Merging changes synced from https://github.com/MicrosoftDocs/azure-docs-pr (branch live)

Author: Learn Build Service GitHub App

.openpublishing.redirection.machine-configuration.json

@@ -48,6 +48,11 @@ Here's an overview of the steps to create a custom neural voice in Speech Studio
 
 You can tune, adjust, and use your custom voice, similarly as you would use a prebuilt neural voice. Convert text into speech in real-time, or generate audio content offline with text input. You use the [REST API](./rest-text-to-speech.md), the [Speech SDK](./get-started-text-to-speech.md), or the [Speech Studio](https://speech.microsoft.com/audiocontentcreation).
 
+> [!TIP]
+> You can also use the Speech SDK and custom voice REST API to train a custom neural voice.
+> 
+> Check out the code samples in the [Speech SDK repository on GitHub](https://github.com/Azure-Samples/cognitive-services-speech-sdk/blob/master/samples/custom-voice/README.md) to see how to use personal voice in your application.
+
 The style and the characteristics of the trained voice model depend on the style and the quality of the recordings from the voice talent used for training. However, you can make several adjustments by using [SSML (Speech Synthesis Markup Language)](./speech-synthesis-markup.md?tabs=csharp) when you make the API calls to your voice model to generate synthetic speech. SSML is the markup language used to communicate with the text to speech service to convert text into audio. The adjustments you can make include change of pitch, rate, intonation, and pronunciation correction. If the voice model is built with multiple styles, you can also use SSML to switch the styles.
 
 ## Components sequence

articles/active-directory-b2c/media/tutorial-create-tenant/create-a-resource.png

@@ -22,6 +22,9 @@ OpenAI text to speech voices are available via two model variants: `Neural` and
 - `Neural`: Optimized for real-time use cases with the lowest latency, but lower quality than `NeuralHD`.
 - `NeuralHD`: Optimized for quality.
 
+For a demonstration of OpenAI voices in Azure OpenAI Studio and Speech Studio, view this [introductory video](https://youtu.be/Ic505XeV3gs).
+> [!VIDEO https://www.youtube.com/embed/Ic505XeV3gs]
+
 ## Available text to speech voices in Azure AI services
 
 You might ask: If I want to use an OpenAI text to speech voice, should I use it via the Azure OpenAI Service or via Azure AI Speech? What are the scenarios that guide me to use one or the other?

articles/active-directory-b2c/media/tutorial-create-tenant/portal-02-create-tenant.png

@@ -63,6 +63,9 @@ To get started, here's a summary of the steps to create a personal voice:
 
 Once you have a personal voice, you can [use it](./personal-voice-how-to-use.md) to synthesize speech in any of the 91 languages supported across 100+ locales. A locale tag isn't required. Personal voice uses automatic language detection at the sentence level. For more information, see [use personal voice in your application](./personal-voice-how-to-use.md).
 
+> [!TIP]
+> Check out the code samples in the [Speech SDK repository on GitHub](https://github.com/Azure-Samples/cognitive-services-speech-sdk/blob/master/samples/custom-voice/README.md) to see how to use personal voice in your application.
+
 ## Reference documentation
 
 > [!div class="nextstepaction"]

articles/ai-services/speech-service/custom-neural-voice.md

@@ -1,33 +1,87 @@
 ---
-title: Monitor quality and safety of deployed applications 
+title: Monitor quality and safety of deployed prompt flow applications 
 titleSuffix: Azure AI Studio
-description: Learn how to monitor quality and safety of deployed applications with Azure AI Studio.
+description: Learn how to monitor quality and safety of deployed prompt flow applications with Azure AI Studio.
 manager: scottpolly
 ms.service: azure-ai-studio
 ms.custom:
   - ignite-2023
 ms.topic: how-to
-ms.date: 11/15/2023
+ms.date: 2/7/2024
 ms.reviewer: fasantia
 ms.author: mopeakande
 author: msakande
 ---
 
-# Monitor quality and safety of deployed applications 
+# Monitor quality and safety of deployed prompt flow applications 
 
-Monitoring models that are deployed in production is an essential part of the generative AI application lifecycle. Changes in data and consumer behavior can influence your application over time, resulting in outdated systems that negatively affect business outcomes and expose organizations to compliance, economic, and reputational risks. 
+Monitoring models that are deployed in production is an essential part of the generative AI application lifecycle. Changes in data and consumer behavior can influence your application over time, resulting in outdated systems that negatively affect business outcomes and expose organizations to compliance, economic, and reputation risks. 
 
 Azure AI model monitoring for generative AI applications makes it easier for you to monitor your applications in production for safety and quality on a cadence to ensure it's delivering maximum business value. 
 
-Capabilities and integrations include: 
-- Collect production data using Model data collector from a prompt flow deployment.
+Capabilities and integrations for monitoring a prompt flow deployment include: 
+- Collect production data using the model data collector.
 - Apply Responsible AI evaluation metrics such as groundedness, coherence, fluency, relevance, and similarity, which are interoperable with prompt flow evaluation metrics. 
 - Preconfigured alerts and defaults to run monitoring on a recurring basis.
 - Consume result and configure advanced behavior in Azure AI Studio.
 
+## Set up monitoring for prompt flow 
+
+Follow these steps to set up monitoring for your prompt flow deployment:
+
+1. Confirm your flow runs successfully, and that the required inputs and outputs are configured for the [metrics you want to assess](#evaluation-metrics). The minimum required parameters of collecting only inputs and outputs provide only two metrics: coherence and fluency. You must configure your flow according to the [flow and metric configuration requirements](#flow-and-metric-configuration-requirements).
+
+    :::image type="content" source="../media/deploy-monitor/monitor/user-experience.png" alt-text="Screenshot of prompt flow editor with deploy button." lightbox = "../media/deploy-monitor/monitor/user-experience.png":::
+
+1. Deploy your flow. By default, both inferencing data collection and application insights are enabled automatically. These are required for the creation of your monitor. 
+
+    :::image type="content" source="../media/deploy-monitor/monitor/basic-settings.png" alt-text="Screenshot of basic settings in the deployment wizard." lightbox = "../media/deploy-monitor/monitor/basic-settings.png":::
+
+1. By default, all outputs of your deployment are collected using Azure AI's Model Data Collector. As an optional step, you can enter the advanced settings to confirm that your desired columns (for example, context of ground truth) are included in the endpoint response. 
+
+    Your deployed flow needs to be configured in the following way:  
+    - Flow inputs & outputs: You need to name your flow outputs appropriately and remember these column names when creating your monitor. In this article, we use the following settings: 
+      - Inputs (required): "prompt" 
+      - Outputs (required): "completion" 
+      - Outputs (optional): "context" and/or "ground truth" 
+
+    - Data collection: The **inferencing data collection** toggle must be enabled using Model Data Collector 
+
+    - Outputs: In the prompt flow deployment wizard, confirm the required outputs are selected (such as completion, context, and ground_truth) that meet your metric configuration requirements.
+
+1. Test your deployment in the deployment **Test** tab. 
+
+    :::image type="content" source="../media/deploy-monitor/monitor/test-deploy.png" alt-text="Screenshot of the deployment test page." lightbox = "../media/deploy-monitor/monitor/test-deploy.png":::
+
+    > [!NOTE]
+    > Monitoring requires the endpoint to be used at least 10 times to collect enough data to provide insights. If you'd like to test sooner, manually send about 50 rows in the 'test' tab before running the monitor.
+
+1. Create your monitor by either enabling from the deployment details page, or the **Monitoring** tab.
+
+    :::image type="content" source="../media/deploy-monitor/monitor/enable-monitoring.png" alt-text="Screenshot of the button to enable monitoring." lightbox = "../media/deploy-monitor/monitor/enable-monitoring.png":::
+
+1. Ensure your columns are mapped from your flow as defined in the previous requirements. 
+
+    :::image type="content" source="../media/deploy-monitor/monitor/column-map.png" alt-text="Screenshot of columns mapped for monitoring metrics." lightbox = "../media/deploy-monitor/monitor/column-map.png":::
+
+1. View your monitor in the **Monitor** tab.  
+
+    :::image type="content" source="../media/deploy-monitor/monitor/monitor-metrics.png" alt-text="Screenshot of the monitoring result metrics." lightbox = "../media/deploy-monitor/monitor/monitor-metrics.png":::
+
+By default, operational metrics such as requests per minute and request latency show up. The default safety and quality monitoring signal are configured with a 10% sample rate and run on your default workspace Azure Open AI connection. 
+
+Your monitor is created with default settings:
+- 10% sample rate
+- 4/5 (thresholds / recurrence)
+- Weekly recurrence on Monday mornings
+- Alerts are delivered to the inbox of the person that triggered the monitor.
+
+To view more details about your monitoring metrics, you can follow the link to navigate to monitoring in Azure Machine Learning studio, which is a separate studio that allows for more customizations. 
+
+
 ## Evaluation metrics 
 
-Metrics are generated by the following state-of-the-art GPT language models configured with specific evaluation instructions (prompt templates) which act as evaluator models for sequence-to-sequence tasks. This technique has shown strong empirical results and high correlation with human judgment when compared to standard generative AI evaluation metrics. For more information about prompt flow evaluation, see [Submit bulk test and evaluate a flow](./flow-bulk-test-evaluation.md) and [evaluation and monitoring metrics for generative AI](../concepts/evaluation-metrics-built-in.md).
+Metrics are generated by the following state-of-the-art GPT language models configured with specific evaluation instructions (prompt templates) which act as evaluator models for sequence-to-sequence tasks. This technique has strong empirical results and high correlation with human judgment when compared to standard generative AI evaluation metrics. For more information about prompt flow evaluation, see [Submit bulk test and evaluate a flow](./flow-bulk-test-evaluation.md) and [evaluation and monitoring metrics for generative AI](../concepts/evaluation-metrics-built-in.md).
 
 These GPT models are supported with monitoring and configured as your Azure OpenAI resource: 
 
@@ -53,7 +107,7 @@ When creating your flow, you need to ensure your column names are mapped. The fo
 |------|------------|----------|
 | Prompt text | The original prompt given (also known as "inputs" or "question") | Required |
 | Completion text | The final completion from the API call that is returned (also known as "outputs" or "answer") | Required |
-| Context text | Any context data that is sent to the API call, together with original prompt. For example, if you hope to get search results only from certain certified information sources/website, you can define in the evaluation steps. This is an optional step that can be configured through prompt flow. | Optional |
+| Context text | Any context data that is sent to the API call, together with original prompt. For example, if you hope to get search results only from certain certified information sources/website, you can define in the evaluation steps. | Optional |
 | Ground truth text | The user-defined text as the "source of truth" | Optional |
 
 What parameters are configured in your data asset dictates what metrics you can produce, according to this table: 
@@ -68,61 +122,6 @@ What parameters are configured in your data asset dictates what metrics you can
 
 For more information, see [question answering metric requirements](evaluate-generative-ai-app.md#question-answering-metric-requirements).
 
-## User Experience 
-
-Confirm your flow runs successfully, and that the required inputs and outputs are configured for the metrics you want to assess. The minimum required parameters of collecting only inputs and outputs provide only two metrics: coherence and fluency. You must configure your flow according to the [prior guidance](#flow-and-metric-configuration-requirements).
-
-:::image type="content" source="../media/deploy-monitor/monitor/user-experience.png" alt-text="Screenshot of prompt flow editor with deploy button." lightbox = "../media/deploy-monitor/monitor/user-experience.png":::
-
-Deploy your flow. By default, both inferencing data collection and application insights are enabled automatically. These are required for the creation of your monitor. 
-
-:::image type="content" source="../media/deploy-monitor/monitor/basic-settings.png" alt-text="Screenshot of basic settings in the deployment wizard." lightbox = "../media/deploy-monitor/monitor/basic-settings.png":::
-
-By default, all outputs of your deployment are collected using Azure AI's Model Data Collector. As an optional step, you can enter the advanced settings to confirm that your desired columns (for example, context of ground truth) are included in the endpoint response. 
-
-In summary, your deployed flow needs to be configured in the following way:  
-
-- Flow inputs & outputs: You need to name your flow outputs appropriately and remember these column names when creating your monitor. In this article, we use the following: 
-    - Inputs (required): "prompt" 
-    - Outputs (required): "completion" 
-    - Outputs (optional): "context" and/or "ground truth" 
-
-- Data collection: in the "Deployment" (Step #2 of the prompt flow deployment wizard), the 'inference data collection' toggle must be enabled using Model Data Collector 
-
-- Outputs: In the Outputs (Step #3 of the prompt flow deployment wizard), confirm you have selected the required outputs listed above (for example, completion | context | ground_truth) that meet your metric configuration requirements.
-
-Test your deployment in the deployment **Test** tab. 
-
-:::image type="content" source="../media/deploy-monitor/monitor/test-deploy.png" alt-text="Screenshot of the deployment test page." lightbox = "../media/deploy-monitor/monitor/test-deploy.png":::
-
- 
-> [!NOTE]
-> Monitoring requires the endpoint to be used at least 10 times to collect enough data to provide insights. If you'd like to test sooner, manually send about 50 rows in the 'test' tab before running the monitor.
-
-Create your monitor by either enabling from the deployment details page, or the **Monitoring** tab.
-
-:::image type="content" source="../media/deploy-monitor/monitor/enable-monitoring.png" alt-text="Screenshot of the button to enable monitoring." lightbox = "../media/deploy-monitor/monitor/enable-monitoring.png":::
-
-Ensure your columns are mapped from your flow as defined in the previous requirements. 
-
-:::image type="content" source="../media/deploy-monitor/monitor/column-map.png" alt-text="Screenshot of columns mapped for monitoring metrics." lightbox = "../media/deploy-monitor/monitor/column-map.png":::
-
-
-View your monitor in the **Monitor** tab.  
-
-:::image type="content" source="../media/deploy-monitor/monitor/monitor-metrics.png" alt-text="Screenshot of the monitoring result metrics." lightbox = "../media/deploy-monitor/monitor/monitor-metrics.png":::
-
-By default, operational metrics such as requests per minute and request latency show up. The default safety and quality monitoring signal are configured with a 10% sample rate and will run on your default workspace Azure Open AI connection. 
-
-Your monitor is created with default settings:
-- 10% sample rate
-- 4/5 (thresholds / recurrence)
-- Weekly recurrence on Monday mornings
-- Alerts are delivered to the inbox of the person that triggered the monitor.
-
-To view more details about your monitoring metrics, you can follow the link to navigate to monitoring in Azure Machine Learning studio, which is a separate studio that allows for more customizations. 
-
-
 ## Next steps
 
 - Learn more about what you can do in [Azure AI Studio](../what-is-ai-studio.md)

articles/ai-services/speech-service/openai-voices.md

@@ -168,7 +168,7 @@
     - name: Deploy open models
       href: how-to/deploy-models-open.md
       displayName: oss, open source
-    - name: Monitor quality and safety of deployed applications
+    - name: Monitor prompt flow deployments
       href: how-to/monitor-quality-safety.md
     - name: Troubleshoot deployments and monitoring
       href: how-to/troubleshoot-deploy-and-monitor.md

articles/ai-services/speech-service/personal-voice-overview.md

@@ -1,26 +1,26 @@
 ---
 title: Configure an internal load balancer (ILB) endpoint
 titleSuffix: Azure Application Gateway
-description: This article provides information on how to configure Application Gateway Standard v1 with a private frontend IP address
+description: This article provides information on how to configure Application Gateway Standard v2 with a private frontend IP address
 services: application-gateway
 author: greg-lindsay
 ms.service: application-gateway
 ms.topic: how-to
-ms.date: 01/11/2022
+ms.date: 02/07/2024
 ms.author: greglin
 ---
 
 # Configure an application gateway with an internal load balancer (ILB) endpoint
 
-Azure Application Gateway Standard v1 can be configured with an Internet-facing VIP or with an internal endpoint that isn't exposed to the Internet. An internal endpoint uses a private IP address for the frontend, which is also known as an *internal load balancer (ILB) endpoint*.
+Azure Application Gateway Standard v2 can be configured with an Internet-facing VIP or with an internal endpoint that isn't exposed to the Internet. An internal endpoint uses a private IP address for the frontend, which is also known as an *internal load balancer (ILB) endpoint*.
 
 Configuring the gateway using a frontend private IP address is useful for internal line-of-business applications that aren't exposed to the Internet. It's also useful for services and tiers within a multi-tier application that are in a security boundary that isn't exposed to the Internet but:
 
 - still require round-robin load distribution
 - session stickiness
 - or Transport Layer Security (TLS) termination (previously known as Secure Sockets Layer (SSL)).
 
-This article guides you through the steps to configure a Standard v1 Application Gateway with an ILB using the Azure portal.
+This article guides you through the steps to configure a Standard v2 Application Gateway with an ILB using the Azure portal.
 
 [!INCLUDE [updated-for-az](../../includes/updated-for-az.md)]
 
@@ -38,7 +38,7 @@ In this example, you create a new virtual network. You can create a virtual netw
 2. Select **Networking** and then select **Application Gateway** in the Featured list.
 3. Enter *myAppGateway* for the name of the application gateway and *myResourceGroupAG* for the new resource group.
 4. For **Region**, select **Central US**.
-5. For **Tier**, select **Standard**.
+5. For **Tier**, select **Standard V2**.
 6. Under **Configure virtual network** select **Create new**, and then enter these values for the virtual network:
    - *myVNet* - for the name of the virtual network.
    - *10.0.0.0/16* - for the virtual network address space.