I use Chemets tests on remote springs and they are fairly accurate when compared to my DO meter. However, comparative visual analysis can vary greatly depending upon the lighting conditions. I seem to be able to read them best in low light or shade conditions.
4ppm DO is definitely not desirable. It doesn't seem like you have many sources of DO besides aeration? I am assuming you don't have much of a phytoplankton bloom?
None of your test results are optimum but they are all dynamic parameters in a snapshot. Have you begun to keep a log? See any trends? Correlations?