shiyue/Microsoft-tts
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: add style selection for TTS and update related functionality

Browse Source
This commit is contained in:
王锦强
2025-03-09 17:21:52 +08:00
parent 4ec09304f6
commit 1cd2ac1624
5 changed files with 70 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
internal/http/handlers/tts.go
View File

@@ -153,6 +153,7 @@ func (h *TTSHandler) HandleTTS(w http.ResponseWriter, r *http.Request) {
Voice: q.Get("v"),
Rate: q.Get("r"),
Pitch: q.Get("p"),
Style: q.Get("s"),
}
case http.MethodPost:
// 从POST JSON体获取
@@ -174,6 +175,7 @@ func (h *TTSHandler) HandleTTS(w http.ResponseWriter, r *http.Request) {
Voice: r.FormValue("voice"),
Rate: r.FormValue("rate"),
Pitch: r.FormValue("pitch"),
Style: r.FormValue("style"),
}
}
default:

1
internal/models/tts.go
View File

@@ -6,6 +6,7 @@ type TTSRequest struct {
Voice string `json:"voice"` // 语音ID
Rate string `json:"rate"` // 语速 (-100% 到 +100%)
Pitch string `json:"pitch"` // 语调 (-100% 到 +100%)
Style string `json:"style"` // 说话风格
}
// TTSResponse 表示一个语音合成响应

9
internal/tts/microsoft/client.go
View File

@@ -25,7 +25,7 @@ const (
ttsEndpoint = "https://%s.tts.speech.microsoft.com/cognitiveservices/v1"
ssmlTemplate = `<speak version='1.0' xmlns='http://www.w3.org/2001/10/synthesis' xmlns:mstts="http://www.w3.org/2001/mstts" xml:lang='%s'>
<voice name='%s'>
<mstts:express-as style="general" styledegree="1.0" role="default">
<mstts:express-as style="%s" styledegree="1.0" role="default">
<prosody rate='%s%%' pitch='%s%%' volume="medium">
%s
</prosody>
@@ -227,6 +227,11 @@ func (c *Client) createTTSRequest(ctx context.Context, req models.TTSRequest) (*
voice = c.defaultVoice
}
style := req.Style
if req.Style == "" {
style = "general"
}
rate := req.Rate
if rate == "" {
rate = c.defaultRate
@@ -249,7 +254,7 @@ func (c *Client) createTTSRequest(ctx context.Context, req models.TTSRequest) (*
escapedText := html.EscapeString(req.Text)
// 准备SSML内容
ssml := fmt.Sprintf(ssmlTemplate, locale, voice, rate, pitch, escapedText)
ssml := fmt.Sprintf(ssmlTemplate, locale, voice, style, rate, pitch, escapedText)
// 获取端点信息
endpoint, err := c.getEndpoint(ctx)

51
web/static/js/app.js
View File

@@ -2,6 +2,7 @@ document.addEventListener('DOMContentLoaded', function() {
// 获取DOM元素
const textInput = document.getElementById('text');
const voiceSelect = document.getElementById('voice');
const styleSelect = document.getElementById('style');
const rateInput = document.getElementById('rate');
const rateValue = document.getElementById('rateValue');
const pitchInput = document.getElementById('pitch');
@@ -15,6 +16,8 @@ document.addEventListener('DOMContentLoaded', function() {
// 保存最后一个音频URL
let lastAudioUrl = '';
// 存储语音数据
let voicesData = [];
// 初始化
initVoicesList();
@@ -37,13 +40,18 @@ document.addEventListener('DOMContentLoaded', function() {
pitchValue.textContent = value + '%';
});
// 语音选择变化时更新可用风格
voiceSelect.addEventListener('change', function() {
updateStyleOptions();
});
// 获取可用语音列表
async function initVoicesList() {
try {
const response = await fetch(`${config.basePath}/voices`);
if (!response.ok) throw new Error('获取语音列表失败');
const voices = await response.json();
voicesData = await response.json();
// 清空并重建选项
voiceSelect.innerHTML = '';
@@ -51,7 +59,7 @@ document.addEventListener('DOMContentLoaded', function() {
// 按语言和名称分组
const voicesByLocale = {};
voices.forEach(voice => {
voicesData.forEach(voice => {
if (!voicesByLocale[voice.locale]) {
voicesByLocale[voice.locale] = [];
}
@@ -78,12 +86,49 @@ document.addEventListener('DOMContentLoaded', function() {
voiceSelect.appendChild(optgroup);
}
// 初始化风格列表
updateStyleOptions();
} catch (error) {
console.error('获取语音列表失败:', error);
voiceSelect.innerHTML = '<option value="">无法加载语音列表</option>';
}
}
// 更新风格选项
function updateStyleOptions() {
// 清空风格选择
styleSelect.innerHTML = '';
// 获取当前选中的语音
const selectedVoice = voiceSelect.value;
const voiceData = voicesData.find(v => v.short_name === selectedVoice);
if (!voiceData || !voiceData.style_list || voiceData.style_list.length === 0) {
// 如果没有可用风格，添加默认选项
const option = document.createElement('option');
option.value = "general";
option.textContent = "普通";
styleSelect.appendChild(option);
return;
}
// 添加可用风格选项
voiceData.style_list.forEach(style => {
const option = document.createElement('option');
option.value = style
option.textContent = style
// 如果是默认风格则选中
if (style === config.defaultStyle ||
(!config.defaultStyle && style === "general")) {
option.selected = true;
}
styleSelect.appendChild(option);
});
}
// 初始化事件监听器
function initEventListeners() {
// 转换按钮点击事件
@@ -137,6 +182,7 @@ document.addEventListener('DOMContentLoaded', function() {
}
const voice = voiceSelect.value;
const style = styleSelect.value;
const rate = rateInput.value;
const pitch = pitchInput.value;
@@ -149,6 +195,7 @@ document.addEventListener('DOMContentLoaded', function() {
const params = new URLSearchParams({
t: text,
v: voice,
s: style,
r: rate,
p: pitch
});

14
web/templates/index.html
View File

@@ -34,15 +34,22 @@
</select>
</div>
<div class="setting-group">
<label for="style">风格:</label>
<select id="style">
<option value="loading">加载中...</option>
</select>
</div>
<div class="setting-group">
<label for="rate">语速:</label>
<input type="range" id="rate" min="-50" max="50" value="0">
<input type="range" id="rate" min="-100" max="100" value="0">
<span id="rateValue">0%</span>
</div>
<div class="setting-group">
<label for="pitch">语调:</label>
<input type="range" id="pitch" min="-50" max="50" value="0">
<input type="range" id="pitch" min="-100" max="100" value="0">
<span id="pitchValue">0%</span>
</div>
</div>
@@ -75,7 +82,8 @@
basePath: "{{.BasePath}}",
defaultVoice: "{{.DefaultVoice}}",
defaultRate: "{{.DefaultRate}}",
defaultPitch: "{{.DefaultPitch}}"
defaultPitch: "{{.DefaultPitch}}",
defaultStyle: "{{.DefaultStyle}}"
};
</script>
<script src="{{.BasePath}}/static/js/app.js"></script>