Best Way to Push to Cloud

FreeRTOS SDK
1

We have tested RTSP/Websocket seems to work well .

We want to push video stream to Cloud as needed , what will be the most efficient way to do that.

Devices are behind NAT so only outbound links would work.

Is RTMP available ? or something similar .

2

Hi @boltsandbytes ,

Currently, we do have an example for cloud streaming using the latest version of WebRTC, kindly check this out. ameba-rtos-pro2/component/example/kvs_webrtc_v2_mmf at main · Ameba-AIoT/ameba-rtos-pro2 · GitHub

Streaming via RTMP is still not yet available in our SDK, but it is possible to set up an external PC to convert the RTSP stream into RTMP before pushing to cloud. Nevertheless, the latency will be much higher than direct streaming through the WebRTC example.

Thank you.

3

Thank you i was able to get it working, works well. One more request , is it possible to do it without AWS , Some areas of deployment do not have internet and run private networks.

Maybe something which we can run ourself like Pion / MetaRTC for ex .

With our own infra we can give fixed pricing and run in offline mode.

4

For local network streaming, currently we only have RTSP and websocket examples, you may consider using them for local streaming without using the Internet.

Thank you.

5

In the example provided , I see audio option also.

I am able to recive audio from amb82 side , does it support sending audio to amb82 ? so that it can be played on speaker.

6

Hi @boltsandbytes ,

Yes, you may refer to the mmf2_example_2way_audio_init example in audio examples. This example is able to stream AAC sound to AmebaPro2’s audio jack via the network and transmit the sound received by AmebaPro2 over the network simultaneously.

Thank you.

7

Hi @boltsandbytes ,

There is another alternative method of pushing video stream to cloud. This method involves setting up a separate node (server / PC) that runs ffmpeg to access RTSP packets from AMB82-mini, then convert them into RTMP before pushing to a cloud. We do have an example implemented using ffmpeg and AWS EC2 virtual server, please let us know if this solution fits your application needs. Thank you.