feat: implement medium effort features - markers, web workers, and bezier automation

Implemented three major medium effort features to enhance the audio editor:

**1. Region Markers System**
- Add marker type definitions supporting point markers and regions
- Create useMarkers hook for marker state management
- Build MarkerTimeline component for visual marker display
- Create MarkerDialog component for adding/editing markers
- Add keyboard shortcuts: M (add marker), Shift+M (next), Shift+Ctrl+M (previous)
- Support marker navigation, editing, and deletion

**2. Web Worker for Computations**
- Create audio worker for offloading heavy computations
- Implement worker functions: generatePeaks, generateMinMaxPeaks, normalizePeaks, analyzeAudio, findPeak
- Build useAudioWorker hook for easy worker integration
- Integrate worker into Waveform component with peak caching
- Significantly improve UI responsiveness during waveform generation

**3. Bezier Curve Automation**
- Enhance interpolateAutomationValue to support Bezier curves
- Implement cubic Bezier interpolation with control handles
- Add createSmoothHandles for auto-smooth curve generation
- Add generateBezierCurvePoints for smooth curve rendering
- Support bezier alongside existing linear and step curves

All features are type-safe and integrate seamlessly with the existing codebase.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>

lib/workers/audio.worker.ts

@@ -0,0 +1,200 @@
+/**
+ * Web Worker for heavy audio computations
+ * Offloads waveform generation, analysis, and normalization to background thread
+ */
+
+export interface WorkerMessage {
+  id: string;
+  type: 'generatePeaks' | 'generateMinMaxPeaks' | 'normalizePeaks' | 'analyzeAudio' | 'findPeak';
+  payload: any;
+}
+
+export interface WorkerResponse {
+  id: string;
+  type: string;
+  result?: any;
+  error?: string;
+}
+
+// Message handler
+self.onmessage = (event: MessageEvent<WorkerMessage>) => {
+  const { id, type, payload } = event.data;
+
+  try {
+    let result: any;
+
+    switch (type) {
+      case 'generatePeaks':
+        result = generatePeaks(
+          payload.channelData,
+          payload.width
+        );
+        break;
+
+      case 'generateMinMaxPeaks':
+        result = generateMinMaxPeaks(
+          payload.channelData,
+          payload.width
+        );
+        break;
+
+      case 'normalizePeaks':
+        result = normalizePeaks(
+          payload.peaks,
+          payload.targetMax
+        );
+        break;
+
+      case 'analyzeAudio':
+        result = analyzeAudio(payload.channelData);
+        break;
+
+      case 'findPeak':
+        result = findPeak(payload.channelData);
+        break;
+
+      default:
+        throw new Error(`Unknown worker message type: ${type}`);
+    }
+
+    const response: WorkerResponse = { id, type, result };
+    self.postMessage(response);
+  } catch (error) {
+    const response: WorkerResponse = {
+      id,
+      type,
+      error: error instanceof Error ? error.message : String(error),
+    };
+    self.postMessage(response);
+  }
+};
+
+/**
+ * Generate waveform peaks from channel data
+ */
+function generatePeaks(channelData: Float32Array, width: number): Float32Array {
+  const peaks = new Float32Array(width);
+  const samplesPerPeak = Math.floor(channelData.length / width);
+
+  for (let i = 0; i < width; i++) {
+    const start = i * samplesPerPeak;
+    const end = Math.min(start + samplesPerPeak, channelData.length);
+
+    let max = 0;
+    for (let j = start; j < end; j++) {
+      const abs = Math.abs(channelData[j]);
+      if (abs > max) {
+        max = abs;
+      }
+    }
+
+    peaks[i] = max;
+  }
+
+  return peaks;
+}
+
+/**
+ * Generate min/max peaks for more detailed waveform visualization
+ */
+function generateMinMaxPeaks(
+  channelData: Float32Array,
+  width: number
+): { min: Float32Array; max: Float32Array } {
+  const min = new Float32Array(width);
+  const max = new Float32Array(width);
+  const samplesPerPeak = Math.floor(channelData.length / width);
+
+  for (let i = 0; i < width; i++) {
+    const start = i * samplesPerPeak;
+    const end = Math.min(start + samplesPerPeak, channelData.length);
+
+    let minVal = 1;
+    let maxVal = -1;
+
+    for (let j = start; j < end; j++) {
+      const val = channelData[j];
+      if (val < minVal) minVal = val;
+      if (val > maxVal) maxVal = val;
+    }
+
+    min[i] = minVal;
+    max[i] = maxVal;
+  }
+
+  return { min, max };
+}
+
+/**
+ * Normalize peaks to a given range
+ */
+function normalizePeaks(peaks: Float32Array, targetMax: number = 1): Float32Array {
+  const normalized = new Float32Array(peaks.length);
+  let max = 0;
+
+  // Find max value
+  for (let i = 0; i < peaks.length; i++) {
+    if (peaks[i] > max) {
+      max = peaks[i];
+    }
+  }
+
+  // Normalize
+  const scale = max > 0 ? targetMax / max : 1;
+  for (let i = 0; i < peaks.length; i++) {
+    normalized[i] = peaks[i] * scale;
+  }
+
+  return normalized;
+}
+
+/**
+ * Analyze audio data for statistics
+ */
+function analyzeAudio(channelData: Float32Array): {
+  peak: number;
+  rms: number;
+  crestFactor: number;
+  dynamicRange: number;
+} {
+  let peak = 0;
+  let sumSquares = 0;
+  let min = 1;
+  let max = -1;
+
+  for (let i = 0; i < channelData.length; i++) {
+    const val = channelData[i];
+    const abs = Math.abs(val);
+
+    if (abs > peak) peak = abs;
+    if (val < min) min = val;
+    if (val > max) max = val;
+
+    sumSquares += val * val;
+  }
+
+  const rms = Math.sqrt(sumSquares / channelData.length);
+  const crestFactor = rms > 0 ? peak / rms : 0;
+  const dynamicRange = max - min;
+
+  return {
+    peak,
+    rms,
+    crestFactor,
+    dynamicRange,
+  };
+}
+
+/**
+ * Find peak value in channel data
+ */
+function findPeak(channelData: Float32Array): number {
+  let peak = 0;
+
+  for (let i = 0; i < channelData.length; i++) {
+    const abs = Math.abs(channelData[i]);
+    if (abs > peak) peak = abs;
+  }
+
+  return peak;
+}