Attaching 2 comparisons using 18-55 lens at 18mm and F5.6. These were hand held at very fast shutter speeds and similarly framed. It is not a perfectly scientific test, just a quick and dirty one. I downsized the 24MP to match the 16MP using photoshops recommended bicubic sharpener. I also had to boost the exposure +.48 in the X-A2 image to match the X-T20. After comparing, if you are using the default LR settings, the X-T20 shows a very slight improvement over the X-A2 in terms of sharpness (apparent in the larger leaf in first comparison, otherwise it really is hard to see an improvement in most of the image). Given the MP difference, I would have expected a more noticeable difference. I took the same image and ran it through Iridient X-Transformer and that shows a much more noticeable improvement in sharpness and detail. If I were to use a LR only workflow, the IQ difference is minimal and in my opinion, would not be reason for upgrade. However, if you use a better demosaicer like Iridient's, that is a different story. The X-T20 wins here. One thing I noticed about colors is that the X-A2 AWB did a better job producing natural looking greens where the X-T20 showed a slight yellow bias. However, using the same neutral point to set WB, the colors looked fairly identical. Bottom line for me anyway, is if I did upgrade from an X-A series camera to an X-T series, I would always use a better demosaicer (probably the Iridient one).