Improved: Parametric filter calculations (command "Filter") are now fully using double precision. This yields a great improvement in signal-to-noise ratio for practically no performance penalty.

Fixed: The method BiQuad::gainAt was returning wrong values. As this method was only used for testing purposes, this change is not user-visible.

filters/BiQuadFilter.h

@@ -26,18 +26,26 @@
 class BiQuadFilter : public IFilter
 {
 public:
-	BiQuadFilter(BiQuad::Type type, double dbGain, double freq, double bandwidthOrQOrS, bool isBandwidth);
+	BiQuadFilter(BiQuad::Type type, double dbGain, double freq, double bandwidthOrQOrS, bool isBandwidth, bool isCornerFreq);
 	virtual ~BiQuadFilter();
 	virtual bool getInPlace() {return true;}
 	virtual std::vector<std::wstring> initialize(float sampleRate, unsigned maxFrameCount, std::vector<std::wstring> channelNames);
 	virtual void process(float** output, float** input, unsigned frameCount);
 
+	BiQuad::Type getType() const;
+	double getDbGain() const;
+	double getFreq() const;
+	double getBandwidthOrQOrS() const;
+	bool getIsBandwidth() const;
+	bool getIsCornerFreq() const;
+
 private:
 	BiQuad::Type type;
 	double dbGain;
 	double freq;
 	double bandwidthOrQOrS;
 	bool isBandwidth;
+	bool isCornerFreq;
 
 	size_t channelCount;
 	BiQuad* biquads;