WebRTC是Google公司的一款跨平台的音视频通话技术,它为我们提供了音视频通信的核心技术,包括音视频的采集、编解码、网络传输、视频显示等功能。
借助这款API,我们可以更容易地实现音视频聊天功能。
准备工作:
1、两个客户端。
2、一台socket长连接服务器,用于交换客户端两端的通信信息(ip地址,端口等),以及管理通信状态等。
3、一台STUN/TURN/ICE Server,用于获取公网IP。
通信步骤:
1、客户端告知WebRTC获取公网地址的服务器的IP,整个过程中WebRTC会不断去获取公网地址,获取公网地址后,通过socket服务器发送给对方。
2、客户端拿到对方发送的公网地址,保存起来,WebRTC会在通信的时候选择最优的公网地址。
3、呼叫方生成一个类型为offer的会话描述,并设置为本地回话描述,然后通过socket服务器发送给接听方。
4、接听方收到offer的会话描述后,把offer设置为了远程回话描述。并生成一个类型为answer的回话描述,发送给呼叫方。
5、呼叫方收到接收方的answer,把answer设置为了远程回话描述。
6、至此,双方P2P连接建立,开始音视频通信。
具体实现:
@property(nonatomic, strong) RTCPeerConnection* peerConnection;
@property(nonatomic, strong) RTCPeerConnectionFactory* peerConnectionFactory;
@property(nonatomic, strong) RTCSessionDescription* localSdp;
@property(nonatomic, strong) RTCSessionDescription* remoteSdp;
@property(nonatomic, strong) RTCMediaConstraints* sdpMediaConstraints;
@property(nonatomic, strong) RTCAudioTrack* audioTrack;
@property(nonatomic, strong) RTCMediaStream* localMediaStream;
@property(nonatomic, strong) NSObject* candidateMutex;
@property(nonatomic, strong) NSMutableArray* queuedRemoteCandidates;
@property(nonatomic, strong) NSObject* sdpMutex;
@property(nonatomic, strong) NSMutableArray* queuedRemoteSdp;
@property(nonatomic, strong) RTCConfiguration *rtcConfig;
// video
@property(nonatomic, strong) RTCEAGLVideoView* localVideoView;
@property(nonatomic, strong) RTCEAGLVideoView* remoteVideoView;
@property(nonatomic, strong) RTCVideoTrack* localVideoTrack;
@property(nonatomic, strong) RTCVideoTrack* remoteVideoTrack;
- RTCPeerConnection:连接管理类。
- RTCPeerConnectionFactory:RTC连接工厂类,负责一些全局配置,也负责RTC对象的- 实例化。
- RTCSessionDescription:会话描述,呼叫方类型为offer;接听方为answer。
- RTCMediaConstraints:媒体信息约束,可以理解为对媒体信息进行配置的类。
- RTCMediaStream: 媒体流。
- RTCAudioTrack:音频轨道,用于添加进RTCMediaStream里,才会有声音传输。
- RTCVideoTrack:视频轨道,用于添加进RTCMediaStream里,才会有视频传输。
- RTCEAGLVideoView:RTCVideoTrack渲染之后,可显示视频画面。
- RTCConfiguration:配置连接信息,配置连接时候的超时信息,ICE服务器的地址等。
接下来进行整体的初始化:
创建ConnectionFactory;
#如果想要更加安全就开启SSL。
[RTCPeerConnectionFactory initializeSSL];
self.peerConnectionFactory = [[RTCPeerConnectionFactory alloc] init];
创建媒体流,并且添加音频轨道和视频轨道
if (self.peerConnectionFactory) {
# 初始化媒体流,ARDAMSa0和ARDAMS为专业标识符。
self.localMediaStream = [self.peerConnectionFactory mediaStreamWithLabel:@"ARDAMS"];
#添加音频轨道
self.audioTrack = [self.peerConnectionFactory audioTrackWithID:@"ARDAMSa0"];
[self.localMediaStream addAudioTrack:self.audioTrack];
if (self.supportVideo) {
#添加视频轨道
self.localVideoTrack = [self createVideoTrack:self.useFrontFacingCamera];
[self.localMediaStream addVideoTrack:self.localVideoTrack];
#把本地视频轨道渲染到localVideoView
[self.localVideoTrack addRenderer:self.localVideoView];
}
}
创建媒体会话描述SDP:
#配置信息的基础单元,以键值对的方式
RTCPair* audio = [[RTCPair alloc] initWithKey:@"OfferToReceiveAudio" value:@"true"];
NSArray* mandatory;
if (self.supportVideo) {
RTCPair* video = [[RTCPair alloc] initWithKey:@"OfferToReceiveVideo" value:@"true"];
mandatory = @[ audio, video ];
} else {
mandatory = @[ audio ];
}
self.sdpMediaConstraints = [[RTCMediaConstraints alloc] initWithMandatoryConstraints:mandatory
optionalConstraints:nil];
#此处将ICEServers的ip地址和登录信息传递给RTC
NSMutableArray *iceServers = [[NSMutableArray alloc] init];
for(NSDictionary * dic in list)
{
NSString *strUsername = [dic objectForKey:@"userName"];
NSString *strPassword = [dic objectForKey:@"password"];
NSURL * strUri= [NSURL URLWithString:[dic objectForKey:@"uri"]];
RTCICEServer * iceServer = [[RTCICEServer alloc] initWithURI:strUri username:strUsername password:strPassword];
[iceServers addObject:iceServer];//jay for test
}
#config设置
self.rtcConfig = [[RTCConfiguration alloc] init];
self.rtcConfig.iceServers = iceServers;
self.rtcConfig.iceConnectionReceivingTimeout = 90000; // 90s
#没有使用DtlsSrtp加密。
RTCPair *dtlssrtp = [[RTCPair alloc] initWithKey:@"DtlsSrtpKeyAgreement" value:@"false"]; //lianwei add for test
RTCMediaConstraints *constraints = [[RTCMediaConstraints alloc] initWithMandatoryConstraints:mandatory optionalConstraints:@[dtlssrtp]];
#创建peerConnection
if (self.peerConnectionFactory) {
self.peerConnection = [self.peerConnectionFactory peerConnectionWithConfiguration:self.rtcConfig
constraints:constraints
delegate:self];
}
#把媒体流添加进peerConnection
if (self.peerConnection) {
[self.peerConnection addStream:self.localMediaStream];
}
当我们将ICEServers告知RTC,且设置好peerConnection的delegate之后,如果获取到公网地址信息(ICECandidate),则会触发- (void)peerConnection:(RTCPeerConnection*)peerConnection gotICECandidate:(RTCICECandidate*)candidate
方法:
- (void)peerConnection:(RTCPeerConnection*)peerConnection gotICECandidate:(RTCICECandidate*)candidate
{
ZUSLOG(@"gotICECandidate,peerConnection=%@,canidate=%@",peerConnection,candidate);
NSDictionary *dicCandidate = @{
@"type" : @"candidate",
@"label" : @(candidate.sdpMLineIndex),
@"id" : candidate.sdpMid,
@"candidate" : candidate.sdp
};
NSString *message = [NSString stringWithDictionary:dicCandidate];
# 将获取到ICECandidate通过socket服务器发送给其他客户端。
[self sendRTCMessage:message withOfferType:NO];
}
而当其他端收到发送过来的ICECandidate,需要存入peerConnection备用:
NSString *type = dict[@"type"];
if ([type isEqualToString:@"candidate"]) {
NSString* mid = MessageDict[@"id"];
NSNumber* sdpLineIndex = MessageDict[@"label"];
NSString* sdp = MessageDict[@"candidate"];
//added by wafer for connection information,2017.06.15
_sdpCandidate = [NSString stringWithFormat:@"%@%@\r\n",_sdpCandidate,sdp];
RTCICECandidate* candidate = [[RTCICECandidate alloc] initWithMid:mid index:sdpLineIndex.intValue sdp:sdp];
#添加到peerConnection里
if (self.peerConnection) {
[self.peerConnection addICECandidate:candidate];
ZUSLOG(@"receiveRTCMessage,addICECandidate");
}
}
到这里,通信之前的准备工作就差不多了,接下来要开始拨打电话了:
呼叫方首先生成一个offer类型的SDP(会话描述):
#判断是否是呼叫方
if (self.initiator) {
#呼叫方生成一个offer类型的SDP
[self.peerConnection createOfferWithDelegate:self constraints:self.sdpMediaConstraints];
}
这里又出现了delegate,于是我们需要在回调方法里处理事件。
这里的protocol一共有两个方法:
1、生成SDP的回调:
- (void)peerConnection:(RTCPeerConnection *)peerConnection didCreateSessionDescription:(RTCSessionDescription *)sdp error:(NSError *)error;
2、这个是设置本地或者远程sdp的时候会调用:
- (void)peerConnection:(RTCPeerConnection *)peerConnection didSetSessionDescriptionWithError:(NSError *)error;
我们先处理第一个方法:
首先我们在成功生成SDP后,
- (void)peerConnection:(RTCPeerConnection*)peerConnection didCreateSessionDescription:(RTCSessionDescription*)origSdp error:(NSError*)error
{
ZUSLOG(@"didCreateSessionDescription,peerConnection=%@,origSdp=%@,error=%@",peerConnection,origSdp,error);
if (error) { #生成失败
return;
}
#(不管是呼叫方还是接收方,处理都是一样的)
self.localSdp = [[RTCSessionDescription alloc] initWithType:origSdp.type sdp:[NSString preferVoiceCodec:origSdp.description voiceCodecType:self.voiceCodecType]];
if (self.peerConnection) {
# 生成成功就把他设为本地SDP
[self.peerConnection setLocalDescriptionWithDelegate:self sessionDescription:self.localSdp];
}
# 然后再把生成的SDP发出去
NSDictionary *dicSdp = @{@"type":self.localSdp.type, @"sdp":self.localSdp.description};
NSString *message = [NSString stringWithDictionary:dicSdp];
[self sendRTCMessage:message withOfferType:YES];
});
}
然后就是接收方的处理(呼叫的信息是通过socket服务器告知的):
接收方在收到呼叫方的offer类型的SDP后,需要设置RemoteDescription为对方的SDP:
if ([type isEqualToString:@"offer"] ) {
NSString *sdpString = MessageDict[@"sdp"];
self.remoteSdp = [[RTCSessionDescription alloc] initWithType:type sdp:[NSString preferVoiceCodec:sdpString voiceCodecType:self.voiceCodecType]];
设置RemoteDescription为对方的SDP
if (self.peerConnection) {
[self.peerConnection setRemoteDescriptionWithDelegate:self
sessionDescription:self.remoteSdp];
}
设置成功之后会触发代理方法:
- (void)peerConnection:(RTCPeerConnection *)peerConnection didSetSessionDescriptionWithError:(NSError *)error;
我们处理接收方设置完成之后的回调,接收方需要生成类型为answer的SDP返回给呼叫方,表示同意接受通信。
- (void)peerConnection:(RTCPeerConnection*)peerConnection didSetSessionDescriptionWithError:(NSError*)error
{
if (error) {
return;
}
#判断是接收方
if (!self.initiator) {
if (self.peerConnection && self.peerConnection.remoteDescription && !self.peerConnection.localDescription) {
#
[self.peerConnection createAnswerWithDelegate:self constraints:self.sdpMediaConstraints];
}
}
}
然后又触发delegate方法
- (void)peerConnection:(RTCPeerConnection *)peerConnection didCreateSessionDescription:(RTCSessionDescription *)sdp error:(NSError *)error;
还是一模一样的,把生成的answer的SDP设为接收方自己的SDP,再把这个SDP再返回给呼叫方。
最后再回到接收方:
接收方在收到回调之后,再把RemoteDescription设置好,至此,连接正式建立
if ([type isEqualToString:@"answer"]) {
NSString *sdpString = MessageDict[@"sdp"];
self.remoteSdp = [[RTCSessionDescription alloc] initWithType:type sdp:[NSString preferVoiceCodec:sdpString voiceCodecType:self.voiceCodecType]];
if (self.peerConnection) {
[self.peerConnection setRemoteDescriptionWithDelegate:self
sessionDescription:self.remoteSdp];
}
}
从代码的角度看,呼叫方和接收方的逻辑公用还有回调之间的跳来跳去是有点晕。
但是总结一下:其实就是呼叫方生成offer的SDP,接收方生成answer的SDP,双方交换,然后设置好两端的localSDP和RemoteSDP。
最后,视频流传输的过程中会触发回调方法,记得渲染到远端视频界面remoteVideoView:
- (void)peerConnection:(RTCPeerConnection*)peerConnection addedStream:(RTCMediaStream*)stream
{
dispatch_async(dispatch_get_main_queue(), ^{
ZUSLOG(@"PCO onAddStream.");
if (self.supportVideo && stream.videoTracks.count) {
self.remoteVideoTrack = stream.videoTracks[0];
[self.remoteVideoTrack addRenderer:self.remoteVideoView]; //渲染
}
});
}
完 ~