add server mode gain

client/demo/package.json

@@ -19,13 +19,13 @@
     "author": "wataru.okada@flect.co.jp",
     "license": "ISC",
     "devDependencies": {
-        "@babel/core": "^7.21.8",
+        "@babel/core": "^7.22.0",
-        "@babel/plugin-transform-runtime": "^7.21.4",
+        "@babel/plugin-transform-runtime": "^7.22.0",
-        "@babel/preset-env": "^7.21.5",
+        "@babel/preset-env": "^7.22.0",
-        "@babel/preset-react": "^7.18.6",
+        "@babel/preset-react": "^7.22.0",
         "@babel/preset-typescript": "^7.21.5",
-        "@types/node": "^20.2.3",
+        "@types/node": "^20.2.4",
-        "@types/react": "^18.2.6",
+        "@types/react": "^18.2.7",
         "@types/react-dom": "^18.2.4",
         "autoprefixer": "^10.4.14",
         "babel-loader": "^9.1.2",
@@ -39,7 +39,7 @@
         "html-loader": "^4.2.0",
         "html-webpack-plugin": "^5.5.1",
         "npm-run-all": "^4.1.5",
-        "postcss-loader": "^7.3.0",
+        "postcss-loader": "^7.3.1",
         "postcss-nested": "^6.0.1",
         "prettier": "^2.8.8",
         "rimraf": "^5.0.1",
@@ -47,12 +47,12 @@
         "ts-loader": "^9.4.3",
         "tsconfig-paths": "^4.2.0",
         "typescript": "^5.0.4",
-        "webpack": "^5.83.1",
+        "webpack": "^5.84.1",
         "webpack-cli": "^5.1.1",
         "webpack-dev-server": "^4.15.0"
     },
     "dependencies": {
-        "@dannadori/voice-changer-client-js": "^1.0.132",
+        "@dannadori/voice-changer-client-js": "^1.0.135",
         "@fortawesome/fontawesome-svg-core": "^6.4.0",
         "@fortawesome/free-brands-svg-icons": "^6.4.0",
         "@fortawesome/free-regular-svg-icons": "^6.4.0",

client/demo/src/components/demo/components/502_GainControlRow.tsx

@@ -9,6 +9,10 @@ export const GainControlRow = (_props: GainControlRowProps) => {
 
 
     const gainControlRow = useMemo(() => {
+        if (appState.serverSetting.serverSetting.enableServerAudio == 1) {
+            return <></>
+        }
+
         return (
             <div className="body-row split-3-2-2-3 left-padding-1 guided">
                 <div className="body-item-title left-padding-1 ">Gain Control</div>
@@ -33,8 +37,48 @@ export const GainControlRow = (_props: GainControlRowProps) => {
     }, [
         appState.clientSetting.clientSetting.inputGain,
         appState.clientSetting.clientSetting.outputGain,
-        appState.clientSetting.updateClientSetting
+        appState.clientSetting.updateClientSetting,
+        appState.serverSetting.serverSetting.enableServerAudio
     ])
 
-    return gainControlRow
+
+
+    const serverGainControlRow = useMemo(() => {
+        if (appState.serverSetting.serverSetting.enableServerAudio == 0) {
+            return <></>
+        }
+
+        return (
+            <div className="body-row split-3-2-2-3 left-padding-1 guided">
+                <div className="body-item-title left-padding-1 ">Server Gain Control</div>
+                <div>
+                    <span className="body-item-input-slider-label">in</span>
+                    <input type="range" className="body-item-input-slider" min="0.1" max="10.0" step="0.1" value={appState.serverSetting.serverSetting.serverInputAudioGain} onChange={(e) => {
+                        appState.serverSetting.updateServerSettings({ ...appState.serverSetting.serverSetting, serverInputAudioGain: Number(e.target.value) })
+                    }}></input>
+                    <span className="body-item-input-slider-val">{appState.serverSetting.serverSetting.serverInputAudioGain}</span>
+                </div>
+                <div>
+                    <span className="body-item-input-slider-label">out</span>
+                    <input type="range" className="body-item-input-slider" min="0.1" max="10.0" step="0.1" value={appState.serverSetting.serverSetting.serverOutputAudioGain} onChange={(e) => {
+                        appState.serverSetting.updateServerSettings({ ...appState.serverSetting.serverSetting, serverOutputAudioGain: Number(e.target.value) })
+                    }}></input>
+                    <span className="body-item-input-slider-val">{appState.serverSetting.serverSetting.serverOutputAudioGain}</span>
+                </div>
+                <div className="body-button-container">
+                </div>
+            </div>
+        )
+    }, [
+        appState.serverSetting.serverSetting.serverInputAudioGain,
+        appState.serverSetting.serverSetting.serverOutputAudioGain,
+        appState.serverSetting.updateServerSettings,
+        appState.serverSetting.serverSetting.enableServerAudio
+    ])
+    return (
+        <>
+            {gainControlRow}
+            {serverGainControlRow}
+        </>
+    )
 }

client/lib/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@dannadori/voice-changer-client-js",
-  "version": "1.0.134",
+  "version": "1.0.135",
   "lockfileVersion": 2,
   "requires": true,
   "packages": {
     "": {
       "name": "@dannadori/voice-changer-client-js",
-      "version": "1.0.134",
+      "version": "1.0.135",
       "license": "ISC",
       "dependencies": {
         "@types/readable-stream": "^2.3.15",

client/lib/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@dannadori/voice-changer-client-js",
-  "version": "1.0.134",
+  "version": "1.0.135",
   "description": "",
   "main": "dist/index.js",
   "directories": {

client/lib/src/const.ts

@@ -107,6 +107,8 @@ export const ServerSettingKey = {
     "serverInputDeviceId": "serverInputDeviceId",
     "serverOutputDeviceId": "serverOutputDeviceId",
     "serverReadChunkSize": "serverReadChunkSize",
+    "serverInputAudioGain": "serverInputAudioGain",
+    "serverOutputAudioGain": "serverOutputAudioGain",
 
     "tran": "tran",
     "noiseScale": "noiseScale",
@@ -162,6 +164,8 @@ export type VoiceChangerServerSetting = {
     serverInputDeviceId: number
     serverOutputDeviceId: number
     serverReadChunkSize: number
+    serverInputAudioGain: number
+    serverOutputAudioGain: number
 
 
     tran: number // so-vits-svc
@@ -268,6 +272,8 @@ export const DefaultServerSetting: ServerInfo = {
     serverInputDeviceId: -1,
     serverOutputDeviceId: -1,
     serverReadChunkSize: 256,
+    serverInputAudioGain: 1.0,
+    serverOutputAudioGain: 1.0,
 
     // VC Specific
     srcId: 0,

server/voice_changer/VoiceChanger.py

@@ -60,11 +60,13 @@ class VoiceChangerSettings:
     # serverOutputAudioSampleRate: int = 48000
     serverInputAudioSampleRate: int = 44100
     serverOutputAudioSampleRate: int = 44100
-    serverInputAudioBufferSize: int = 1024 * 24
+    # serverInputAudioBufferSize: int = 1024 * 24
-    serverOutputAudioBufferSize: int = 1024 * 24
+    # serverOutputAudioBufferSize: int = 1024 * 24
     serverInputDeviceId: int = -1
     serverOutputDeviceId: int = -1
     serverReadChunkSize: int = 256
+    serverInputAudioGain: float = 1.0
+    serverOutputAudioGain: float = 1.0
     performance: list[int] = field(default_factory=lambda: [0, 0, 0, 0])
 
     # ↓mutableな物だけ列挙
@@ -77,15 +79,20 @@ class VoiceChangerSettings:
             "serverAudioStated",
             "serverInputAudioSampleRate",
             "serverOutputAudioSampleRate",
-            "serverInputAudioBufferSize",
+            # "serverInputAudioBufferSize",
-            "serverOutputAudioBufferSize",
+            # "serverOutputAudioBufferSize",
             "serverInputDeviceId",
             "serverOutputDeviceId",
             "serverReadChunkSize",
         ]
     )
     floatData: list[str] = field(
-        default_factory=lambda: ["crossFadeOffsetRate", "crossFadeEndRate"]
+        default_factory=lambda: [
+            "crossFadeOffsetRate",
+            "crossFadeEndRate",
+            "serverInputAudioGain",
+            "serverOutputAudioGain",
+        ]
     )
     strData: list[str] = field(default_factory=lambda: [])
 
@@ -105,6 +112,7 @@ class VoiceChanger:
         self, indata: np.ndarray, outdata: np.ndarray, frames, times, status
     ):
         try:
+            indata = indata * self.settings.serverInputAudioGain
             with Timer("all_inference_time") as t:
                 unpackedData = librosa.to_mono(indata.T) * 32768.0
                 out_wav, times = self.on_request(unpackedData)
@@ -113,6 +121,7 @@ class VoiceChanger:
                     np.repeat(out_wav, outputChunnels).reshape(-1, outputChunnels)
                     / 32768.0
                 )
+                outdata[:] = outdata * self.settings.serverOutputAudioGain
             all_inference_time = t.secs
             performance = [all_inference_time] + times
             if self.emitTo is not None: