add RTP/MPEG-4 Video decoder and encoder (#266)

README.md

@@ -44,8 +44,8 @@ Features:
 * Utilities
   * Parse RTSP elements
   * Encode/decode format-specific frames into/from RTP packets. The following formats are supported:
-    * Video: H264, H265, M-JPEG, VP8, VP9
+    * Video: H264, H265, M-JPEG, VP8, VP9, MPEG-4 Video (H263, DivX)
-    * Audio: G711 (PCMA, PCMU), G722, LPCM, MPEG-2 audio (MP3), MPEG-4 audio (AAC), Opus
+    * Audio: G711 (PCMA, PCMU), G722, LPCM, MPEG-2 Audio (MP3), MPEG-4 Audio (AAC), Opus
 
 ## Table of contents
 

pkg/formats/mpeg2_audio.go

@@ -6,7 +6,7 @@ import (
 	"github.com/bluenviron/gortsplib/v3/pkg/formats/rtpmpeg2audio"
 )
 
-// MPEG2Audio is a RTP format that uses a MPEG-1 or MPEG-2 audio codec.
+// MPEG2Audio is a RTP format that uses a MPEG-1 or MPEG-2 Audio codec.
 // Specification: https://datatracker.ietf.org/doc/html/rfc2250
 type MPEG2Audio struct{}
 

pkg/formats/mpeg4_audio_generic.go

@@ -15,7 +15,7 @@ import (
 // MPEG4Audio is an alias for MPEG4AudioGeneric.
 type MPEG4Audio = MPEG4AudioGeneric
 
-// MPEG4AudioGeneric is a RTP format that uses a MPEG-4 audio codec.
+// MPEG4AudioGeneric is a RTP format that uses a MPEG-4 Audio codec.
 // Specification: https://datatracker.ietf.org/doc/html/rfc3640
 type MPEG4AudioGeneric struct {
 	PayloadTyp       uint8

pkg/formats/mpeg4_audio_latm.go

@@ -9,7 +9,7 @@ import (
 	"github.com/pion/rtp"
 )
 
-// MPEG4AudioLATM is a RTP format that uses a MPEG-4 audio codec.
+// MPEG4AudioLATM is a RTP format that uses a MPEG-4 Audio codec.
 // Specification: https://datatracker.ietf.org/doc/html/rfc6416#section-7.3
 type MPEG4AudioLATM struct {
 	PayloadTyp     uint8

pkg/formats/mpeg4_video_es.go

@@ -7,12 +7,14 @@ import (
 	"strings"
 
 	"github.com/pion/rtp"
+
+	"github.com/bluenviron/gortsplib/v3/pkg/formats/rtpmpeg4video"
 )
 
 // MPEG4Video is an alias for MPEG4VideoES.
 type MPEG4Video = MPEG4VideoES
 
-// MPEG4VideoES is a RTP format that uses the video codec defined in MPEG-4 part 2.
+// MPEG4VideoES is a RTP format that uses a MPEG-4 Video codec.
 // Specification: https://datatracker.ietf.org/doc/html/rfc6416#section-7.1
 type MPEG4VideoES struct {
 	PayloadTyp     uint8
@@ -83,3 +85,19 @@ func (f *MPEG4VideoES) FMTP() map[string]string {
 func (f *MPEG4VideoES) PTSEqualsDTS(*rtp.Packet) bool {
 	return true
 }
+
+// CreateDecoder creates a decoder able to decode the content of the format.
+func (f *MPEG4VideoES) CreateDecoder() *rtpmpeg4video.Decoder {
+	d := &rtpmpeg4video.Decoder{}
+	d.Init()
+	return d
+}
+
+// CreateEncoder creates an encoder able to encode the content of the format.
+func (f *MPEG4VideoES) CreateEncoder() *rtpmpeg4video.Encoder {
+	e := &rtpmpeg4video.Encoder{
+		PayloadType: f.PayloadTyp,
+	}
+	e.Init()
+	return e
+}

pkg/formats/mpeg4_video_es_test.go

@@ -17,3 +17,19 @@ func TestMPEG4VideoESAttributes(t *testing.T) {
 	require.Equal(t, 90000, format.ClockRate())
 	require.Equal(t, true, format.PTSEqualsDTS(&rtp.Packet{}))
 }
+
+func TestMPEG4VideoESDecEncoder(t *testing.T) {
+	format := &MPEG4VideoES{
+		PayloadTyp: 96,
+	}
+
+	enc := format.CreateEncoder()
+	pkts, err := enc.Encode([]byte{0x01, 0x02, 0x03, 0x04}, 0)
+	require.NoError(t, err)
+	require.Equal(t, format.PayloadType(), pkts[0].PayloadType)
+
+	dec := format.CreateDecoder()
+	byts, _, err := dec.Decode(pkts[0])
+	require.NoError(t, err)
+	require.Equal(t, []byte{0x01, 0x02, 0x03, 0x04}, byts)
+}

pkg/formats/rtph264/decoder.go

@@ -52,7 +52,7 @@ func (d *Decoder) Init() {
 	d.timeDecoder = rtptime.NewDecoder(rtpClockRate)
 }
 
-// Decode decodes NALUs from a RTP/H264 packet.
+// Decode decodes NALUs from a RTP packet.
 func (d *Decoder) Decode(pkt *rtp.Packet) ([][]byte, time.Duration, error) {
 	if d.PacketizationMode >= 2 {
 		return nil, 0, fmt.Errorf("PacketizationMode >= 2 is not supported")
@@ -167,7 +167,7 @@ func (d *Decoder) Decode(pkt *rtp.Packet) ([][]byte, time.Duration, error) {
 	return nalus, d.timeDecoder.Decode(pkt.Timestamp), nil
 }
 
-// DecodeUntilMarker decodes NALUs from a RTP/H264 packet and puts them in a buffer.
+// DecodeUntilMarker decodes NALUs from a RTP packet and puts them in a buffer.
 // When a packet has the marker flag (meaning that all the NALUs with the same PTS have
 // been received), the buffer is returned.
 func (d *Decoder) DecodeUntilMarker(pkt *rtp.Packet) ([][]byte, time.Duration, error) {

pkg/formats/rtph264/encoder_test.go

@@ -3,7 +3,6 @@ package rtph264
 import (
 	"bytes"
 	"testing"
-	"time"
 
 	"github.com/pion/rtp"
 	"github.com/stretchr/testify/require"
@@ -28,7 +27,6 @@ func mergeBytes(vals ...[]byte) []byte {
 var cases = []struct {
 	name  string
 	nalus [][]byte
-	pts   time.Duration
 	pkts  []*rtp.Packet
 }{
 	{
@@ -39,7 +37,6 @@ var cases = []struct {
 				bytes.Repeat([]byte{0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07}, 8),
 			),
 		},
-		25 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -47,7 +44,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289528607,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -65,7 +62,6 @@ var cases = []struct {
 				bytes.Repeat([]byte{0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07}, 512),
 			),
 		},
-		55 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -73,7 +69,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -88,7 +84,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17646,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -104,7 +100,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17647,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -131,7 +127,6 @@ var cases = []struct {
 				0x00, 0x00, 0x6d, 0x40,
 			},
 		},
-		0,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -177,7 +172,6 @@ var cases = []struct {
 				bytes.Repeat([]byte{0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07}, 175),
 			),
 		},
-		0,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -227,7 +221,6 @@ var cases = []struct {
 			{0x09, 0xF0},
 			{0x09, 0xF0},
 		},
-		0,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -297,7 +290,7 @@ func TestEncode(t *testing.T) {
 			}
 			e.Init()
 
-			pkts, err := e.Encode(ca.nalus, ca.pts)
+			pkts, err := e.Encode(ca.nalus, 0)
 			require.NoError(t, err)
 			require.Equal(t, ca.pkts, pkts)
 		})

pkg/formats/rtph265/decoder.go

@@ -50,7 +50,7 @@ func (d *Decoder) Init() {
 	d.timeDecoder = rtptime.NewDecoder(rtpClockRate)
 }
 
-// Decode decodes NALUs from a RTP/H265 packet.
+// Decode decodes NALUs from a RTP packet.
 func (d *Decoder) Decode(pkt *rtp.Packet) ([][]byte, time.Duration, error) {
 	if d.MaxDONDiff != 0 {
 		return nil, 0, fmt.Errorf("MaxDONDiff != 0 is not supported (yet)")
@@ -159,7 +159,7 @@ func (d *Decoder) Decode(pkt *rtp.Packet) ([][]byte, time.Duration, error) {
 	return nalus, d.timeDecoder.Decode(pkt.Timestamp), nil
 }
 
-// DecodeUntilMarker decodes NALUs from a RTP/H265 packet and puts them in a buffer.
+// DecodeUntilMarker decodes NALUs from a RTP packet and puts them in a buffer.
 // When a packet has the marker flag (meaning that all the NALUs with the same PTS have
 // been received), the buffer is returned.
 func (d *Decoder) DecodeUntilMarker(pkt *rtp.Packet) ([][]byte, time.Duration, error) {

pkg/formats/rtph265/encoder_test.go

@@ -3,7 +3,6 @@ package rtph265
 import (
 	"bytes"
 	"testing"
-	"time"
 
 	"github.com/pion/rtp"
 	"github.com/stretchr/testify/require"
@@ -28,13 +27,11 @@ func mergeBytes(vals ...[]byte) []byte {
 var cases = []struct {
 	name  string
 	nalus [][]byte
-	pts   time.Duration
 	pkts  []*rtp.Packet
 }{
 	{
 		"single",
 		[][]byte{{0x01, 0x02, 0x03, 0x04, 0x05}},
-		25 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -42,7 +39,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289528607,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: []byte{0x01, 0x02, 0x03, 0x04, 0x05},
@@ -56,7 +53,6 @@ var cases = []struct {
 			{0x08, 0x08},
 			{0x09, 0x09},
 		},
-		0,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -79,7 +75,6 @@ var cases = []struct {
 		[][]byte{
 			bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 1024),
 		},
-		55 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -87,7 +82,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -102,7 +97,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17646,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -116,7 +111,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17647,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -148,7 +143,7 @@ func TestEncode(t *testing.T) {
 			}
 			e.Init()
 
-			pkts, err := e.Encode(ca.nalus, ca.pts)
+			pkts, err := e.Encode(ca.nalus, 0)
 			require.NoError(t, err)
 			require.Equal(t, ca.pkts, pkts)
 		})

pkg/formats/rtplpcm/encoder_test.go

@@ -3,7 +3,6 @@ package rtplpcm
 import (
 	"bytes"
 	"testing"
-	"time"
 
 	"github.com/pion/rtp"
 	"github.com/stretchr/testify/require"
@@ -12,13 +11,11 @@ import (
 var cases = []struct {
 	name    string
 	samples []byte
-	pts     time.Duration
 	pkts    []*rtp.Packet
 }{
 	{
 		"single",
 		[]byte{0x01, 0x02, 0x03, 0x04, 0x05, 0x06},
-		25 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -26,7 +23,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289527557,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: []byte{0x01, 0x02, 0x03, 0x04, 0x05, 0x06},
@@ -36,7 +33,6 @@ var cases = []struct {
 	{
 		"splitted",
 		bytes.Repeat([]byte{0x41, 0x42, 0x43}, 680),
-		25 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -44,7 +40,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289527557,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: bytes.Repeat([]byte{0x41, 0x42, 0x43}, 486),
@@ -55,7 +51,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17646,
-					Timestamp:      2289527800,
+					Timestamp:      2289526600,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: bytes.Repeat([]byte{0x41, 0x42, 0x43}, 194),
@@ -87,7 +83,7 @@ func TestEncode(t *testing.T) {
 			}
 			e.Init()
 
-			pkts, err := e.Encode(ca.samples, ca.pts)
+			pkts, err := e.Encode(ca.samples, 0)
 			require.NoError(t, err)
 			require.Equal(t, ca.pkts, pkts)
 		})

pkg/formats/rtpmjpeg/decoder.go

@@ -119,7 +119,7 @@ func (d *Decoder) Init() {
 	d.timeDecoder = rtptime.NewDecoder(rtpClockRate)
 }
 
-// Decode decodes an image from a RTP/M-JPEG packet.
+// Decode decodes an image from a RTP packet.
 func (d *Decoder) Decode(pkt *rtp.Packet) ([]byte, time.Duration, error) {
 	byts := pkt.Payload
 

pkg/formats/rtpmjpeg/decoder_test.go

@@ -14,13 +14,12 @@ func TestDecode(t *testing.T) {
 			d.Init()
 
 			for _, pkt := range ca.pkts {
-				image, pts, err := d.Decode(pkt)
+				image, _, err := d.Decode(pkt)
 				if err == ErrMorePacketsNeeded {
 					continue
 				}
 
 				require.NoError(t, err)
-				require.Equal(t, ca.pts, pts)
 				require.Equal(t, ca.image, image)
 			}
 		})

pkg/formats/rtpmjpeg/encoder_test.go

@@ -2,7 +2,6 @@ package rtpmjpeg
 
 import (
 	"testing"
-	"time"
 
 	"github.com/pion/rtp"
 	"github.com/stretchr/testify/require"
@@ -11,7 +10,6 @@ import (
 var cases = []struct {
 	name  string
 	image []byte
-	pts   time.Duration
 	pkts  []*rtp.Packet
 }{
 	{
@@ -276,7 +274,6 @@ var cases = []struct {
 			0x59, 0x54, 0xda, 0xb6, 0xb3, 0x2e, 0xb3, 0x7f,
 			0xe7, 0x7f, 0xaa, 0xff, 0xff, 0xd9,
 		},
-		0,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -527,7 +524,7 @@ func TestEncode(t *testing.T) {
 			}
 			e.Init()
 
-			pkts, err := e.Encode(ca.image, ca.pts)
+			pkts, err := e.Encode(ca.image, 0)
 			require.NoError(t, err)
 			require.Equal(t, ca.pkts, pkts)
 		})

pkg/formats/rtpmpeg2audio/decoder.go

@@ -30,7 +30,7 @@ func joinFragments(fragments [][]byte, size int) []byte {
 	return ret
 }
 
-// Decoder is a RTP/MPEG2-audio decoder.
+// Decoder is a RTP/MPEG-2 Audio decoder.
 // Specification: https://datatracker.ietf.org/doc/html/rfc2250
 type Decoder struct {
 	timeDecoder         *rtptime.Decoder
@@ -45,7 +45,7 @@ func (d *Decoder) Init() {
 	d.timeDecoder = rtptime.NewDecoder(90000)
 }
 
-// Decode decodes frames from a RTP/MPEG2-audio packet.
+// Decode decodes frames from a RTP packet.
 func (d *Decoder) Decode(pkt *rtp.Packet) ([][]byte, time.Duration, error) {
 	if len(pkt.Payload) < 5 {
 		d.fragments = d.fragments[:0] // discard pending fragmented packets

pkg/formats/rtpmpeg2audio/encoder.go

@@ -28,7 +28,7 @@ func lenAggregated(frames [][]byte, frame []byte) int {
 	return l
 }
 
-// Encoder is a RTP/MPEG2-audio encoder.
+// Encoder is a RTP/MPEG-2 Audio encoder.
 // Specification: https://datatracker.ietf.org/doc/html/rfc2250
 type Encoder struct {
 	// SSRC of packets (optional).
@@ -73,7 +73,7 @@ func (e *Encoder) Init() {
 	e.timeEncoder = rtptime.NewEncoder(90000, *e.InitialTimestamp)
 }
 
-// Encode encodes frames into RTP/MPEG2-audio packets.
+// Encode encodes frames into RTP packets.
 func (e *Encoder) Encode(frames [][]byte, pts time.Duration) ([]*rtp.Packet, error) {
 	var rets []*rtp.Packet
 	var batch [][]byte

pkg/formats/rtpmpeg2audio/rtpmpeg2audio.go

@@ -1,2 +1,2 @@
-// Package rtpmpeg2audio contains a RTP/MPEG2-audio decoder and encoder.
+// Package rtpmpeg2audio contains a RTP/MPEG-2 Audio decoder and encoder.
 package rtpmpeg2audio

pkg/formats/rtpmpeg4audio/decoder.go

@@ -24,7 +24,7 @@ func joinFragments(fragments [][]byte, size int) []byte {
 	return ret
 }
 
-// Decoder is a RTP/MPEG4-audio decoder.
+// Decoder is a RTP/MPEG-4 Audio decoder.
 // Specification: https://datatracker.ietf.org/doc/html/rfc3640
 type Decoder struct {
 	// sample rate of input packets.
@@ -51,7 +51,7 @@ func (d *Decoder) Init() {
 	d.timeDecoder = rtptime.NewDecoder(d.SampleRate)
 }
 
-// Decode decodes AUs from a RTP/MPEG4-audio packet.
+// Decode decodes AUs from a RTP packet.
 // It returns the AUs and the PTS of the first AU.
 // The PTS of subsequent AUs can be calculated by adding time.Second*mpeg4audio.SamplesPerAccessUnit/clockRate.
 func (d *Decoder) Decode(pkt *rtp.Packet) ([][]byte, time.Duration, error) {

pkg/formats/rtpmpeg4audio/encoder.go

@@ -81,7 +81,7 @@ func (e *Encoder) Init() {
 	e.timeEncoder = rtptime.NewEncoder(e.SampleRate, *e.InitialTimestamp)
 }
 
-// Encode encodes AUs into RTP/MPEG4-audio packets.
+// Encode encodes AUs into RTP packets.
 func (e *Encoder) Encode(aus [][]byte, pts time.Duration) ([]*rtp.Packet, error) {
 	var rets []*rtp.Packet
 	var batch [][]byte

pkg/formats/rtpmpeg4audio/encoder_test.go

@@ -3,7 +3,6 @@ package rtpmpeg4audio
 import (
 	"bytes"
 	"testing"
-	"time"
 
 	"github.com/pion/rtp"
 	"github.com/stretchr/testify/require"
@@ -31,7 +30,6 @@ var cases = []struct {
 	indexLength      int
 	indexDeltaLength int
 	aus              [][]byte
-	pts              time.Duration
 	pkts             []*rtp.Packet
 }{
 	{
@@ -87,7 +85,6 @@ var cases = []struct {
 				0xaf, 0x7,
 			},
 		},
-		20 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -95,7 +92,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: []byte{
@@ -158,7 +155,6 @@ var cases = []struct {
 			{0x04, 0x05, 0x06, 0x07},
 			{0x08, 0x09, 0x0A, 0x0B},
 		},
-		0,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -177,7 +173,7 @@ var cases = []struct {
 			},
 		},
 	},
-	{
+	{ //nolint:dupl
 		"fragmented",
 		13,
 		3,
@@ -185,7 +181,6 @@ var cases = []struct {
 		[][]byte{
 			bytes.Repeat([]byte{0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07}, 512),
 		},
-		0,
 		[]*rtp.Packet{ //nolint:dupl
 			{
 				Header: rtp.Header{
@@ -242,7 +237,6 @@ var cases = []struct {
 			{0x08, 0x09, 0x0A, 0x0B},
 			bytes.Repeat([]byte{0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07}, 256),
 		},
-		0,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -297,7 +291,6 @@ var cases = []struct {
 		[][]byte{
 			{0x01, 0x02, 0x03, 0x04},
 		},
-		20 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -305,7 +298,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: []byte{
@@ -323,7 +316,6 @@ var cases = []struct {
 		[][]byte{
 			{0x01, 0x02, 0x03, 0x04},
 		},
-		20 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -331,7 +323,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: []byte{
@@ -350,7 +342,6 @@ var cases = []struct {
 			{0x01, 0x02, 0x03, 0x04},
 			{0x05, 0x06, 0x07, 0x08},
 		},
-		20 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -358,7 +349,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: []byte{
@@ -377,7 +368,6 @@ var cases = []struct {
 		[][]byte{
 			bytes.Repeat([]byte{0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07}, 512),
 		},
-		20 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -385,7 +375,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -400,7 +390,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17646,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -415,7 +405,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17647,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -425,7 +415,7 @@ var cases = []struct {
 			},
 		},
 	},
-	{
+	{ //nolint:dupl
 		"fragmented, custom sized, padded",
 		13,
 		0,
@@ -433,7 +423,6 @@ var cases = []struct {
 		[][]byte{
 			bytes.Repeat([]byte{0x08, 0x09, 0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F}, 512),
 		},
-		20 * time.Millisecond,
 		[]*rtp.Packet{ //nolint:dupl
 			{
 				Header: rtp.Header{
@@ -441,7 +430,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -455,7 +444,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17646,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -469,7 +458,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17647,
-					Timestamp:      2289527317,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes(
@@ -505,7 +494,7 @@ func TestEncode(t *testing.T) {
 			}
 			e.Init()
 
-			pkts, err := e.Encode(ca.aus, ca.pts)
+			pkts, err := e.Encode(ca.aus, 0)
 			require.NoError(t, err)
 			require.Equal(t, ca.pkts, pkts)
 		})

pkg/formats/rtpmpeg4audio/rtpmpeg4audio.go

@@ -1,2 +1,2 @@
-// Package rtpmpeg4audio contains a RTP/MPEG4-audio decoder and encoder.
+// Package rtpmpeg4audio contains a RTP/MPEG-4 Audio decoder and encoder.
 package rtpmpeg4audio

pkg/formats/rtpmpeg4video/decoder.go

@@ -0,0 +1,72 @@
+package rtpmpeg4video
+
+import (
+	"errors"
+	"fmt"
+	"time"
+
+	"github.com/pion/rtp"
+
+	"github.com/bluenviron/gortsplib/v3/pkg/rtptime"
+)
+
+// ErrMorePacketsNeeded is returned when more packets are needed.
+var ErrMorePacketsNeeded = errors.New("need more packets")
+
+const (
+	maxFrameSize = 1 * 1024 * 1024
+)
+
+func joinFragments(fragments [][]byte, size int) []byte {
+	ret := make([]byte, size)
+	n := 0
+	for _, p := range fragments {
+		n += copy(ret[n:], p)
+	}
+	return ret
+}
+
+// Decoder is a RTP/MPEG-4 Video decoder.
+// Specification: https://datatracker.ietf.org/doc/html/rfc6416
+type Decoder struct {
+	timeDecoder    *rtptime.Decoder
+	fragments      [][]byte
+	fragmentedSize int
+}
+
+// Init initializes the decoder.
+func (d *Decoder) Init() {
+	d.timeDecoder = rtptime.NewDecoder(90000)
+}
+
+// Decode decodes a frame from a RTP packet.
+func (d *Decoder) Decode(pkt *rtp.Packet) ([]byte, time.Duration, error) {
+	var frame []byte
+
+	if len(d.fragments) == 0 {
+		if pkt.Marker {
+			frame = pkt.Payload
+		} else {
+			d.fragmentedSize = len(pkt.Payload)
+			d.fragments = append(d.fragments, pkt.Payload)
+			return nil, 0, ErrMorePacketsNeeded
+		}
+	} else {
+		d.fragmentedSize += len(pkt.Payload)
+		if d.fragmentedSize > maxFrameSize {
+			d.fragments = d.fragments[:0] // discard pending fragmented packets
+			return nil, 0, fmt.Errorf("frame size (%d) is too big (maximum is %d)", d.fragmentedSize, maxFrameSize)
+		}
+
+		d.fragments = append(d.fragments, pkt.Payload)
+
+		if !pkt.Marker {
+			return nil, 0, ErrMorePacketsNeeded
+		}
+
+		frame = joinFragments(d.fragments, d.fragmentedSize)
+		d.fragments = d.fragments[:0]
+	}
+
+	return frame, d.timeDecoder.Decode(pkt.Timestamp), nil
+}

pkg/formats/rtpmpeg4video/decoder_test.go

@@ -0,0 +1,62 @@
+package rtpmpeg4video
+
+import (
+	"testing"
+
+	"github.com/pion/rtp"
+	"github.com/stretchr/testify/require"
+)
+
+func TestDecode(t *testing.T) {
+	for _, ca := range cases {
+		t.Run(ca.name, func(t *testing.T) {
+			d := &Decoder{}
+			d.Init()
+
+			var frame []byte
+			var err error
+
+			for _, pkt := range ca.pkts {
+				frame, _, err = d.Decode(pkt)
+				if err == ErrMorePacketsNeeded {
+					continue
+				}
+
+				require.NoError(t, err)
+			}
+
+			require.Equal(t, ca.frame, frame)
+		})
+	}
+}
+
+func FuzzDecoder(f *testing.F) {
+	f.Fuzz(func(t *testing.T, a []byte, am bool, b []byte, bm bool) {
+		d := &Decoder{}
+		d.Init()
+
+		d.Decode(&rtp.Packet{
+			Header: rtp.Header{
+				Version:        2,
+				Marker:         am,
+				PayloadType:    96,
+				SequenceNumber: 17645,
+				Timestamp:      2289527317,
+				SSRC:           0x9dbb7812,
+			},
+			Payload: a,
+		})
+
+		d.Decode(&rtp.Packet{
+			Header: rtp.Header{
+				Version:        2,
+				Marker:         bm,
+				PayloadType:    96,
+				SequenceNumber: 17646,
+				Timestamp:      2289527317,
+				SSRC:           0x9dbb7812,
+			},
+			Payload: b,
+		})
+	})
+}

pkg/formats/rtpmpeg4video/encoder.go

@@ -0,0 +1,111 @@
+package rtpmpeg4video
+
+import (
+	"crypto/rand"
+	"time"
+
+	"github.com/pion/rtp"
+
+	"github.com/bluenviron/gortsplib/v3/pkg/rtptime"
+)
+
+const (
+	rtpVersion = 2
+)
+
+func randUint32() uint32 {
+	var b [4]byte
+	rand.Read(b[:])
+	return uint32(b[0])<<24 | uint32(b[1])<<16 | uint32(b[2])<<8 | uint32(b[3])
+}
+
+// Encoder is a RTP/MPEG-4 Video encoder.
+// Specification: https://datatracker.ietf.org/doc/html/rfc6416
+type Encoder struct {
+	// payload type of packets.
+	PayloadType uint8
+
+	// SSRC of packets (optional).
+	// It defaults to a random value.
+	SSRC *uint32
+
+	// initial sequence number of packets (optional).
+	// It defaults to a random value.
+	InitialSequenceNumber *uint16
+
+	// initial timestamp of packets (optional).
+	// It defaults to a random value.
+	InitialTimestamp *uint32
+
+	// maximum size of packet payloads (optional).
+	// It defaults to 1460.
+	PayloadMaxSize int
+
+	sequenceNumber uint16
+	timeEncoder    *rtptime.Encoder
+}
+
+// Init initializes the encoder.
+func (e *Encoder) Init() {
+	if e.SSRC == nil {
+		v := randUint32()
+		e.SSRC = &v
+	}
+	if e.InitialSequenceNumber == nil {
+		v := uint16(randUint32())
+		e.InitialSequenceNumber = &v
+	}
+	if e.InitialTimestamp == nil {
+		v := randUint32()
+		e.InitialTimestamp = &v
+	}
+	if e.PayloadMaxSize == 0 {
+		e.PayloadMaxSize = 1460 // 1500 (UDP MTU) - 20 (IP header) - 8 (UDP header) - 12 (RTP header)
+	}
+
+	e.sequenceNumber = *e.InitialSequenceNumber
+	e.timeEncoder = rtptime.NewEncoder(90000, *e.InitialTimestamp)
+}
+
+// Encode encodes a frame into RTP packets.
+func (e *Encoder) Encode(frame []byte, pts time.Duration) ([]*rtp.Packet, error) {
+	availPerPacket := e.PayloadMaxSize
+	le := len(frame)
+	packetCount := le / availPerPacket
+	lastPacketSize := le % availPerPacket
+	if lastPacketSize > 0 {
+		packetCount++
+	}
+
+	pos := 0
+	ret := make([]*rtp.Packet, packetCount)
+	encPTS := e.timeEncoder.Encode(pts)
+
+	for i := range ret {
+		var le int
+		if i != (packetCount - 1) {
+			le = availPerPacket
+		} else {
+			le = lastPacketSize
+		}
+
+		payload := make([]byte, le)
+		pos += copy(payload, frame[pos:])
+
+		ret[i] = &rtp.Packet{
+			Header: rtp.Header{
+				Version:        rtpVersion,
+				PayloadType:    e.PayloadType,
+				SequenceNumber: e.sequenceNumber,
+				Timestamp:      encPTS,
+				SSRC:           *e.SSRC,
+				Marker:         (i == len(ret)-1),
+			},
+			Payload: payload,
+		}
+
+		e.sequenceNumber++
+	}
+
+	return ret, nil
+}

pkg/formats/rtpmpeg4video/encoder_test.go

@@ -0,0 +1,101 @@
+package rtpmpeg4video
+
+import (
+	"bytes"
+	"testing"
+
+	"github.com/pion/rtp"
+	"github.com/stretchr/testify/require"
+)
+
+var cases = []struct {
+	name  string
+	frame []byte
+	pkts  []*rtp.Packet
+}{
+	{
+		"single",
+		[]byte{0x01, 0x02, 0x03, 0x04},
+		[]*rtp.Packet{
+			{
+				Header: rtp.Header{
+					Version:        2,
+					Marker:         true,
+					PayloadType:    96,
+					SequenceNumber: 17645,
+					Timestamp:      2289526357,
+					SSRC:           0x9dbb7812,
+				},
+				Payload: []byte{
+					0x01, 0x02, 0x03, 0x04,
+				},
+			},
+		},
+	},
+	{
+		"fragmented",
+		bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 150/4),
+		[]*rtp.Packet{
+			{
+				Header: rtp.Header{
+					Version:        2,
+					Marker:         false,
+					PayloadType:    96,
+					SequenceNumber: 17645,
+					Timestamp:      2289526357,
+					SSRC:           0x9dbb7812,
+				},
+				Payload: bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 100/4),
+			},
+			{
+				Header: rtp.Header{
+					Version:        2,
+					Marker:         true,
+					PayloadType:    96,
+					SequenceNumber: 17646,
+					Timestamp:      2289526357,
+					SSRC:           0x9dbb7812,
+				},
+				Payload: bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 50/4),
+			},
+		},
+	},
+}
+
+func TestEncode(t *testing.T) {
+	for _, ca := range cases {
+		t.Run(ca.name, func(t *testing.T) {
+			e := &Encoder{
+				PayloadType: 96,
+				SSRC: func() *uint32 {
+					v := uint32(0x9dbb7812)
+					return &v
+				}(),
+				InitialSequenceNumber: func() *uint16 {
+					v := uint16(0x44ed)
+					return &v
+				}(),
+				InitialTimestamp: func() *uint32 {
+					v := uint32(0x88776655)
+					return &v
+				}(),
+				PayloadMaxSize: 100,
+			}
+			e.Init()
+
+			pkts, err := e.Encode(ca.frame, 0)
+			require.NoError(t, err)
+			require.Equal(t, ca.pkts, pkts)
+		})
+	}
+}
+
+func TestEncodeRandomInitialState(t *testing.T) {
+	e := &Encoder{
+		PayloadType: 96,
+	}
+	e.Init()
+	require.NotEqual(t, nil, e.SSRC)
+	require.NotEqual(t, nil, e.InitialSequenceNumber)
+	require.NotEqual(t, nil, e.InitialTimestamp)
+}

pkg/formats/rtpmpeg4video/rtpmpeg4video.go

@@ -0,0 +1,2 @@
+// Package rtpmpeg4video contains a RTP/MPEG-4 Video decoder and encoder.
+package rtpmpeg4video

pkg/formats/rtpsimpleaudio/encoder_test.go

@@ -2,7 +2,6 @@ package rtpsimpleaudio
 
 import (
 	"testing"
-	"time"
 
 	"github.com/pion/rtp"
 	"github.com/stretchr/testify/require"
@@ -11,20 +10,18 @@ import (
 var cases = []struct {
 	name  string
 	frame []byte
-	pts   time.Duration
 	pkt   *rtp.Packet
 }{
 	{
 		"single",
 		[]byte{0x01, 0x02, 0x03, 0x04},
-		25 * time.Millisecond,
 		&rtp.Packet{
 			Header: rtp.Header{
 				Version:        2,
 				Marker:         false,
 				PayloadType:    0,
 				SequenceNumber: 17645,
-				Timestamp:      2289526557,
+				Timestamp:      2289526357,
 				SSRC:           0x9dbb7812,
 			},
 			Payload: []byte{0x01, 0x02, 0x03, 0x04},
@@ -53,7 +50,7 @@ func TestEncode(t *testing.T) {
 			}
 			e.Init()
 
-			pkt, err := e.Encode(ca.frame, ca.pts)
+			pkt, err := e.Encode(ca.frame, 0)
 			require.NoError(t, err)
 			require.Equal(t, ca.pkt, pkt)
 		})

pkg/formats/rtpvp8/decoder.go

@@ -43,7 +43,7 @@ func (d *Decoder) Init() {
 	d.timeDecoder = rtptime.NewDecoder(rtpClockRate)
 }
 
-// Decode decodes a VP8 frame from a RTP/VP8 packet.
+// Decode decodes a VP8 frame from a RTP packet.
 func (d *Decoder) Decode(pkt *rtp.Packet) ([]byte, time.Duration, error) {
 	var vpkt codecs.VP8Packet
 	_, err := vpkt.Unmarshal(pkt.Payload)

pkg/formats/rtpvp8/encoder_test.go

@@ -3,7 +3,6 @@ package rtpvp8
 import (
 	"bytes"
 	"testing"
-	"time"
 
 	"github.com/pion/rtp"
 	"github.com/stretchr/testify/require"
@@ -28,13 +27,11 @@ func mergeBytes(vals ...[]byte) []byte {
 var cases = []struct {
 	name  string
 	frame []byte
-	pts   time.Duration
 	pkts  []*rtp.Packet
 }{
 	{
 		"single",
 		[]byte{0x01, 0x02, 0x03, 0x04},
-		25 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -42,7 +39,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289528607,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: []byte{0x10, 0x01, 0x02, 0x03, 0x04},
@@ -52,7 +49,6 @@ var cases = []struct {
 	{
 		"fragmented",
 		bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 4096/4),
-		55 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -60,7 +56,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes([]byte{0x10}, bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 364), []byte{0x01, 0x02, 0x03}),
@@ -71,7 +67,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17646,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes([]byte{0x00, 0x04}, bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 364), []byte{0x01, 0x02}),
@@ -82,7 +78,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17647,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes([]byte{0x00, 0x03, 0x04}, bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 294)),
@@ -111,7 +107,7 @@ func TestEncode(t *testing.T) {
 			}
 			e.Init()
 
-			pkts, err := e.Encode(ca.frame, ca.pts)
+			pkts, err := e.Encode(ca.frame, 0)
 			require.NoError(t, err)
 			require.Equal(t, ca.pkts, pkts)
 		})

pkg/formats/rtpvp9/decoder.go

@@ -43,7 +43,7 @@ func (d *Decoder) Init() {
 	d.timeDecoder = rtptime.NewDecoder(rtpClockRate)
 }
 
-// Decode decodes a VP9 frame from a RTP/VP9 packet.
+// Decode decodes a VP9 frame from a RTP packet.
 func (d *Decoder) Decode(pkt *rtp.Packet) ([]byte, time.Duration, error) {
 	var vpkt codecs.VP9Packet
 	_, err := vpkt.Unmarshal(pkt.Payload)

pkg/formats/rtpvp9/encoder_test.go

@@ -3,7 +3,6 @@ package rtpvp9
 import (
 	"bytes"
 	"testing"
-	"time"
 
 	"github.com/pion/rtp"
 	"github.com/stretchr/testify/require"
@@ -28,13 +27,11 @@ func mergeBytes(vals ...[]byte) []byte {
 var cases = []struct {
 	name  string
 	frame []byte
-	pts   time.Duration
 	pkts  []*rtp.Packet
 }{
 	{
 		"single",
 		[]byte{0x01, 0x02, 0x03, 0x04},
-		25 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -42,7 +39,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289528607,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: []byte{0x9c, 0xb5, 0xaf, 0x01, 0x02, 0x03, 0x04},
@@ -52,7 +49,6 @@ var cases = []struct {
 	{
 		"fragmented",
 		bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 4096/4),
-		55 * time.Millisecond,
 		[]*rtp.Packet{
 			{
 				Header: rtp.Header{
@@ -60,7 +56,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17645,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes([]byte{0x98, 0xb5, 0xaf}, bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 364), []byte{0x01}),
@@ -71,7 +67,7 @@ var cases = []struct {
 					Marker:         false,
 					PayloadType:    96,
 					SequenceNumber: 17646,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes([]byte{0x90, 0xb5, 0xaf, 0x02, 0x03, 0x04},
@@ -83,7 +79,7 @@ var cases = []struct {
 					Marker:         true,
 					PayloadType:    96,
 					SequenceNumber: 17647,
-					Timestamp:      2289531307,
+					Timestamp:      2289526357,
 					SSRC:           0x9dbb7812,
 				},
 				Payload: mergeBytes([]byte{0x94, 0xb5, 0xaf, 0x03, 0x04}, bytes.Repeat([]byte{0x01, 0x02, 0x03, 0x04}, 295)),
@@ -116,7 +112,7 @@ func TestEncode(t *testing.T) {
 			}
 			e.Init()
 
-			pkts, err := e.Encode(ca.frame, ca.pts)
+			pkts, err := e.Encode(ca.frame, 0)
 			require.NoError(t, err)
 			require.Equal(t, ca.pkts, pkts)
 		})