support G711 multiple channels and custom sample rates (#497)

client_play_test.go

@@ -3384,8 +3384,13 @@ func TestClientPlayBackChannel(t *testing.T) {
 			Body: mediasToSDP([]*description.Media{
 				testH264Media,
 				{
-					Type:          description.MediaTypeAudio,
+					Type: description.MediaTypeAudio,
-					Formats:       []format.Format{&format.G711{}},
+					Formats: []format.Format{&format.G711{
+						PayloadTyp:   8,
+						MULaw:        false,
+						SampleRate:   8000,
+						ChannelCount: 1,
+					}},
 					IsBackChannel: true,
 				},
 			}),

examples/client-record-format-g711/main.go

@@ -38,8 +38,13 @@ func main() {
 	// create a description that contains a G711 format
 	desc := &description.Session{
 		Medias: []*description.Media{{
-			Type:    description.MediaTypeVideo,
+			Type: description.MediaTypeVideo,
-			Formats: []format.Format{&format.G711{}},
+			Formats: []format.Format{&format.G711{
+				PayloadTyp:   8,
+				MULaw:        false,
+				SampleRate:   8000,
+				ChannelCount: 1,
+			}},
 		}},
 	}
 

pkg/description/session_test.go

@@ -72,7 +72,10 @@ var casesSession = []struct {
 					Type:    MediaTypeAudio,
 					Control: "rtsp://10.0.100.50/profile5/media.smp/trackID=a",
 					Formats: []format.Format{&format.G711{
-						MULaw: true,
+						PayloadTyp:   0,
+						MULaw:        true,
+						SampleRate:   8000,
+						ChannelCount: 1,
 					}},
 				},
 				{
@@ -140,7 +143,10 @@ var casesSession = []struct {
 					Type:    MediaTypeAudio,
 					Control: "trackID=2",
 					Formats: []format.Format{&format.G711{
-						MULaw: true,
+						PayloadTyp:   0,
+						MULaw:        true,
+						SampleRate:   8000,
+						ChannelCount: 1,
 					}},
 				},
 				{
@@ -324,10 +330,16 @@ var casesSession = []struct {
 							ClockRat:   8000,
 						},
 						&format.G711{
-							MULaw: true,
+							PayloadTyp:   0,
+							MULaw:        true,
+							SampleRate:   8000,
+							ChannelCount: 1,
 						},
 						&format.G711{
-							MULaw: false,
+							PayloadTyp:   8,
+							MULaw:        false,
+							SampleRate:   8000,
+							ChannelCount: 1,
 						},
 						&format.Generic{
 							PayloadTyp: 106,
@@ -516,13 +528,23 @@ var casesSession = []struct {
 				{
 					Type:    MediaTypeAudio,
 					Control: "rtsp://192.168.0.1/audio",
-					Formats: []format.Format{&format.G711{MULaw: true}},
+					Formats: []format.Format{&format.G711{
+						PayloadTyp:   0,
+						MULaw:        true,
+						SampleRate:   8000,
+						ChannelCount: 1,
+					}},
 				},
 				{
 					Type:          MediaTypeAudio,
 					IsBackChannel: true,
 					Control:       "rtsp://192.168.0.1/audioback",
-					Formats:       []format.Format{&format.G711{MULaw: true}},
+					Formats: []format.Format{&format.G711{
+						PayloadTyp:   0,
+						MULaw:        true,
+						SampleRate:   8000,
+						ChannelCount: 1,
+					}},
 				},
 			},
 		},
@@ -671,9 +693,14 @@ var casesSession = []struct {
 			},
 			Medias: []*Media{
 				{
-					ID:      "1",
+					ID:   "1",
-					Type:    MediaTypeAudio,
+					Type: MediaTypeAudio,
-					Formats: []format.Format{&format.G711{MULaw: true}},
+					Formats: []format.Format{&format.G711{
+						PayloadTyp:   0,
+						MULaw:        true,
+						SampleRate:   8000,
+						ChannelCount: 1,
+					}},
 				},
 				{
 					ID:   "2",

pkg/format/format.go

@@ -137,6 +137,9 @@ func Unmarshal(mediaType string, payloadType uint8, rtpMap string, fmtp map[stri
 				codec == "aal2-g726-40") && clock == "8000":
 				return &G726{}
 
+			case codec == "pcma", codec == "pcmu":
+				return &G711{}
+
 			case codec == "l8", codec == "l16", codec == "l24":
 				return &LPCM{}
 			}

pkg/format/format_test.go

@@ -27,27 +27,65 @@ var casesFormat = []struct {
 	encFmtp     map[string]string
 }{
 	{
-		"audio g711 pcma",
+		"audio g711 pcma static payload type",
 		"audio",
 		8,
 		"",
 		nil,
-		&G711{},
+		&G711{
+			PayloadTyp:   8,
+			MULaw:        false,
+			SampleRate:   8000,
+			ChannelCount: 1,
+		},
 		"PCMA/8000",
 		nil,
 	},
 	{
-		"audio g711 pcmu",
+		"audio g711 pcmu static payload type",
 		"audio",
 		0,
 		"",
 		nil,
 		&G711{
-			MULaw: true,
+			PayloadTyp:   0,
+			MULaw:        true,
+			SampleRate:   8000,
+			ChannelCount: 1,
 		},
 		"PCMU/8000",
 		nil,
 	},
+	{
+		"audio g711 pcma dynamic payload type",
+		"audio",
+		96,
+		"PCMA/16000/2",
+		nil,
+		&G711{
+			PayloadTyp:   96,
+			MULaw:        false,
+			SampleRate:   16000,
+			ChannelCount: 2,
+		},
+		"PCMA/16000/2",
+		nil,
+	},
+	{
+		"audio g711 pcmu dynamic payload type",
+		"audio",
+		96,
+		"PCMU/16000/2",
+		nil,
+		&G711{
+			PayloadTyp:   96,
+			MULaw:        true,
+			SampleRate:   16000,
+			ChannelCount: 2,
+		},
+		"PCMU/16000/2",
+		nil,
+	},
 	{
 		"audio g722",
 		"audio",
@@ -125,7 +163,7 @@ var casesFormat = []struct {
 		nil,
 	},
 	{
-		"audio lpcm 8",
+		"audio lpcm 8 dynamic payload type",
 		"audio",
 		97,
 		"L8/48000/2",
@@ -140,7 +178,7 @@ var casesFormat = []struct {
 		nil,
 	},
 	{
-		"audio lpcm 16",
+		"audio lpcm 16 dynamic payload type",
 		"audio",
 		97,
 		"L16/96000/2",
@@ -155,7 +193,7 @@ var casesFormat = []struct {
 		nil,
 	},
 	{
-		"audio lpcm 16 rfc3551 stereo",
+		"audio lpcm 16 static payload type",
 		"audio",
 		10,
 		"",
@@ -170,7 +208,7 @@ var casesFormat = []struct {
 		nil,
 	},
 	{
-		"audio lpcm 16 rfc3551 mono",
+		"audio lpcm 16 static payload type",
 		"audio",
 		11,
 		"",

pkg/format/g711.go

@@ -1,20 +1,60 @@
 package format
 
 import (
+	"strconv"
+	"strings"
+
 	"github.com/pion/rtp"
 
-	"github.com/bluenviron/gortsplib/v4/pkg/format/rtpsimpleaudio"
+	"github.com/bluenviron/gortsplib/v4/pkg/format/rtplpcm"
 )
 
 // G711 is the RTP format for the G711 codec, encoded with mu-law or A-law.
 // Specification: https://datatracker.ietf.org/doc/html/rfc3551
 type G711 struct {
-	// whether to use mu-law. Otherwise, A-law is used.
+	PayloadTyp   uint8
-	MULaw bool
+	MULaw        bool
+	SampleRate   int
+	ChannelCount int
 }
 
 func (f *G711) unmarshal(ctx *unmarshalContext) error {
-	f.MULaw = (ctx.payloadType == 0)
+	f.PayloadTyp = ctx.payloadType
+
+	if ctx.payloadType == 0 {
+		f.MULaw = true
+		f.SampleRate = 8000
+		f.ChannelCount = 1
+		return nil
+	}
+
+	if ctx.payloadType == 8 {
+		f.MULaw = false
+		f.SampleRate = 8000
+		f.ChannelCount = 1
+		return nil
+	}
+
+	f.MULaw = (ctx.codec == "pcmu")
+
+	tmp := strings.SplitN(ctx.clock, "/", 2)
+
+	tmp1, err := strconv.ParseUint(tmp[0], 10, 31)
+	if err != nil {
+		return err
+	}
+	f.SampleRate = int(tmp1)
+
+	if len(tmp) >= 2 {
+		tmp1, err := strconv.ParseUint(tmp[1], 10, 31)
+		if err != nil {
+			return err
+		}
+		f.ChannelCount = int(tmp1)
+	} else {
+		f.ChannelCount = 1
+	}
+
 	return nil
 }
 
@@ -30,18 +70,26 @@ func (f *G711) ClockRate() int {
 
 // PayloadType implements Format.
 func (f *G711) PayloadType() uint8 {
-	if f.MULaw {
+	return f.PayloadTyp
-		return 0
-	}
-	return 8
 }
 
 // RTPMap implements Format.
 func (f *G711) RTPMap() string {
+	ret := ""
+
 	if f.MULaw {
-		return "PCMU/8000"
+		ret += "PCMU"
+	} else {
+		ret += "PCMA"
 	}
-	return "PCMA/8000"
+
+	ret += "/" + strconv.FormatInt(int64(f.SampleRate), 10)
+
+	if f.ChannelCount != 1 {
+		ret += "/" + strconv.FormatInt(int64(f.ChannelCount), 10)
+	}
+
+	return ret
 }
 
 // FMTP implements Format.
@@ -55,8 +103,11 @@ func (f *G711) PTSEqualsDTS(*rtp.Packet) bool {
 }
 
 // CreateDecoder creates a decoder able to decode the content of the format.
-func (f *G711) CreateDecoder() (*rtpsimpleaudio.Decoder, error) {
+func (f *G711) CreateDecoder() (*rtplpcm.Decoder, error) {
-	d := &rtpsimpleaudio.Decoder{}
+	d := &rtplpcm.Decoder{
+		BitDepth:     8,
+		ChannelCount: f.ChannelCount,
+	}
 
 	err := d.Init()
 	if err != nil {
@@ -67,9 +118,11 @@ func (f *G711) CreateDecoder() (*rtpsimpleaudio.Decoder, error) {
 }
 
 // CreateEncoder creates an encoder able to encode the content of the format.
-func (f *G711) CreateEncoder() (*rtpsimpleaudio.Encoder, error) {
+func (f *G711) CreateEncoder() (*rtplpcm.Encoder, error) {
-	e := &rtpsimpleaudio.Encoder{
+	e := &rtplpcm.Encoder{
-		PayloadType: f.PayloadType(),
+		PayloadType:  f.PayloadType(),
+		BitDepth:     8,
+		ChannelCount: f.ChannelCount,
 	}
 
 	err := e.Init()

pkg/format/g711_test.go

@@ -8,32 +8,49 @@ import (
 )
 
 func TestG711Attributes(t *testing.T) {
-	format := &G711{}
+	t.Run("pcma", func(t *testing.T) {
-	require.Equal(t, "G711", format.Codec())
+		format := &G711{
-	require.Equal(t, 8000, format.ClockRate())
+			PayloadTyp:   8,
-	require.Equal(t, true, format.PTSEqualsDTS(&rtp.Packet{}))
+			MULaw:        false,
+			SampleRate:   8000,
+			ChannelCount: 1,
+		}
+		require.Equal(t, "G711", format.Codec())
+		require.Equal(t, 8000, format.ClockRate())
+		require.Equal(t, true, format.PTSEqualsDTS(&rtp.Packet{}))
+	})
 
-	format = &G711{
+	t.Run("pcmu", func(t *testing.T) {
-		MULaw: true,
+		format := &G711{
-	}
+			PayloadTyp:   0,
-	require.Equal(t, "G711", format.Codec())
+			MULaw:        true,
-	require.Equal(t, 8000, format.ClockRate())
+			SampleRate:   8000,
+			ChannelCount: 1,
+		}
+		require.Equal(t, "G711", format.Codec())
+		require.Equal(t, 8000, format.ClockRate())
+	})
 }
 
 func TestG711DecEncoder(t *testing.T) {
-	format := &G711{}
+	format := &G711{
+		PayloadTyp:   8,
+		MULaw:        false,
+		SampleRate:   8000,
+		ChannelCount: 1,
+	}
 
 	enc, err := format.CreateEncoder()
 	require.NoError(t, err)
 
-	pkt, err := enc.Encode([]byte{0x01, 0x02, 0x03, 0x04})
+	pkts, err := enc.Encode([]byte{0x01, 0x02, 0x03, 0x04})
 	require.NoError(t, err)
-	require.Equal(t, format.PayloadType(), pkt.PayloadType)
+	require.Equal(t, format.PayloadType(), pkts[0].PayloadType)
 
 	dec, err := format.CreateDecoder()
 	require.NoError(t, err)
 
-	byts, err := dec.Decode(pkt)
+	byts, err := dec.Decode(pkts[0])
 	require.NoError(t, err)
 	require.Equal(t, []byte{0x01, 0x02, 0x03, 0x04}, byts)
 }