2

我正在尝试使用 WebRTC 实现音频通话。我拥有的豆荚是：

pod 'Starscream', '~> 2.0.4'
pod 'libjingle_peerconnection'

当我尝试发送收到的 SDP 作为答案时，它失败了。我在这里提供选定的代码行，因为它太长了。

  1. 连接插座

    self.socket?.connect()

  2. 随着套接字获得连接状态，初始化WebRTC

    func initalizeWebRTC() -> Void {
print("----------------initalizeWebRTC----------------")

RTCPeerConnectionFactory.initializeSSL()
self.webRtcClient  = RTCPeerConnectionFactory.init()
let stunServer = self.defaultStunServer()
let defaultConstraint = self.createDefaultConstraint()
self.peerConnection = self.webRtcClient?.peerConnection(withICEServers: [stunServer], constraints: defaultConstraint, delegate: self)

self.localVideoView.delegate = self
self.remoteVideoView.delegate = self
// webrtc initalized local rendering of video on
self.addLocalMediaStrem()

}

  3. 这是我的配置

    func defaultStunServer() -> RTCICEServer {
print("----------------defaultStunServer----------------")
let url = URL.init(string: stunServer);
let iceServer = RTCICEServer.init(uri: url, username: "", password: "")
return iceServer!
}

func createAudioVideoConstraints() -> RTCMediaConstraints{
print("----------------createAudioVideoConstraints----------------")
let audioOffer : RTCPair = RTCPair(key: "OfferToReceiveAudio", value: "true")
let videoOffer : RTCPair = RTCPair(key: "OfferToReceiveVideo", value: "false")
let dtlsSrtpKeyAgreement : RTCPair = RTCPair(key: "DtlsSrtpKeyAgreement", value: "true")

let connectConstraints : RTCMediaConstraints = RTCMediaConstraints.init(mandatoryConstraints: [audioOffer,videoOffer], optionalConstraints: [dtlsSrtpKeyAgreement])

return connectConstraints
}

func createDefaultConstraint() -> RTCMediaConstraints {
print("----------------createDefaultConstraint----------------")
  let dtlsSrtpKeyAgreement : RTCPair = RTCPair(key: "DtlsSrtpKeyAgreement", value: "true")
let connectConstraints : RTCMediaConstraints = RTCMediaConstraints.init(mandatoryConstraints: nil, optionalConstraints: [dtlsSrtpKeyAgreement])

return connectConstraints
}

  4. 使用 WebRTC 创建报价

    self.peerConnection?.createOffer(with: self, constraints: constraint)

  5. 将数据写入套接字

    let offerDict =  ["id":"joinRoomPresenter","name":fileName,"roomName":fileName,"isFrontCamera":"false","isMicroPhone":"false","isPhoneAudio":"false","isChat":"false","isOneToOneSession":"false","accessToken":"9289010e-d2d5-42e8-a95c-212f06aa9238","userId":"3459","fileName":fileName,"portalId":voiceCallID,"personName":fileName,"photoURL":"","sdpOffer":sdp.description] as [String : Any]
socket?.write(string: offerDict.json)

  6. 成功后，我收到了来自套接字的 sip 消息。

[“sdpAnswer”：v=0

o=- 544328641767753251 2 IN IP4 104.248.181.233

s=影音室2623889018315199

t=0 0

a=组：捆绑音频

a=msid 语义：WMS janus

m=音频 9 UDP/TLS/RTP/SAVPF 111

c=IN IP4 104.248.181.233

a=recvonly

a=中：音频

a=rtcp 多路复用器

a=冰-ufrag:SJ8U

a=ice-pwd:72CaTpbKHUzwWj7rX64cGn

a=冰选项：涓流

a=指纹：sha-256 13:2A:57:AA:FC:AE:2B:44:86:1A:FD:FD:77:4A:11:BD:78:60:A3:4E:D1:98 :3C:43:1C:71:82:C7:88:EE:01:CC

a=设置：活动

a=rtpmap:111 作品/48000/2

a=extmap:1 urn:ietf:params:rtp-hdrext:ssrc-audio-level

a=msid:janus janusa0

a=ssrc:1319411670 cname:janus

a=ssrc:1319411670 msid:janus janusa0

a=ssrc:1319411670 mslabel:janus

a=ssrc:1319411670 标签：janusa0

a=候选人：1 1 udp 2013266431 104.248.181.233 57307 类型主机

a=候选：2 1 udp 2013266430 10.46.0.9 56250 类型主机

a=候选人结束

, "id": joinRoomPresenterResponse, "response": 接受]

  1. 我发送相同的 sdp 作为答案

    self.peerConnection?.setRemoteDescriptionWith(self, sessionDescription: rtcSessionDesc!)

  2. 我将 iceGathering 状态更改为 2（正在收集）。但几秒钟后，iceConnection 状态变为 4（失败）

我哪里错了？

4

1 回答 1

3

对于初学者来说，libjingle 已经很老了（2016 年）。我会尝试 GoogleWebRTC pod，它会从他们的仓库中不断更新。

在套接字发送中，您有 "isMicroPhone":"false","isPhoneAudio":"false" 如果是声音，那不应该是真的吗？

我会尝试谷歌的主要示例应用程序并从那里开始。有时很难通过 webrtc 连接正确完成所有步骤。

于 2019-08-12T19:07:22.357 回答