Skip to content

Commit e69876c

Browse files
Court72Court72
authored
Merge pull request #117655 from filipw/main
Updated the C# and Python OpenAI text-to-speech samples to latest code/SDK/API versions
2 parents 0e91d09 + 2f2295b commit e69876c

File tree

3 files changed

+165
-162
lines changed

3 files changed

+165
-162
lines changed

‎articles/ai-services/speech-service/includes/common/environment-variables-openai.md

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -14,6 +14,7 @@ Your application must be authenticated to access Azure AI services resources. Fo
1414
To set the environment variables, open a console window, and follow the instructions for your operating system and development environment.
1515
- To set the `OPEN_AI_KEY` environment variable, replace `your-openai-key` with one of the keys for your resource.
1616
- To set the `OPEN_AI_ENDPOINT` environment variable, replace `your-openai-endpoint` with one of the regions for your resource.
17+
- To set the `OPEN_AI_DEPLOYMENT_NAME` environment variable, replace `your-openai-deployment-name` with one of the regions for your resource.
1718
- To set the `SPEECH_KEY` environment variable, replace `your-speech-key` with one of the keys for your resource.
1819
- To set the `SPEECH_REGION` environment variable, replace `your-speech-region` with one of the regions for your resource.
1920

@@ -22,6 +23,7 @@ To set the environment variables, open a console window, and follow the instruct
2223
```console
2324
setx OPEN_AI_KEY your-openai-key
2425
setx OPEN_AI_ENDPOINT your-openai-endpoint
26+
setx OPEN_AI_DEPLOYMENT_NAME=your-openai-deployment-name
2527
setx SPEECH_KEY your-speech-key
2628
setx SPEECH_REGION your-speech-region
2729
```
@@ -36,6 +38,7 @@ After you add the environment variables, you may need to restart any running pro
3638
```bash
3739
export OPEN_AI_KEY=your-openai-key
3840
export OPEN_AI_ENDPOINT=your-openai-endpoint
41+
export OPEN_AI_DEPLOYMENT_NAME=your-openai-deployment-name
3942
export SPEECH_KEY=your-speech-key
4043
export SPEECH_REGION=your-speech-region
4144
```
@@ -51,6 +54,7 @@ Edit your .bash_profile, and add the environment variables:
5154
```bash
5255
export OPEN_AI_KEY=your-openai-key
5356
export OPEN_AI_ENDPOINT=your-openai-endpoint
57+
export OPEN_AI_DEPLOYMENT_NAME=your-openai-deployment-name
5458
export SPEECH_KEY=your-speech-key
5559
export SPEECH_REGION=your-speech-region
5660
```

‎articles/ai-services/speech-service/includes/quickstarts/openai-speech/csharp.md

Lines changed: 126 additions & 130 deletions
Original file line numberDiff line numberDiff line change
@@ -19,7 +19,7 @@ The Speech SDK is available as a [NuGet package](https://www.nuget.org/packages/
1919

2020
### Set environment variables
2121

22-
This example requires environment variables named `OPEN_AI_KEY`, `OPEN_AI_ENDPOINT`, `SPEECH_KEY`, and `SPEECH_REGION`.
22+
This example requires environment variables named `OPEN_AI_KEY`, `OPEN_AI_ENDPOINT`, `OPEN_AI_DEPLOYMENT_NAME`, `SPEECH_KEY`, and `SPEECH_REGION`.
2323

2424
[!INCLUDE [Environment variables](../../common/environment-variables-openai.md)]
2525

@@ -42,156 +42,152 @@ Follow these steps to create a new console application.
4242
1. Replace the contents of `Program.cs` with the following code.
4343
4444
```csharp
45-
using System;
46-
using System.IO;
47-
using System.Threading.Tasks;
45+
using System.Text;
4846
using Microsoft.CognitiveServices.Speech;
4947
using Microsoft.CognitiveServices.Speech.Audio;
5048
using Azure;
5149
using Azure.AI.OpenAI;
52-
using static System.Environment;
53-
54-
class Program
50+
51+
// This example requires environment variables named "OPEN_AI_KEY", "OPEN_AI_ENDPOINT" and "OPEN_AI_DEPLOYMENT_NAME"
52+
// Your endpoint should look like the following https://YOUR_OPEN_AI_RESOURCE_NAME.openai.azure.com/
53+
string openAIKey = Environment.GetEnvironmentVariable("OPEN_AI_KEY") ??
54+
throw new ArgumentException("Missing OPEN_AI_KEY");
55+
string openAIEndpoint = Environment.GetEnvironmentVariable("OPEN_AI_ENDPOINT") ??
56+
throw new ArgumentException("Missing OPEN_AI_ENDPOINT");
57+
58+
// Enter the deployment name you chose when you deployed the model.
59+
string engine = Environment.GetEnvironmentVariable("OPEN_AI_DEPLOYMENT_NAME") ??
60+
throw new ArgumentException("Missing OPEN_AI_DEPLOYMENT_NAME");
61+
62+
// This example requires environment variables named "SPEECH_KEY" and "SPEECH_REGION"
63+
string speechKey = Environment.GetEnvironmentVariable("SPEECH_KEY") ??
64+
throw new ArgumentException("Missing SPEECH_KEY");
65+
string speechRegion = Environment.GetEnvironmentVariable("SPEECH_REGION") ??
66+
throw new ArgumentException("Missing SPEECH_REGION");
67+
68+
// Sentence end symbols for splitting the response into sentences.
69+
List<string> sentenceSaperators = new() { ".", "!", "?", ";", "。", "！", "？", "；", "\n" };
70+
71+
try
72+
{
73+
await ChatWithOpenAI();
74+
}
75+
catch (Exception ex)
76+
{
77+
Console.WriteLine(ex);
78+
}
79+
80+
// Prompts Azure OpenAI with a request and synthesizes the response.
81+
async Task AskOpenAI(string prompt)
5582
{
56-
// This example requires environment variables named "OPEN_AI_KEY" and "OPEN_AI_ENDPOINT"
57-
// Your endpoint should look like the following https://YOUR_OPEN_AI_RESOURCE_NAME.openai.azure.com/
58-
static string openAIKey = Environment.GetEnvironmentVariable("OPEN_AI_KEY");
59-
static string openAIEndpoint = Environment.GetEnvironmentVariable("OPEN_AI_ENDPOINT");
60-
61-
// Enter the deployment name you chose when you deployed the model.
62-
static string engine = "gpt-35-turbo-instruct";
63-
64-
// This example requires environment variables named "SPEECH_KEY" and "SPEECH_REGION"
65-
static string speechKey = Environment.GetEnvironmentVariable("SPEECH_KEY");
66-
static string speechRegion = Environment.GetEnvironmentVariable("SPEECH_REGION");
67-
68-
// Sentence end symbols for splitting the response into sentences.
69-
static List<string> sentenceSaperators = new() { ".", "!", "?", ";", "。", "！", "？", "；", "\n" };
70-
71-
private static object consoleLock = new();
72-
73-
// Prompts Azure OpenAI with a request and synthesizes the response.
74-
async static Task AskOpenAI(string prompt)
83+
object consoleLock = new();
84+
var speechConfig = SpeechConfig.FromSubscription(speechKey, speechRegion);
85+
86+
// The language of the voice that speaks.
87+
speechConfig.SpeechSynthesisVoiceName = "en-US-JennyMultilingualNeural";
88+
var audioOutputConfig = AudioConfig.FromDefaultSpeakerOutput();
89+
using var speechSynthesizer = new SpeechSynthesizer(speechConfig, audioOutputConfig);
90+
speechSynthesizer.Synthesizing += (sender, args) =>
7591
{
76-
var speechConfig = SpeechConfig.FromSubscription(speechKey, speechRegion);
77-
// The language of the voice that speaks.
78-
speechConfig.SpeechSynthesisVoiceName = "en-US-JennyMultilingualNeural";
79-
var audioOutputConfig = AudioConfig.FromDefaultSpeakerOutput();
80-
using var speechSynthesizer = new SpeechSynthesizer(speechConfig, audioOutputConfig);
81-
speechSynthesizer.Synthesizing += (sender, args) =>
92+
lock (consoleLock)
8293
{
83-
lock (consoleLock)
84-
{
85-
Console.ForegroundColor = ConsoleColor.Yellow;
86-
Console.Write($"[Audio]");
87-
Console.ResetColor();
88-
}
89-
};
90-
91-
// Ask Azure OpenAI
92-
OpenAIClient client = new(new Uri(openAIEndpoint), new AzureKeyCredential(openAIKey));
93-
var completionsOptions = new CompletionsOptions()
94+
Console.ForegroundColor = ConsoleColor.Yellow;
95+
Console.Write($"[Audio]");
96+
Console.ResetColor();
97+
}
98+
};
99+
100+
// Ask Azure OpenAI
101+
OpenAIClient client = new(new Uri(openAIEndpoint), new AzureKeyCredential(openAIKey));
102+
var completionsOptions = new ChatCompletionsOptions()
103+
{
104+
DeploymentName = engine,
105+
Messages = { new ChatRequestUserMessage(prompt) },
106+
MaxTokens = 100,
107+
};
108+
var responseStream = await client.GetChatCompletionsStreamingAsync(completionsOptions);
109+
110+
StringBuilder gptBuffer = new();
111+
await foreach (var completionUpdate in responseStream)
112+
{
113+
var message = completionUpdate.ContentUpdate;
114+
if (string.IsNullOrEmpty(message))
94115
{
95-
Prompts = { prompt },
96-
MaxTokens = 100,
97-
98-
};
99-
var responseStream = await client.GetCompletionsStreamingAsync(engine, completionsOptions);
100-
using var streamingCompletions = responseStream.Value;
101-
StringBuilder gptBuffer = new();
102-
await foreach (var choice in streamingCompletions.GetChoicesStreaming())
116+
continue;
117+
}
118+
119+
lock (consoleLock)
103120
{
104-
await foreach (var message in choice.GetTextStreaming())
105-
{
106-
if (string.IsNullOrEmpty(message))
107-
{
108-
continue;
109-
}
110-
111-
lock (consoleLock)
112-
{
113-
Console.ForegroundColor = ConsoleColor.DarkBlue;
114-
Console.Write($"{message}");
115-
Console.ResetColor();
116-
}
117-
118-
gptBuffer.Append(message);
119-
120-
if (sentenceSaperators.Any(message.Contains))
121-
{
122-
var sentence = gptBuffer.ToString().Trim();
123-
if (!string.IsNullOrEmpty(sentence))
124-
{
125-
await speechSynthesizer.SpeakTextAsync(sentence).ConfigureAwait(true);
126-
gptBuffer.Clear();
127-
}
128-
}
129-
}
121+
Console.ForegroundColor = ConsoleColor.DarkBlue;
122+
Console.Write($"{message}");
123+
Console.ResetColor();
130124
}
131-
}
132-
133-
// Continuously listens for speech input to recognize and send as text to Azure OpenAI
134-
async static Task ChatWithOpenAI()
135-
{
136-
// Should be the locale for the speaker's language.
137-
var speechConfig = SpeechConfig.FromSubscription(speechKey, speechRegion);
138-
speechConfig.SpeechRecognitionLanguage = "en-US";
139-
140-
using var audioConfig = AudioConfig.FromDefaultMicrophoneInput();
141-
using var speechRecognizer = new SpeechRecognizer(speechConfig, audioConfig);
142-
var conversationEnded = false;
143-
144-
while(!conversationEnded)
125+
126+
gptBuffer.Append(message);
127+
128+
if (sentenceSaperators.Any(message.Contains))
145129
{
146-
Console.WriteLine("Azure OpenAI is listening. Say 'Stop' or press Ctrl-Z to end the conversation.");
147-
148-
// Get audio from the microphone and then send it to the TTS service.
149-
var speechRecognitionResult = await speechRecognizer.RecognizeOnceAsync();
150-
151-
switch (speechRecognitionResult.Reason)
130+
var sentence = gptBuffer.ToString().Trim();
131+
if (!string.IsNullOrEmpty(sentence))
152132
{
153-
case ResultReason.RecognizedSpeech:
154-
if (speechRecognitionResult.Text == "Stop.")
155-
{
156-
Console.WriteLine("Conversation ended.");
157-
conversationEnded = true;
158-
}
159-
else
160-
{
161-
Console.WriteLine($"Recognized speech: {speechRecognitionResult.Text}");
162-
await AskOpenAI(speechRecognitionResult.Text).ConfigureAwait(true);
163-
}
164-
break;
165-
case ResultReason.NoMatch:
166-
Console.WriteLine($"No speech could be recognized: ");
167-
break;
168-
case ResultReason.Canceled:
169-
var cancellationDetails = CancellationDetails.FromResult(speechRecognitionResult);
170-
Console.WriteLine($"Speech Recognition canceled: {cancellationDetails.Reason}");
171-
if (cancellationDetails.Reason == CancellationReason.Error)
172-
{
173-
Console.WriteLine($"Error details={cancellationDetails.ErrorDetails}");
174-
}
175-
break;
133+
await speechSynthesizer.SpeakTextAsync(sentence);
134+
gptBuffer.Clear();
176135
}
177136
}
178137
}
179-
180-
async static Task Main(string[] args)
138+
}
139+
140+
// Continuously listens for speech input to recognize and send as text to Azure OpenAI
141+
async Task ChatWithOpenAI()
142+
{
143+
// Should be the locale for the speaker's language.
144+
var speechConfig = SpeechConfig.FromSubscription(speechKey, speechRegion);
145+
speechConfig.SpeechRecognitionLanguage = "en-US";
146+
147+
using var audioConfig = AudioConfig.FromDefaultMicrophoneInput();
148+
using var speechRecognizer = new SpeechRecognizer(speechConfig, audioConfig);
149+
var conversationEnded = false;
150+
151+
while (!conversationEnded)
181152
{
182-
try
183-
{
184-
await ChatWithOpenAI().ConfigureAwait(true);
185-
}
186-
catch (Exception ex)
153+
Console.WriteLine("Azure OpenAI is listening. Say 'Stop' or press Ctrl-Z to end the conversation.");
154+
155+
// Get audio from the microphone and then send it to the TTS service.
156+
var speechRecognitionResult = await speechRecognizer.RecognizeOnceAsync();
157+
158+
switch (speechRecognitionResult.Reason)
187159
{
188-
Console.WriteLine(ex.Message);
160+
case ResultReason.RecognizedSpeech:
161+
if (speechRecognitionResult.Text == "Stop.")
162+
{
163+
Console.WriteLine("Conversation ended.");
164+
conversationEnded = true;
165+
}
166+
else
167+
{
168+
Console.WriteLine($"Recognized speech: {speechRecognitionResult.Text}");
169+
await AskOpenAI(speechRecognitionResult.Text);
170+
}
171+
172+
break;
173+
case ResultReason.NoMatch:
174+
Console.WriteLine($"No speech could be recognized: ");
175+
break;
176+
case ResultReason.Canceled:
177+
var cancellationDetails = CancellationDetails.FromResult(speechRecognitionResult);
178+
Console.WriteLine($"Speech Recognition canceled: {cancellationDetails.Reason}");
179+
if (cancellationDetails.Reason == CancellationReason.Error)
180+
{
181+
Console.WriteLine($"Error details={cancellationDetails.ErrorDetails}");
182+
}
183+
184+
break;
189185
}
190186
}
191187
}
192188
```
193189
194-
1. To increase or decrease the number of tokens returned by Azure OpenAI, change the `MaxTokens` property in the `CompletionsOptions` class instance. For more information tokens and cost implications, see [Azure OpenAI tokens](/azure/ai-services/openai/overview#tokens) and [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/).
190+
1. To increase or decrease the number of tokens returned by Azure OpenAI, change the `MaxTokens` property in the `ChatCompletionsOptions` class instance. For more information tokens and cost implications, see [Azure OpenAI tokens](/azure/ai-services/openai/overview#tokens) and [Azure OpenAI pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/).
195191
196192
Run your new console application to start speech recognition from a microphone:
197193
@@ -200,7 +196,7 @@ dotnet run
200196
```
201197

202198
> [!IMPORTANT]
203-
> Make sure that you set the `OPEN_AI_KEY`, `OPEN_AI_ENDPOINT`, `SPEECH_KEY` and `SPEECH_REGION` environment variables as described [previously](#set-environment-variables). If you don't set these variables, the sample will fail with an error message.
199+
> Make sure that you set the `OPEN_AI_KEY`, `OPEN_AI_ENDPOINT`, `OPEN_AI_DEPLOYMENT_NAME`, `SPEECH_KEY` and `SPEECH_REGION` environment variables as described [previously](#set-environment-variables). If you don't set these variables, the sample will fail with an error message.
204200
205201
Speak into your microphone when prompted. The console output includes the prompt for you to begin speaking, then your request as text, and then the response from Azure OpenAI as text. The response from Azure OpenAI should be converted from text to speech and then output to the default speaker.
206202

0 commit comments

Comments
 (0)