WebRTC: 推流，如何只推视频或者只推音频

[shinoi2]

描述(Description)

描述你遇到了什么问题(Please description your issue here) 希望Webrtc 推流时，可以只推视频或者只推音频。 SRS版本(Version): v4.0.95 SRS的配置如下(Config): rtc2rtmp.conf

重现(Replay)

重现Bug的步骤(How to replay bug?) 例如我只有麦克风，没有摄像头，可以选择只推音频。这里以只推音频为例： 1、将 srs.sdk.js 中调用摄像头相关代码注释，启动 srs。

    self.constraints = {
        audio: true,
        // video: {
        //     width: {ideal: 320, max: 576}
        // }
    };
    // self.pc.addTransceiver("video", {direction: "sendonly"});
    // self.pc.addTransceiver("video", {direction: "recvonly"});

2、此时webrtc可以实现只推音频，播放纯音频。但是 rtmp 流无法播放，第三方webrtc(jswebrtc)无法播放。 3、如果注释掉音频，只推送视频。此时webrtc可以实现只推视频，播放纯视频。但是 rtmp 流播放延迟大(约6s)，第三方webrtc(jswebrtc)无法播放，并且导致srs服务崩溃。

期望行为(Expect)

可以正常只推视频或者只推音频，并且可以正常播放。

[shinoi2]

The reason for the error is that when webrtc only pushes audio, encoding fails when encountering NACK. It should be that NACK is not handled properly.

TRANS_BY_GPT3

[gyf19]

I also want to know if the app's Flutter version has this bug. Thank you.

TRANS_BY_GPT3

[winlinvip]

Welcome Patch.

TRANS_BY_GPT3

[strangemk2]

Can you only push audio to HLS? It seems like there is no logic for segmenting and generating AAC in the code.

TRANS_BY_GPT3

[SuperChrisliu]

Has this problem been resolved now? Test it out. Webrtc only pushes audio, rtmp pulling is possible, but it takes about 7 seconds for rtmp to receive data.

TRANS_BY_GPT3