This input contribution reports the results of EE2-3.10. The test includes a new motion compensation (MC) filter for luma with sharper filter properties compared to the existing MC filter in ECM. In subtest a, the proposed filter is used for all bi-predicted blocks, while the existing MC filter is used for all uni-predicted blocks. In subtest b, the proposed filter is used for bi-predicted blocks with both width and height greater than 4, while the existing MC filter is used for all other blocks. The filters for chroma and the filters for RPR are unchanged. EE2-3.10a gives an average BD-Rate(PSNR-Y/U/V) impact of -0.06%/-0.01%/-0.05% for RA and -0.03%/0.11%/-0.18% for LDB. EE2-3.10b gives an average BD-Rate(PSNR-Y/U/V) impact of -0.07%/0.00%/0.00% for RA and -0.10%/-0.1%/-0.1% for LDB. In the first version of the proposal, 2 QP points are copied from the anchor for subtest EE2-3.10b and some test classes are from a heterogenous cluster (yielding somewhat unreliable encoding and decoding times).