Fix the memory leak caused by allocating memory for m_cYuvPredTempDMVRL0 and m_cYuvPredTempDMVRL1 multiple times at decoder.

Changed computation of deltaIntMvX and deltaIntMvY based on un-clipped
MVs instead of clipped MVs

When MC for luma only is desired, use existing luma/chroma parameters
instead of encoder-specific field in data structure

Also take care that luma only MC doesn't imply IBC

- all elements of SPSNext are moved into SPS
- syntax elements of SPS are sorted according to draft (as far as possible)
- comments were added where software and draft diverge

 - CE9.2.1g + RefinedMV only for MC and TMVP + SAD cost

M0192 modifications - affine chroma MV is average of two luma sub-block MVs instead of four; MV rounding is unified

Two changes are included: (1) disable fractional MVD for UHD sequences (M0312 & M0255) (2) encoder optimization for MMVD

Missing clipping for MV storage in affine

- sps_sao_enabled_flag
- sps_alf_enabled_flag
- pcm_enabled_flag
- pcm_sample_bit_depth_luma_minus1
- pcm_sample_bit_depth_chroma_minus1
- log2_min_pcm_luma_coding_block_size_minus3
- log2_diff_max_min_pcm_luma_coding_block_size
- pcm_loop_filter_disabled_flag
- sps_ref_wraparound_enabled_flag
- sps_ref_wraparound_offset
- sps_temporal_mvp_enabled_flag
- sps_sbtmvp_enabled_flag
- sps_bdof_enabled_flag