GenericByte and byte raw data A/V

MIP
, , ,
1

Dear support.

Thanks to the documentation you provided, the GenericByte format is more clear now.

But I still have a doubt on the payload structure. Let’s me explain that.

Audio sub-format

Let’s consider the AAC/ADTS codec for the audio coming from camera. In this case, MUST the payload of a single Audio Stream Packet be a complete ADTS frame? In my idea, I’d think of splitting the audio stream from the camera in generic chunks of bytes (for example X bytes, without caring of ADTS frame structure) and thus making Audio Stream sub-format Packets adding the mandatory header. Is it possible?

Video sub-format

As for the above question, Must I consider for example a well-formed GOP to build a Video Stream Packet or may I just consider video stream as a sequence of raw bytes and make Packets unaware of what GOP is at that level?

Audio/Video format

The camera send an mpeg audio/video coded stream (H264 and AAC) synchronized. Do I need to separate audio/video in my Driver and send to the Registraton Server some Multi-Packets with audio and video stream packets insiede? Is there another way to do so avoiding to demultiplex audio and video data stream? If the splitting is the only one possibility, is there a specific API to use in Driver Framework in order to demultiplex MPEG A/V stream and obtain audio and video frames from it?

In general, what the registration server is able to care of about Audio/Video data from camera?

Thanks a lot,

BR.

2

Goodmorning,

Sorry for reminding but I’d need a technical answer on this topic.

Could you please give me some suggestion?

Thank in you in advance.

BR,

Lorenzo.

3

I am trying to consult some of my colleagues in Milestone Development and hope you will hear from us soon..

4

Hi Bo. Thanks a lot.

5

The devices normally send one frame at a time and this is what we are expecting. It is also dependent on a specific driver, the normal connections (RTSP/HTTP) provide packets with frames in them which we split into frames. There is no mechanism to be able to process a whole pack of a GOP or a bulk AudioFrame.

Demuxing can be done in a specific driver, however this you need to implement explicitly.

6

Hi Bo.

Thanks a lot for your answer.

Maybe I am insisting too much on this topic but I think it would be much simpler … and I am probably not aware of the problems that could exist in allowing it

Is there an example of source code (of a driver) that your support can make available in order to understand how to integrate an RTSP camera in milestone, how to manage the received flow and communicate Audio/Video Frame to the registration server? Are there maybe privacy or similar related problems on doing that? Unfortunately, I think the training time necessary to understand how to develop a driver for Milestone is not efficient and too long for beginners.

It would be a value added support.

Thanks a lot in advance.

BR

Lorenzo.