[BUG] H265 encode main 10 is not properly supported as it should

[dabrain34]

Describe the bug

The testing framework reports not supported when running ./tests/vvs_test_runner.py -t encode_h265_main10_profile

     ERROR: Input is not supported. GetVideoCapabilities() result: 0xc464dc25
     *** Video encode capabilities not supported by hardware/driver (result) ***
     InitDeviceCapabilties failed
     
     ERROR: InitEncoder() failed with ret(-9)
     Can't create the video encoder: unsupported feature

Setup

• OS: Linux
• GPU: nVidia
• Driver Version: 595.44.00
• Sample Application: Encoder
• Sample Application Version: 0.4.4

Expected Behavior

The test should pass

Observed Behavior

The test is passing with #171 but the result is incorrect when testing the bitstream with another player. Its failing as AV1 main10 profile test so need #222 as well. Then the bistream is correcly by the external player such as GStreamer

Steps to Reproduce

./tests/vvs_test_runner.py -t encode_h265_main10_profile

Screenshots

Related non-duplicate issues

Additional Information

[dabrain34]

@srinathkr-nv did you have any chance to look into the comments of #171 ?

[srinathkr-nv]

Yes, I reviewed those comments towards the end of last week and I've started working on more changes to address them; I should have that ready in the next couple of days.

To make sure that I understand the description in the "Observed behavior" section:

• the encode_h265_main10_profile test currently fails stating that the required capabilities are not supported
• when the changes in encode: codec profile and level fixes #171 are applied, the test passes but the generated output is seen to be incorrect because the required MSB shift was not specified when encoding the input, and this gets fixed by encoder: default msbShift for >8-bit input #222

Is the above correct?

[crmurillo]

@dabrain34: this bug can be closed. I verified the test passes now and generates a correct video.
Relevant PRs are merged: #171 and #222. Test results are also updated in #239.