindia.customvoice.api.speech.microsoft.com`|
37
43
| South Central US |`https://southcentralus.customvoice.api.speech.microsoft.com`|
38
-
| Southeast Asia |`https://southeastasia.customvoice.api.speech.microsoft.com`|
39
-
| UK South |`https://uksouth.customvoice.api.speech.microsoft.com`|
40
-
| West Europe |`https://westeurope.customvoice.api.speech.microsoft.com`|
44
+
| Southeast Asia |`https://southeastasia.customvoice.api.speech.microsoft.com`|
45
+
| UK South |`https://uksouth.customvoice.api.speech.microsoft.com`|
46
+
| West Europe |`https://westeurope.customvoice.api.speech.microsoft.com`|
41
47
42
48
## Voices list
43
49
@@ -47,11 +53,12 @@ The Long Audio API is limited to the set of voices returned by a GET request to
47
53
48
54
## Text inputs
49
55
50
-
Batch synthesis text inputs are sent in a JSON payload of up to 500 kilobytes.
56
+
Batch synthesis text inputs are sent in a JSON payload of up to 2 megabytes.
51
57
52
58
Long Audio API text inputs are uploaded from a file that meets the following requirements:
53
-
* One plain text (.txt) or SSML text (.txt) file encoded as [UTF-8 with Byte Order Mark (BOM)](https://www.w3.org/International/questions/qa-utf8-bom.en#bom). Don't use compressed files such as ZIP. If you have more than one input file, you must submit multiple requests.
54
-
* Contains more than 400 characters for plain text or 400 [billable characters](./text-to-speech.md#pricing-note) for SSML text, and less than 10,000 paragraphs. For plain text, each paragraph is separated by a new line. For SSML text, each SSML piece is considered a paragraph. Separate SSML pieces by different paragraphs.
59
+
60
+
- One plain text (.txt) or SSML text (.txt) file encoded as [UTF-8 with Byte Order Mark (BOM)](https://www.w3.org/International/questions/qa-utf8-bom.en#bom). Don't use compressed files such as ZIP. If you have more than one input file, you must submit multiple requests.
61
+
- Contains more than 400 characters for plain text or 400 [billable characters](./text-to-speech.md#pricing-note) for SSML text, and less than 10,000 paragraphs. For plain text, each paragraph is separated by a new line. For SSML text, each SSML piece is considered a paragraph. Separate SSML pieces by different paragraphs.
55
62
56
63
With Batch synthesis API, you can use any of the [supported SSML elements](speech-synthesis-markup.md), including the `audio`, `mstts:backgroundaudio`, and `lexicon` elements. The long audio API doesn't support the `audio`, `mstts:backgroundaudio`, and `lexicon` elements.
57
64
@@ -61,28 +68,28 @@ Batch synthesis API supports all [text to speech audio output formats](rest-text
61
68
62
69
The Long Audio API is limited to the following set of audio output formats. The sample rate for long audio voices is 24kHz, not 48kHz. Other sample rates can be obtained through upsampling or downsampling when synthesizing.
63
70
64
-
* riff-8khz-16bit-mono-pcm
65
-
* riff-16khz-16bit-mono-pcm
66
-
* riff-24khz-16bit-mono-pcm
67
-
* riff-48khz-16bit-mono-pcm
68
-
* audio-16khz-32kbitrate-mono-mp3
69
-
* audio-16khz-64kbitrate-mono-mp3
70
-
* audio-16khz-128kbitrate-mono-mp3
71
-
* audio-24khz-48kbitrate-mono-mp3
72
-
* audio-24khz-96kbitrate-mono-mp3
73
-
* audio-24khz-160kbitrate-mono-mp3
71
+
- riff-8khz-16bit-mono-pcm
72
+
- riff-16khz-16bit-mono-pcm
73
+
- riff-24khz-16bit-mono-pcm
74
+
- riff-48khz-16bit-mono-pcm
75
+
- audio-16khz-32kbitrate-mono-mp3
76
+
- audio-16khz-64kbitrate-mono-mp3
77
+
- audio-16khz-128kbitrate-mono-mp3
78
+
- audio-24khz-48kbitrate-mono-mp3
79
+
- audio-24khz-96kbitrate-mono-mp3
80
+
- audio-24khz-160kbitrate-mono-mp3
74
81
75
82
## Getting results
76
83
77
-
With batch synthesis API, use the URL from the `outputs.result` property of the HTTP GET batch synthesis response. The [results](batch-synthesis.md#batch-synthesis-results) are in a ZIP file that contains the audio (such as `0001.wav`), summary, and debug details.
84
+
With batch synthesis API, use the URL from the `outputs.result` property of the HTTP GET batch synthesis response. The [results](batch-synthesis.md#batch-synthesis-results) are in a ZIP file that contains the audio (such as `0001.wav`), summary, and debug details.
78
85
79
86
Long Audio API text inputs and results are returned via two separate content URLs as shown in the following example. The one with `"kind": "LongAudioSynthesisScript"` is the input script submitted. The other one with `"kind": "LongAudioSynthesisResult"` is the result of this request. Both ZIP files can be downloaded from the URL in their `links.contentUrl` property.
80
87
81
88
## Cleaning up resources
82
89
83
-
Batch synthesis API supports up to 200 batch synthesis jobs that don't have a status of "Succeeded" or "Failed". The Speech service keeps each synthesis history for up to 31 days, or the duration of the request `timeToLive` property, whichever comes sooner. The date and time of automatic deletion (for synthesis jobs with a status of "Succeeded" or "Failed") is equal to the `lastActionDateTime` + `timeToLive` properties.
90
+
Batch synthesis API supports up to 300 batch synthesis jobs that don't have a status of "Succeeded" or "Failed". The Speech service keeps each synthesis history for up to 31 days, or the duration of the request `timeToLiveInHours` property, whichever comes sooner. The date and time of automatic deletion (for synthesis jobs with a status of "Succeeded" or "Failed") is equal to the `lastActionDateTime` + `timeToLiveInHours` properties.
84
91
85
-
The Long Audio API is limited to 20,000 requests for each Azure subscription account. The Speech service doesn't remove job history automatically. You must remove the previous job run history before making new requests that would otherwise exceed the limit.
92
+
The Long Audio API is limited to 20,000 requests for each Azure subscription account. The Speech service doesn't remove job history automatically. You must remove the previous job run history before making new requests that would otherwise exceed the limit.
Copy file name to clipboardExpand all lines: articles/ai-services/speech-service/regions.md
+62-62Lines changed: 62 additions & 62 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -17,9 +17,9 @@ The Speech service allows your application to convert audio to text, perform spe
17
17
18
18
Keep in mind the following points:
19
19
20
-
* If your application uses a [Speech SDK](speech-sdk.md), you provide the region identifier, such as `westus`, when you create a `SpeechConfig`. Make sure the region matches the region of your subscription.
21
-
* If your application uses one of the Speech service REST APIs, the region is part of the endpoint URI you use when making requests.
22
-
* Keys created for a region are valid only in that region. If you attempt to use them with other regions, you get authentication errors.
20
+
- If your application uses a [Speech SDK](speech-sdk.md), you provide the region identifier, such as `westus`, when you create a `SpeechConfig`. Make sure the region matches the region of your subscription.
21
+
- If your application uses one of the Speech service REST APIs, the region is part of the endpoint URI you use when making requests.
22
+
- Keys created for a region are valid only in that region. If you attempt to use them with other regions, you get authentication errors.
23
23
24
24
> [!NOTE]
25
25
> Speech service doesn't store or process customer data outside the region the customer deploys the service instance in.
@@ -28,44 +28,44 @@ Keep in mind the following points:
28
28
29
29
The following regions are supported for Speech service features such as speech to text, text to speech, pronunciation assessment, and translation. The geographies are listed in alphabetical order.
30
30
31
-
| Geography | Region | Region identifier |
32
-
| ----- | ----- | ----- |
33
-
| Africa | South Africa North |`southafricanorth` <sup>6</sup>|
34
-
| Asia Pacific | East Asia |`eastasia` <sup>5</sup>|
35
-
| Asia Pacific | Southeast Asia |`southeastasia` <sup>1,2,3,4,5,7,9</sup>|
36
-
| Asia Pacific | Australia East |`australiaeast` <sup>1,2,3,4,7</sup>|
37
-
| Asia Pacific | Central India |`centralindia` <sup>1,2,3,4,5</sup>|
38
-
| Asia Pacific | Japan East |`japaneast` <sup>2,5</sup>|
39
-
| Asia Pacific | Japan West |`japanwest`|
40
-
| Asia Pacific | Korea Central |`koreacentral` <sup>2</sup>|
41
-
| Canada | Canada Central |`canadacentral` <sup>1</sup>|
42
-
| Europe | North Europe |`northeurope` <sup>1,2,4,5,7</sup>|
43
-
| Europe | West Europe |`westeurope` <sup>1,2,3,4,5,7,9</sup>|
44
-
| Europe | France Central |`francecentral`|
45
-
| Europe | Germany West Central |`germanywestcentral`|
46
-
| Europe | Norway East |`norwayeast`|
47
-
| Europe | Sweden Central |`swedencentral`<sup>8</sup> |
48
-
| Europe | Switzerland North |`switzerlandnorth` <sup>6</sup>|
49
-
| Europe | Switzerland West |`switzerlandwest`|
50
-
| Europe | UK South |`uksouth` <sup>1,2,3,4,7</sup>|
51
-
| Middle East | UAE North |`uaenorth` <sup>6</sup>|
52
-
| South America | Brazil South |`brazilsouth` <sup>6</sup>|
53
-
| Qatar | Qatar Central |`qatarcentral`<sup>8</sup> |
54
-
| US | Central US |`centralus`|
55
-
| US | East US |`eastus` <sup>1,2,3,4,5,7,9</sup>|
56
-
| US | East US 2 |`eastus2` <sup>1,2,4,5</sup>|
57
-
| US | North Central US |`northcentralus` <sup>4,6</sup>|
58
-
| US | South Central US |`southcentralus` <sup>1,2,3,4,5,6,7</sup>|
59
-
| US | West Central US |`westcentralus` <sup>5</sup>|
| Africa | South Africa North |`southafricanorth` <sup>6</sup>|
34
+
| Asia Pacific | East Asia |`eastasia` <sup>5</sup>|
35
+
| Asia Pacific | Southeast Asia |`southeastasia` <sup>1,2,4,5,7,9</sup>|
36
+
| Asia Pacific | Australia East |`australiaeast` <sup>1,2,4,7</sup>|
37
+
| Asia Pacific | Central India |`centralindia` <sup>1,2,4,5</sup>|
38
+
| Asia Pacific | Japan East |`japaneast` <sup>2,5</sup>|
39
+
| Asia Pacific | Japan West |`japanwest` <sup>3</sup> |
40
+
| Asia Pacific | Korea Central |`koreacentral` <sup>2</sup>|
41
+
| Canada | Canada Central |`canadacentral` <sup>1</sup>|
42
+
| Europe | North Europe |`northeurope` <sup>1,2,4,5,7</sup>|
43
+
| Europe | West Europe |`westeurope` <sup>1,2,4,5,7,9</sup>|
44
+
| Europe | France Central |`francecentral`|
45
+
| Europe | Germany West Central |`germanywestcentral`|
46
+
| Europe | Norway East |`norwayeast`|
47
+
| Europe | Sweden Central |`swedencentral`<sup>8</sup>|
48
+
| Europe | Switzerland North |`switzerlandnorth` <sup>6</sup>|
49
+
| Europe | Switzerland West |`switzerlandwest` <sup>3</sup> |
50
+
| Europe | UK South |`uksouth` <sup>1,2,4,7</sup>|
51
+
| Middle East | UAE North |`uaenorth` <sup>6</sup>|
52
+
| South America | Brazil South |`brazilsouth` <sup>6</sup>|
53
+
| Qatar | Qatar Central |`qatarcentral`<sup>3,8</sup>|
54
+
| US | Central US |`centralus`|
55
+
| US | East US |`eastus` <sup>1,2,4,5,7,9</sup>|
56
+
| US | East US 2 |`eastus2` <sup>1,2,4,5</sup>|
57
+
| US | North Central US |`northcentralus` <sup>4,6</sup>|
58
+
| US | South Central US |`southcentralus` <sup>1,2,4,5,6,7</sup>|
59
+
| US | West Central US |`westcentralus` <sup>3,5</sup>|
60
+
| US | West US |`westus` <sup>2,5</sup>|
61
+
| US | West US 2 |`westus2` <sup>1,2,4,5,7</sup>|
62
+
| US | West US 3 |`westus3` <sup>3</sup> |
63
63
64
64
<sup>1</sup> The region has dedicated hardware for custom speech training. If you plan to train a custom model with audio data, use one of the regions with dedicated hardware for faster training. Then you can [copy the trained model](how-to-custom-speech-train-model.md#copy-a-model) to another region.
65
65
66
66
<sup>2</sup> The region is available for custom neural voice training. You can copy a trained neural voice model to other regions for deployment.
67
67
68
-
<sup>3</sup> The Long Audio API is available in the region.
68
+
<sup>3</sup> The region doesn't support Batch Synthesis API.
69
69
70
70
<sup>4</sup> The region supports custom keyword advanced models.
71
71
@@ -84,36 +84,36 @@ The following regions are supported for Speech service features such as speech t
84
84
Available regions for intent recognition via the Speech SDK are in the following table.
| North America | South Central US |`southcentralus`|
96
+
| North America | West Central US |`westcentralus`|
97
+
| North America | West US |`westus`|
98
+
| North America | West US 2 |`westus2`|
99
+
| South America | Brazil South |`brazilsouth`|
100
100
101
101
This is a subset of the publishing regions supported by the [Language Understanding service (LUIS)](../luis/luis-reference-regions.md).
102
102
103
103
## Voice assistants
104
104
105
105
The [Speech SDK](speech-sdk.md) supports voice assistant capabilities through [Direct Line Speech](./direct-line-speech.md) for regions in the following table.
Copy file name to clipboardExpand all lines: articles/ai-studio/concepts/evaluation-approach-gen-ai.md
+3-3Lines changed: 3 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -2,15 +2,15 @@
2
2
title: Evaluation of generative AI applications with Azure AI Studio
3
3
titleSuffix: Azure AI Studio
4
4
description: Explore the broader domain of monitoring and evaluating large language models through the establishment of precise metrics, the development of test sets for measurement, and the implementation of iterative testing.
0 commit comments