Merge pull request #6527 from andreasjl/andreasl/SmallRealtimeApiFixes

Small fixes for Realtime API

Author: prmerger-automator[bot]

articles/ai-foundry/openai/quotas-limits.md

@@ -230,7 +230,7 @@ The following section provides you with a quick guide to the default quotas and
 
 ### gpt-4o audio
 
-The rate limits for each `gpt-4o` audio model deployment are 100,000 tokens per minute and 1,000 requests per minute. During the preview, [Azure AI Foundry portal](https://ai.azure.com/?cid=learnDocs) and APIs might inaccurately show different rate limits. Even if you try to set a different rate limit, the actual rate limit is 100,000 tokens per minute and 1,000 requests per minute.
+During the preview, the rate limits for each `gpt-4o` realtime model deployment is at least 100,000 tokens per minute and 1,000 requests per minute, even if a lower limit is shown in [Azure AI Foundry portal](https://ai.azure.com/?cid=learnDocs).
 
 | Model|Tier| Quota limit in tokens per minute | Requests per minute |
 |---|---|:---:|:---:|

articles/ai-foundry/openai/realtime-audio-reference.md

@@ -1658,10 +1658,10 @@ Currently, only 'function' tools are supported.
 | type | [RealtimeTurnDetectionType](#realtimeturndetectiontype) | The type of turn detection.<br><br>Allowed values: `semantic_vad` or `server_vad` |
 | threshold | number | The activation threshold for the server VAD (`server_vad`) turn detection. In noisy environments, you might need to increase the threshold to avoid false positives. In quiet environments, you might need to decrease the threshold to avoid false negatives.<br><br>Defaults to `0.5`. You can set the threshold to a value between `0.0` and `1.0`.<br/><br>This property is only applicable for `server_vad` turn detection. |
 | prefix_padding_ms | string | The duration of speech audio (in milliseconds) to include before the start of detected speech.<br><br>Defaults to `300` milliseconds.<br/><br>This property is only applicable for `server_vad` turn detection. |
-| silence_duration_ms | string | The duration of silence (in milliseconds) to detect the end of speech. You want to detect the end of speech as soon as possible, but not too soon to avoid cutting off the last part of the speech.<br><br>The model will respond more quickly if you set this value to a lower number, but it might cut off the last part of the speech. If you set this value to a higher number, the model will wait longer to detect the end of speech, but it might take longer to respond.<br><br>Defaults to `500` milliseconds.<br/><br>This property is only applicable for `server_vad` turn detection. |
+| silence_duration_ms | string | The duration of silence (in milliseconds) to detect the end of speech. You want to detect the end of speech as soon as possible, but not too soon to avoid cutting off the last part of the speech.<br><br>The model will respond more quickly if you set this value to a lower number, but it might cut off the last part of the speech. If you set this value to a higher number, the model will wait longer to detect the end of speech, but it might take longer to respond.<br><br>Defaults to `200` milliseconds.<br/><br>This property is only applicable for `server_vad` turn detection. |
 | create_response | boolean | Indicates whether the server will automatically create a response when VAD is enabled and speech stops.<br><br>Defaults to `true`. |
 | interrupt_response | boolean | Indicates whether the server will automatically interrupt any ongoing response with output to the default (`auto`) conversation when a VAD start event occurs.<br><br>Defaults to `true`. |
-| eagerness | string | The eagerness of the model to respond and interrupt the user. Specify `low` to wait longer for the user to continue speaking. Specify `high` to chunk the audio as soon as possible for quicker responses. The default value is `auto` that's equivalent to medium.<br/><br>This property is only applicable for `server_vad` turn detection. |
+| eagerness | string | The eagerness of the model to respond and interrupt the user. Specify `low` to wait longer for the user to continue speaking. Specify `high` to chunk the audio as soon as possible for quicker responses. The default value is `auto` that's equivalent to medium.<br/><br>This property is only applicable for `semantic_vad` turn detection. |
 
 ### RealtimeTurnDetectionType