当前位置：首页 > PHP

php怎么实现语音

2026-02-16 13:39:34PHP

PHP实现语音功能的方法

PHP本身是服务器端脚本语言，不直接处理语音功能，但可以通过调用外部服务或库实现语音相关功能。以下是几种常见方法：

使用第三方语音API

许多云服务提供商提供语音识别和合成API，可以通过HTTP请求调用：

$text = "你好，这是一段测试语音";
$apiKey = "YOUR_API_KEY";
$url = "https://texttospeech.googleapis.com/v1/text:synthesize";

$data = [
    'input' => ['text' => $text],
    'voice' => ['languageCode' => 'zh-CN', 'name' => 'cmn-CN-Wavenet-A'],
    'audioConfig' => ['audioEncoding' => 'MP3']
];

$options = [
    'http' => [
        'header' => "Content-Type: application/json\r\nAuthorization: Bearer $apiKey\r\n",
        'method' => 'POST',
        'content' => json_encode($data)
    ]
];

$context = stream_context_create($options);
$result = file_get_contents($url, false, $context);
$response = json_decode($result, true);

file_put_contents('output.mp3', base64_decode($response['audioContent']));

使用本地语音合成引擎

通过系统命令调用本地安装的语音合成工具：

$text = "Hello world";
$outputFile = "speech.wav";
$command = "espeak -v en \"$text\" --stdout > $outputFile";
shell_exec($command);

实现语音识别功能

使用语音识别API将用户上传的语音转换为文本：

if(isset($_FILES['audio'])){
    $audioFile = $_FILES['audio']['tmp_name'];
    $apiUrl = "https://speech.googleapis.com/v1/speech:recognize";
    $audioContent = base64_encode(file_get_contents($audioFile));

    $data = [
        'config' => [
            'encoding' => 'LINEAR16',
            'sampleRateHertz' => 16000,
            'languageCode' => 'zh-CN'
        ],
        'audio' => ['content' => $audioContent]
    ];

    $options = [
        'http' => [
            'header' => "Content-Type: application/json\r\nAuthorization: Bearer $apiKey\r\n",
            'method' => 'POST',
            'content' => json_encode($data)
        ]
    ];

    $context = stream_context_create($options);
    $result = file_get_contents($apiUrl, false, $context);
    $transcript = json_decode($result, true);
    echo $transcript['results'][0]['alternatives'][0]['transcript'];
}

网页语音API集成

在PHP生成的HTML页面中嵌入Web Speech API实现浏览器端语音功能：

echo '<script>
    function startSpeechRecognition() {
        const recognition = new webkitSpeechRecognition();
        recognition.lang = "zh-CN";
        recognition.onresult = function(event) {
            document.getElementById("result").value = event.results[0][0].transcript;
        };
        recognition.start();
    }
</script>
<button onclick="startSpeechRecognition()">开始语音输入</button>
<input type="text" id="result">';