I correctly grouped the frames into [ADE] and [BCF]. But I attributed each group to the wrong camera!
FWIW here were my observations as I examined the frames:
A: more colour-free details in the cleaver scratches; less blown knife highlights; whiter oof diffraction patterns
B: hot pixel in yellow background; chromatics on oof diffraction patterns
C: chroma noise, chromatics in oof diffraction patterns
D: speckly lumi noise, whiter oof diffraction patterns
E: more colour-free details in the sweets highlights; zero colour in white tablecloth texture; speckly lumi noise
F: more colour in oof diffraction patterns, sweets highlights & tablecloth texture
-> Hassy 50D= A, D, E ? [Wrong!!]
The interesting thing is that the Canon's optics had better chromatic correction (this shows regardless of the downsampling, sharpening etc.).
Sensor-wise, the H50D seems to have less luminance noise but more chroma noise (colour moire may explain this in the sharply focussed details, but it can't explain it in the oof regions).