feat: 语音增加预览功能 (liou666#12)

0x7c01 · Apr 20, 2023 · a663a6b · a663a6b
1 parent 09c66c1
commit a663a6b
Show file tree

Hide file tree

Showing 3 changed files with 67 additions and 31 deletions.
diff --git a/src/hooks/useSpeechService.ts b/src/hooks/useSpeechService.ts
@@ -6,8 +6,11 @@ import {
   SpeechRecognizer,
   SpeechSynthesizer,
 } from 'microsoft-cognitiveservices-speech-sdk'
-
-export const useSpeechService = (langs = <const>['fr-FR', 'ja-JP', 'en-US', 'zh-CN', 'zh-HK', 'ko-KR', 'de-DE']) => {
+interface Config {
+  langs?: readonly['fr-FR', 'ja-JP', 'en-US', 'zh-CN', 'zh-HK', 'ko-KR', 'de-DE']
+  isFetchAllVoice?: boolean
+}
+export const useSpeechService = ({ langs = <const>['fr-FR', 'ja-JP', 'en-US', 'zh-CN', 'zh-HK', 'ko-KR', 'de-DE'], isFetchAllVoice = true }: Config = {}) => {
   const { azureKey, azureRegion } = useGlobalSetting()
   const languages = ref(langs)
   const language = ref<typeof langs[number]>(langs[0])
@@ -118,16 +121,17 @@ export const useSpeechService = (langs = <const>['fr-FR', 'ja-JP', 'en-US', 'zh-
     })
   }
 
-  const ssmlToSpeak = async (text: string) => {
+  const ssmlToSpeak = async (text: string, { voice, voiceRate, lang }: { voice?: string; voiceRate?: number; lang?: string } = {}) => {
     isSynthesizing.value = true
 
-    const lang = speechConfig.value.speechSynthesisLanguage
-    const voice = speechConfig.value.speechSynthesisVoiceName
+    const targetLang = lang || speechConfig.value.speechSynthesisLanguage
+    const targetVoice = voice || speechConfig.value.speechSynthesisVoiceName
+    const targetRate = voiceRate || rate.value
 
     const ssml = `
-    <speak version="1.0" xmlns="https://www.w3.org/2001/10/synthesis" xml:lang="${lang}">
-      <voice name="${voice}">
-        <prosody rate="${rate.value}">
+    <speak version="1.0" xmlns="https://www.w3.org/2001/10/synthesis" xml:lang="${targetLang}">
+      <voice name="${targetVoice}">
+        <prosody rate="${targetRate}">
           ${text}
         </prosody>
       </voice>
@@ -158,18 +162,20 @@ export const useSpeechService = (langs = <const>['fr-FR', 'ja-JP', 'en-US', 'zh-
   }
 
   onMounted(async () => {
-    try {
-      isFetchAllVoices.value = true
-      allVoices.value = await getVoices()
-      // fr-FR 法语 ja-JP 日语 en-US 英语 zh-CN 中文 zh-HK 粤语 ko-KR 韩语 de-DE 德语
-      for (const lang of languages.value)
-        languageMap.value[lang] = allVoices.value.filter(x => lang === x.locale)
-      console.log(languageMap)
-      isFetchAllVoices.value = false
-    }
-    catch (error) {
-      isFetchAllVoices.value = false
-      allVoices.value = []
+    if (isFetchAllVoice) {
+      try {
+        isFetchAllVoices.value = true
+        allVoices.value = await getVoices()
+        // fr-FR 法语 ja-JP 日语 en-US 英语 zh-CN 中文 zh-HK 粤语 ko-KR 韩语 de-DE 德语
+        for (const lang of languages.value)
+          languageMap.value[lang] = allVoices.value.filter(x => lang === x.locale)
+        console.log(languageMap)
+        isFetchAllVoices.value = false
+      }
+      catch (error) {
+        isFetchAllVoices.value = false
+        allVoices.value = []
+      }
     }
   })
 

diff --git a/src/pages/Home/components/Content.vue b/src/pages/Home/components/Content.vue
@@ -27,7 +27,7 @@ const {
   stopRecognizeSpeech,
   ssmlToSpeak,
   isSynthesizing,
-} = useSpeechService(store.allLanguage as any)
+} = useSpeechService({ langs: store.allLanguage as any })
 
 // states
 const message = ref('') // input message

diff --git a/src/pages/Home/components/NewChat.vue b/src/pages/Home/components/NewChat.vue
@@ -13,14 +13,15 @@ const avatarList = ref<string[]>(Object.keys(modules).map(path => path.replace('
 const currentAvatarIndex = ref(Math.random() * avatarList.value.length | 0)
 
 const store = useConversationStore()
-
+const { ssmlToSpeak, isSynthesizing } = useSpeechService({ isFetchAllVoice: false })
 const allLanguages = computed(() => [...new Set(allVoices.map(v => v.locale))].filter(l => Object.keys(supportLanguageMap).includes(l)))
 const selectLanguage = ref('')
 const filterVoices = ref<VoiceInfo[]>([])
 const selectVoiceName = ref('')
 const desc = ref('')
 const name = ref('')
 const rate = ref('1.0')
+const previewText = ref('hello wrold')
 
 const canAdd = computed(() => !!(selectLanguage.value && selectVoiceName.value && desc.value && name.value))
 
@@ -58,6 +59,9 @@ const addChat = (event: any) => {
 const changeAvatar = () => {
   currentAvatarIndex.value = avatarList.value.length - 1 === currentAvatarIndex.value ? 0 : currentAvatarIndex.value + 1
 }
+const previewSpeech = () => {
+  ssmlToSpeak(previewText.value, { voice: selectVoiceName.value, lang: selectLanguage.value, voiceRate: +rate })
+}
 </script>
 
 <script>
@@ -74,31 +78,45 @@ const changeAvatar = () => {
       <Avatar v-model:image-url="imageUrl" />
     </div>
     <div flex>
-      <label center-y justify-end mr-2 for="">姓名</label>
-      <input v-model="name" w-50 p-2 type="text">
+      <label for="">姓名</label>
+      <input v-model="name" type="text">
     </div>
     <div flex>
-      <label center-y justify-end mr-2 for="">备注</label>
-      <input v-model="desc" w-50 p-2 type="text">
+      <label for="">备注</label>
+      <input v-model="desc" type="text">
     </div>
     <div flex>
-      <label center-y justify-end mr-2 for="">语言</label>
-      <select v-model="selectLanguage" w-55 select-settings>
+      <label for="">语言</label>
+      <select v-model="selectLanguage">
         <option v-for="item in allLanguages" :key="item" :value="item">
           {{ supportLanguageMap[item] }}
         </option>
       </select>
     </div>
     <div flex>
-      <label center-y justify-end mr-2 for="">音色</label>
-      <select v-model="selectVoiceName" w-55 select-settings>
+      <label for="">音色</label>
+      <select v-model="selectVoiceName">
         <option v-for="item in filterVoices" :key="item.shortName" :value="item.shortName">
           {{ `${item.locale} / ${item.gender === 1 ? 'Female' : 'Male'} / ${item.localName}` }}
         </option>
       </select>
     </div>
+    <div relative center-y>
+      <div flex>
+        <label for="">语音预览</label>
+        <input v-model="previewText" placeholder="输入文字预览语音" type="text">
+      </div>
+      <div absolute left-77>
+        <button v-if="!isSynthesizing" :disabled="!previewText" center-y ml-2 @click="previewSpeech()">
+          <i icon-btn i-ic:baseline-volume-up />
+        </button>
+        <button v-else center-y ml-2>
+          <i icon-btn i-eos-icons:bubble-loading />
+        </button>
+      </div>
+    </div>
     <div flex>
-      <label center-y justify-end mr-2 for="">语速</label>
+      <label for="">语速</label>
       <div w-55 flex>
         <input v-model="rate" flex-1 type="range" step="0.1" min="0.1" max="2.0">
         <span w-4 ml-1>{{ Number(rate).toFixed(1) }}</span>
@@ -119,3 +137,15 @@ const changeAvatar = () => {
     </button>
   </div>
 </template>
+
+<style scoped>
+  label{
+    @apply center-y justify-end mr-2 w-20
+  }
+  input{
+    @apply w-50 p-2
+  }
+  select{
+    @apply w-55 select-settings
+  }
+</style>