Questions about temporal scalability

[Kz_Liao]

Hi,

I've met some problem in utilizing the temporal scalability. The 1st question was also asked in https://software.intel.com/en-us/forums/topic/515572.

1) It's written in https://software.intel.com/en-us/articles/video-conferencing-features-of-intel-media-software-development-kit/ that "Two consecutive temporal layers must have integer ratio of frame rates". Since 3 is an integer, I suppose the ratio can be 3. So I configured the initializing parameter in the video conferencing conference sample like this:

else if ( 0== msdk_strcmp(strInput, MSDK_STRING("-ts")))
        {
            CHECK_OPTION_ARGS(1);
            params.nTemporalScalabilityBase = 3; // ratio of 3
            lexical_cast(strInput[++i], params.nTemporalScalabilityLayers);
            //...
        }

By debugging, the m_temporalScale looked like the follow:

m_temporalScale.Layer[0].Scale == 1;
m_temporalScale.Layer[1].Scale == 3;
m_temporalScale.Layer[2].Scale == 9;
m_temporalScale.Layer[3].Scale == 0;

But the m_encoder->Init(&m_mfxEncParams) returns MFX_ERR_INVALID_VIDEO_PARAM. Is the ratio other than 2 not supported?

2) If the ratio other than 2 is not supported, can I use the mfxExtAVCRefListCtrl at encoding time to simulate the fame dependencies hierarchies like the temporal scalability do?