code-optimus

huailiang · May 1, 2020 · 8dc392e · 8dc392e
1 parent 7e5a56d
commit 8dc392e
Show file tree

Hide file tree

Showing 5 changed files with 11 additions and 6 deletions.
diff --git a/.gitignore b/.gitignore
@@ -19,6 +19,7 @@ Assets/FMODBuild*
 Assets/Plugins*
 Assets/StreamingAssets*
 fmod_editor.log
+**/cn*
 
 # Autogenerated VS/MD solution and project files
 *.unityproj

diff --git a/Assets/LipSync/Editor/LipSyncOfflineRecognizer.cs b/Assets/LipSync/Editor/LipSyncOfflineRecognizer.cs
@@ -41,7 +41,7 @@ public string[] RecognizeAllByAudioClip(AudioClip audioClip)
                 {
                     MathToolBox.Convolute(currentAudioSpectrum, gaussianFilter, MathToolBox.EPaddleType.Repeat, smoothedAudioSpectrum);
                     MathToolBox.FindLocalLargestPeaks(smoothedAudioSpectrum, peakValues, peakPositions);
-                    frequencyUnit = audioClip.frequency / 2 / windowSize;
+                    frequencyUnit = audioClip.frequency / windowSize;
                     for (int l = 0; l < formantArray.Length; ++l)
                     {
                         formantArray[l] = peakPositions[l] * frequencyUnit;

diff --git a/Assets/LipSync/Scripts/Core/LipSyncRecognizer.cs b/Assets/LipSync/Scripts/Core/LipSyncRecognizer.cs
@@ -11,9 +11,9 @@ public class LipSyncRecognizer
         protected const int FORMANT_COUNT = 1;
 
         protected string[] vowelsByFormantJP = { "i", "u", "e", "o", "a" };
-        protected float[] vowelFormantFloorJP = { 0.0f, 250.0f, 300.0f, 450.0f, 600.0f };
+        protected float[] vowelFormantFloorJP = { 0.0f, 500.0f, 600.0f, 900.0f, 1200.0f };
         protected string[] vowelsByFormantCN = { "i", "v", "u", "e", "o", "a" };
-        protected float[] vowelFormantFloorCN = { 0.0f, 100.0f, 250.0f, 300.0f, 450.0f, 600.0f };
+        protected float[] vowelFormantFloorCN = { 0.0f, 200.0f, 500.0f, 600.0f, 900.0f, 1200.0f };
 
         protected string[] currentVowels;
         protected float[] currentVowelFormantCeilValues;

diff --git a/Assets/LipSync/Scripts/Core/LipSyncRuntimeRecognizer.cs b/Assets/LipSync/Scripts/Core/LipSyncRuntimeRecognizer.cs
@@ -26,7 +26,7 @@ private void Recognize(ref string result, int sampleRate)
             {
                 MathToolBox.Convolute(playingAudioSpectrum, gaussianFilter, MathToolBox.EPaddleType.Repeat, smoothedAudioSpectrum);
                 MathToolBox.FindLocalLargestPeaks(smoothedAudioSpectrum, peakValues, peakPositions);
-                frequencyUnit = sampleRate / 2 / windowSize;
+                frequencyUnit = sampleRate / windowSize;
                 for (int i = 0; i < formantArray.Length; ++i)
                 {
                     formantArray[i] = peakPositions[i] * frequencyUnit;

diff --git a/Assets/LipSync/Scripts/Core/MathToolBox.cs b/Assets/LipSync/Scripts/Core/MathToolBox.cs
@@ -80,14 +80,15 @@ public static void Convolute(float[] data, float[] filter, EPaddleType paddleTyp
         /// <param name="data">Source data.</param>
         /// <param name="peakValue">Array to store peak values.</param>
         /// <param name="peakPosition">Array to store peak values' positions.</param>
-        public static void FindLocalLargestPeaks(float[] data, float[] peakValue, int[] peakPosition)
+        public static void FindLocalLargestPeaks(float[] data,  float[] peakValue, int[] peakPosition)
         {
             int peakNum = 0;
             float lastPeak = 0.0f;
             int lastPeakPosition = 0;
             bool isIncreasing = false;
             bool isPeakIncreasing = false;
 
+            string str = "";
             for (int i = 0; i < data.Length - 1; ++i)
             {
                 if (data[i] < data[i + 1])
@@ -96,8 +97,9 @@ public static void FindLocalLargestPeaks(float[] data, float[] peakValue, int[]
                 }
                 else
                 {
-                    if (isIncreasing) 
+                    if (isIncreasing)
                     {
+                        str +=  (i * 86) + "-";
                         if (lastPeak < data[i]) // Peak found. 找到峰值, 一般fft窗口有两三个峰值
                         {
                             isPeakIncreasing = true;
@@ -121,11 +123,13 @@ public static void FindLocalLargestPeaks(float[] data, float[] peakValue, int[]
 
                     isIncreasing = false;
                 }
+
                 if (peakNum >= peakValue.Length)
                 {
                     break;
                 }
             }
+            Debug.Log(str + " " + peakNum + "  " + peakPosition[0] * 86);
         }
 
         /// <summary>