This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TDA2EVM5777: [Vision-SDK] Using both H264 and JPEG encoding with IVA

Part Number: TDA2EVM5777

Hi,

We are using TDA2X platform for one of our ADAS Soluton using Vision SDK Framework with Linux on A15.

And here we are capturing and recording Video from 2 cameras,  encoding into h264 and dumping into file system using a chain somewhat like:

Capture -> Encode -> Null

Now we also need to be able to take a snapshot from both the cameras(only 1 frame from each camera) and store them as a JPEG files upon User Input.

How can we realize this same while the video recording is continuously running. We wish to use the JPEG Encoding capability of IVA-HD.

Can we make use of 2 instances of Encode Link to realize this ?

Or, is it possible to invoke JPEG encoding functionality of IVA-HD outside of VSDK Framework from A15 whenever needed ?

Best Regards,

Ranjith Bhide.

  • Hi Ranjith,

    It is possible to use encLink for both H264 encoding and Jpeg encoding. 

    What is the resolution and fps of the video being captured here?

    >>Or, is it possible to invoke JPEG encoding functionality of IVA-HD outside of VSDK Framework from A15 whenever needed ?

    Gstreamer framework can be used for this. but there are challenges are there.

    1) The SDK doesn't have support for both encLink and IPUMM togerther.

    2) if IPUMM is used for encoding, your caputred buffers are at M4 and they need to be passed to A15 and need to use a pseduo source plugin to pump the caputued buffers to jpegenc.

    Thanks

    RamPrasad

  • Hi Ramprasad,

    Thanks for your response.

    >>What is the resolution and fps of the video being captured here?

    Both Videos are 1920x1080 streams at 30 FPS.

    Is it possible to achieve our requirement with Encode Link itself for the given Resolution and frame rate?

    How can we invoke the JPEG encoding only when we require, rather than trying to continuously encode and discard by default?

    Best Regards,

    Ranjith Bhide

  • Hi Ranjit,
    I checked the visionSDK links-chain architechture. There is only one instance possible for encLink but you can use different channels for H264 and MJPEG.

    You can use SELECT + GATE links to conditionally forward the buffers to MJPEG.

    Please have a look at chains_vipSingleCamConnectorLinksA15SgxDisplay usecase on how to use these to links.

    Thanks
    RamPrasad