support publishing VP9 tracks with RTMP (#2247)

README.md

@@ -26,7 +26,7 @@ Live streams can be published to the server with:
 |[WebRTC servers](#webrtc-servers)|WHEP|AV1, VP9, VP8, H264|Opus, G722, G711|
 |[RTSP clients](#rtsp-clients)|UDP, TCP, RTSPS|AV1, VP9, VP8, H265, H264, MPEG-4 Video (H263, Xvid), MPEG-1/2 Video, M-JPEG and any RTP-compatible codec|Opus, MPEG-4 Audio (AAC), MPEG-1/2 Audio (MP3), G726, G722, G711, LPCM and any RTP-compatible codec|
 |[RTSP cameras and servers](#rtsp-cameras-and-servers)|UDP, UDP-Multicast, TCP, RTSPS|AV1, VP9, VP8, H265, H264, MPEG-4 Video (H263, Xvid), MPEG-1/2 Video, M-JPEG and any RTP-compatible codec|Opus, MPEG-4 Audio (AAC), MPEG-1/2 Audio (MP3), G726, G722, G711, LPCM and any RTP-compatible codec|
-|[RTMP clients](#rtmp-clients)|RTMP, RTMPS, Enhanced RTMP|AV1, H265, H264|MPEG-4 Audio (AAC), MPEG-1/2 Audio (MP3)|
+|[RTMP clients](#rtmp-clients)|RTMP, RTMPS, Enhanced RTMP|AV1, VP9, H265, H264|MPEG-4 Audio (AAC), MPEG-1/2 Audio (MP3)|
 |[RTMP cameras and servers](#rtmp-cameras-and-servers)|RTMP, RTMPS, Enhanced RTMP|H264|MPEG-4 Audio (AAC), MPEG-1/2 Audio (MP3)|
 |[HLS cameras and servers](#hls-cameras-and-servers)|Low-Latency HLS, MP4-based HLS, legacy HLS|AV1, VP9, H265, H264|Opus, MPEG-4 Audio (AAC)|
 |[UDP/MPEG-TS](#udpmpeg-ts)|Unicast, broadcast, multicast|H265, H264|Opus, MPEG-4 Audio (AAC), MPEG-1/2 Audio (MP3)|
@@ -46,7 +46,6 @@ And can be read from the server with:
 
 * Publish live streams to the server
 * Read live streams from the server
-* Proxy streams from other servers or cameras, always or on-demand
 * Streams are automatically converted from a protocol to another. For instance, it's possible to publish a stream with RTSP and read it with HLS
 * Serve multiple streams at once in separate paths
 * Authenticate users; use internal or external authentication
@@ -102,7 +101,7 @@ _rtsp-simple-server_ has been rebranded as _MediaMTX_. The reason is pretty obvi
     * [RTSP](#rtsp)
     * [RTMP](#rtmp)
     * [HLS](#hls)
-* [Features](#features)
+* [Other features](#other-features)
   * [Configuration](#configuration)
   * [Authentication](#authentication)
   * [Encrypt the configuration](#encrypt-the-configuration)
@@ -977,7 +976,7 @@ To decrease the latency, you can:
     ffmpeg -i rtsp://original-stream -pix_fmt yuv420p -c:v libx264 -preset ultrafast -b:v 600k -max_muxing_queue_size 1024 -g 30 -f rtsp rtsp://localhost:$RTSP_PORT/compressed
     ```
 
-## Features
+## Other features
 
 ### Configuration
 

go.mod

@@ -4,7 +4,7 @@ go 1.20
 
 require (
 	code.cloudfoundry.org/bytefmt v0.0.0
-	github.com/abema/go-mp4 v0.12.0
+	github.com/abema/go-mp4 v0.13.0
 	github.com/alecthomas/kong v0.8.0
 	github.com/bluenviron/gohlslib v1.0.0
 	github.com/bluenviron/gortsplib/v3 v3.10.0

go.sum

@@ -1,5 +1,5 @@
-github.com/abema/go-mp4 v0.12.0 h1:XI9PPt1BpjB3wFl18oFiX6C99uesx7F/X13Z+ga8bYY=
-github.com/abema/go-mp4 v0.12.0/go.mod h1:vPl9t5ZK7K0x68jh12/+ECWBCXoWuIDtNgPtU2f04ws=
+github.com/abema/go-mp4 v0.13.0 h1:gjEZLt7g0ePpYA5sUDrI2r8X+WuI8o+USkgG5wMgmkI=
+github.com/abema/go-mp4 v0.13.0/go.mod h1:vPl9t5ZK7K0x68jh12/+ECWBCXoWuIDtNgPtU2f04ws=
 github.com/alecthomas/assert/v2 v2.1.0 h1:tbredtNcQnoSd3QBhQWI7QZ3XHOVkw1Moklp2ojoH/0=
 github.com/alecthomas/kong v0.8.0 h1:ryDCzutfIqJPnNn0omnrgHLbAggDQM2VWHikE1xqK7s=
 github.com/alecthomas/kong v0.8.0/go.mod h1:n1iCIO2xS46oE8ZfYCNDqdR0b0wZNrXAIAqro/2132U=

internal/core/rtmp_conn.go

@@ -633,6 +633,17 @@ func (c *rtmpConn) runPublish(conn *rtmp.Conn, u *url.URL) error {
 				})
 			})
 
+		case *formats.VP9:
+			r.OnDataVP9(func(pts time.Duration, frame []byte) {
+				stream.WriteUnit(videoMedia, videoFormat, &formatprocessor.UnitVP9{
+					BaseUnit: formatprocessor.BaseUnit{
+						NTP: time.Now(),
+					},
+					PTS:   pts,
+					Frame: frame,
+				})
+			})
+
 		case *formats.H265:
 			r.OnDataH265(func(pts time.Duration, au [][]byte) {
 				stream.WriteUnit(videoMedia, videoFormat, &formatprocessor.UnitH265{
@@ -654,6 +665,9 @@ func (c *rtmpConn) runPublish(conn *rtmp.Conn, u *url.URL) error {
 					AU:  au,
 				})
 			})
+
+		default:
+			return fmt.Errorf("unsupported video codec: %T", videoFormat)
 		}
 	}
 
@@ -686,6 +700,9 @@ func (c *rtmpConn) runPublish(conn *rtmp.Conn, u *url.URL) error {
 					Frames: [][]byte{frame},
 				})
 			})
+
+		default:
+			return fmt.Errorf("unsupported audio codec: %T", audioFormat)
 		}
 	}
 

internal/core/rtmp_source.go

@@ -113,11 +113,6 @@ func (s *rtmpSource) runReader(u *url.URL, nconn net.Conn) error {
 
 	videoFormat, audioFormat := mc.Tracks()
 
-	switch videoFormat.(type) {
-	case *formats.H265, *formats.AV1:
-		return fmt.Errorf("proxying H265 or AV1 tracks with RTMP is not supported")
-	}
-
 	var medias media.Medias
 	var stream *stream.Stream
 
@@ -128,7 +123,8 @@ func (s *rtmpSource) runReader(u *url.URL, nconn net.Conn) error {
 		}
 		medias = append(medias, videoMedia)
 
-		if _, ok := videoFormat.(*formats.H264); ok {
+		switch videoFormat.(type) {
+		case *formats.H264:
 			mc.OnDataH264(func(pts time.Duration, au [][]byte) {
 				stream.WriteUnit(videoMedia, videoFormat, &formatprocessor.UnitH264{
 					BaseUnit: formatprocessor.BaseUnit{
@@ -138,6 +134,9 @@ func (s *rtmpSource) runReader(u *url.URL, nconn net.Conn) error {
 					AU:  au,
 				})
 			})
+
+		default:
+			return fmt.Errorf("unsupported video codec: %T", videoFormat)
 		}
 	}
 
@@ -170,6 +169,9 @@ func (s *rtmpSource) runReader(u *url.URL, nconn net.Conn) error {
 					Frames: [][]byte{frame},
 				})
 			})
+
+		default:
+			return fmt.Errorf("unsupported audio codec: %T", audioFormat)
 		}
 	}
 

internal/rtmp/reader.go

@@ -21,6 +21,9 @@ import (
 // OnDataAV1Func is the prototype of the callback passed to OnDataAV1().
 type OnDataAV1Func func(pts time.Duration, tu [][]byte)
 
+// OnDataVP9Func is the prototype of the callback passed to OnDataVP9().
+type OnDataVP9Func func(pts time.Duration, frame []byte)
+
 // OnDataH26xFunc is the prototype of the callback passed to OnDataH26x().
 type OnDataH26xFunc func(pts time.Duration, au [][]byte)
 
@@ -252,7 +255,13 @@ func tracksFromMetadata(conn *Conn, payload []interface{}) (formats.Format, form
 					videoTrack = &formats.AV1{}
 
 				default: // VP9
-					return nil, nil, fmt.Errorf("VP9 is not supported yet")
+					var vpcc mp4.VpcC
+					_, err := mp4.Unmarshal(bytes.NewReader(tmsg.Config), uint64(len(tmsg.Config)), &vpcc, mp4.Context{})
+					if err != nil {
+						return nil, nil, fmt.Errorf("invalid VP9 configuration: %v", err)
+					}
+
+					videoTrack = &formats.VP9{}
 				}
 			}
 
@@ -449,6 +458,16 @@ func (r *Reader) OnDataAV1(cb OnDataAV1Func) {
 	}
 }
 
+// OnDataVP9 sets a callback that is called when VP9 data is received.
+func (r *Reader) OnDataVP9(cb OnDataVP9Func) {
+	r.onDataVideo = func(msg message.Message) error {
+		if msg, ok := msg.(*message.ExtendedCodedFrames); ok {
+			cb(msg.DTS, msg.Payload)
+		}
+		return nil
+	}
+}
+
 // OnDataH265 sets a callback that is called when H265 data is received.
 func (r *Reader) OnDataH265(cb OnDataH26xFunc) {
 	r.onDataVideo = func(msg message.Message) error {