add flac support

Author: neurlang

cmd/tomel/main.go

@@ -3,7 +3,8 @@ package main
 import (
 	"fmt"
 	"os"
-	"localhost/gomel/mel" // make sure this is the correct import path for your `libmel` package
+	"github.com/neurlang/gomel/mel"
+	"strings"
 )
 
 func main() {
@@ -27,12 +28,32 @@ func main() {
 	m.Window = 1024
 	m.Resolut = 8192
 
-	// Generate the mel spectrogram and save it as a PNG file
-	inputFile := filename + ".wav"
-	outputFile := filename + ".png"
-	err := m.ToMelWav(inputFile, outputFile)
-	if err != nil {
-		fmt.Printf("Error generating mel spectrogram: %v\n", err)
-		os.Exit(1)
+	if strings.HasSuffix(filename, ".flac") {
+		// Generate the mel spectrogram and save it as a PNG file
+		inputFile := filename
+		outputFile := filename + ".png"
+		err := m.ToMelFlac(inputFile, outputFile)
+		if err != nil {
+			fmt.Printf("Error generating mel spectrogram: %v\n", err)
+			os.Exit(1)
+		}
+	} else if strings.HasSuffix(filename, ".wav") {
+		// Generate the mel spectrogram and save it as a PNG file
+		inputFile := filename
+		outputFile := filename + ".png"
+		err := m.ToMelWav(inputFile, outputFile)
+		if err != nil {
+			fmt.Printf("Error generating mel spectrogram: %v\n", err)
+			os.Exit(1)
+		}
+	} else {
+		// Generate the mel spectrogram and save it as a PNG file
+		inputFile := filename + ".wav"
+		outputFile := filename + ".png"
+		err := m.ToMelWav(inputFile, outputFile)
+		if err != nil {
+			fmt.Printf("Error generating mel spectrogram: %v\n", err)
+			os.Exit(1)
+		}
 	}
 }

go.mod

@@ -1,4 +1,4 @@
-module localhost/gomel
+module github.com/neurlang/gomel
 
 go 1.13
 
@@ -8,6 +8,7 @@ require (
 )
 
 require (
+	github.com/mewkiz/flac v1.0.7
 	github.com/mjibson/go-dsp v0.0.0-20180508042940-11479a337f12 // indirect
 	github.com/pkg/errors v0.9.1 // indirect
 	github.com/r9y9/go-dsp v0.0.0-20140323125827-6ccdfeca7723 // indirect

go.sum

@@ -10,13 +10,16 @@ github.com/go-audio/wav v1.0.0/go.mod h1:3yoReyQOsiARkvPl3ERCi8JFjihzG6WhjYpZCf5
 github.com/hajimehoshi/go-mp3 v0.3.0/go.mod h1:qMJj/CSDxx6CGHiZeCgbiq2DSUkbK0UbtXShQcnfyMM=
 github.com/hajimehoshi/oto v0.6.1/go.mod h1:0QXGEkbuJRohbJaxr7ZQSxnju7hEhseiPx2hrh6raOI=
 github.com/hajimehoshi/oto v0.7.1/go.mod h1:wovJ8WWMfFKvP587mhHgot/MBr4DnNy9m6EepeVGnos=
+github.com/icza/bitio v1.0.0 h1:squ/m1SHyFeCA6+6Gyol1AxV9nmPPlJFT8c2vKdj3U8=
 github.com/icza/bitio v1.0.0/go.mod h1:0jGnlLAx8MKMr9VGnn/4YrvZiprkvBelsVIbA9Jjr9A=
 github.com/icza/mighty v0.0.0-20180919140131-cfd07d671de6/go.mod h1:xQig96I1VNBDIWGCdTt54nHt6EeI639SmHycLYL7FkA=
 github.com/jfreymuth/oggvorbis v1.0.1/go.mod h1:NqS+K+UXKje0FUYUPosyQ+XTVvjmVjps1aEZH1sumIk=
 github.com/jfreymuth/vorbis v1.0.0/go.mod h1:8zy3lUAm9K/rJJk223RKy6vjCZTWC61NA2QD06bfOE0=
 github.com/lucasb-eyer/go-colorful v1.0.2/go.mod h1:0MS4r+7BZKSJ5mw4/S5MPN+qHFF1fYclkSPilDOKW0s=
 github.com/mattn/go-runewidth v0.0.4/go.mod h1:LwmH8dsx7+W8Uxz3IHJYH5QSwggIsqBzpuz5H//U1FU=
+github.com/mewkiz/flac v1.0.7 h1:uIXEjnuXqdRaZttmSFM5v5Ukp4U6orrZsnYGGR3yow8=
 github.com/mewkiz/flac v1.0.7/go.mod h1:yU74UH277dBUpqxPouHSQIar3G1X/QIclVbFahSd1pU=
+github.com/mewkiz/pkg v0.0.0-20190919212034-518ade7978e2 h1:EyTNMdePWaoWsRSGQnXiSoQu0r6RS1eA557AwJhlzHU=
 github.com/mewkiz/pkg v0.0.0-20190919212034-518ade7978e2/go.mod h1:3E2FUC/qYUfM8+r9zAwpeHJzqRVVMIYnpzD/clwWxyA=
 github.com/mjibson/go-dsp v0.0.0-20180508042940-11479a337f12 h1:dd7vnTDfjtwCETZDrRe+GPYNLA1jBtbZeyfyE8eZCyk=
 github.com/mjibson/go-dsp v0.0.0-20180508042940-11479a337f12/go.mod h1:i/KKcxEWEO8Yyl11DYafRPKOPVYTrhxiTRigjtEEXZU=

mel/impl.go

@@ -3,8 +3,10 @@ package libmel
 import "image"
 import "image/png"
 import "os"
+import "io"
 import "image/color"
 import "github.com/faiface/beep/wav"
+import "github.com/mewkiz/flac"
 import "math"
 
 func dumpimage(name string, buf [][2]float64, mels int, reverse bool) error {
@@ -87,7 +89,35 @@ func loadwav(name string) (out []float64) {
 			out = append(out, samples[i][0])
 		}
 	}
+	return
+}
+
+func loadflac(name string) (out []float64) {
+	// Open love.flac for audio streaming without parsing metadata.
+	stream, err := flac.Open(name)
+	if err != nil {
+		println(err.Error())
+		return nil
+	}
+	defer stream.Close()
 
+	for {
+		// Parse one frame of audio samples at the time, each frame containing one
+		// subframe per audio channel.
+		frame, err := stream.ParseNext()
+		if err != nil {
+			if err == io.EOF {
+				break
+			}
+		}
+
+		for _, subframe := range frame.Subframes {
+			for _, sample := range subframe.Samples {
+				out = append(out, float64(sample)/(256*256))
+			}
+			//break
+		}
+	}
 	return
 }
 

mel/mel.go

@@ -31,13 +31,9 @@ func NewMel() *Mel {
 
 var ErrFileNotLoaded = errors.New("wavNotLoaded")
 
-// ToMel generates a mel spectrogram from an input audio file and saves it as a PNG image.
-func (m *Mel) ToMelWav(inputFile, outputFile string) error {
+// ToMel generates a mel spectrogram from a wave buffer and returns the mel buffer.
+func (m *Mel) ToMel(buf []float64) ([][2]float64, error) {
 
-	var buf = loadwav(inputFile)
-	if len(buf) == 0 {
-		return ErrFileNotLoaded
-	}
 
 	buf = pad(buf, m.Window)
 
@@ -67,6 +63,42 @@ func (m *Mel) ToMelWav(inputFile, outputFile string) error {
 	ospectrum = domel(m.Resolut/2, m.NumMels, ospectrum, m.MelFmin, m.MelFmax)
 
 	spectral_normalize(ospectrum)
+
+	return ospectrum, nil
+
+}
+
+// ToMel generates a mel spectrogram from an input FLAC audio file and saves it as a PNG image.
+func (m *Mel) ToMelFlac(inputFile, outputFile string) error {
+
+	var buf = loadflac(inputFile)
+	if len(buf) == 0 {
+		return ErrFileNotLoaded
+	}
+
+	ospectrum, err := m.ToMel(buf)
+	if err != nil {
+		return err
+	}
+
+	dumpimage(outputFile, ospectrum, m.NumMels, m.YReverse)
+
+	return nil
+}
+
+// ToMel generates a mel spectrogram from an input WAV audio file and saves it as a PNG image.
+func (m *Mel) ToMelWav(inputFile, outputFile string) error {
+
+	var buf = loadwav(inputFile)
+	if len(buf) == 0 {
+		return ErrFileNotLoaded
+	}
+
+	ospectrum, err := m.ToMel(buf)
+	if err != nil {
+		return err
+	}
+
 	dumpimage(outputFile, ospectrum, m.NumMels, m.YReverse)
 
 	return nil