I was initially thrown by the waveform display, which reflects the levels of the underlying log-encoded raw data at all times. Due to the log encoding most of the data was compressed toward the middle of the dynamic range, which is typical for a log curve, but that made it difficult to see what was happening to individual objects, like faces. Over time I learned that all I had to pay attention to was highlight clipping. If I set the exposure by eye based on what I saw in the onboard LCD and then checked that the clipped highlights were the ones I expected to be clipped (there’s always something clipped in a dark forest, usually the sky), I knew I had more than enough to work with in post. And I really liked the quality of the clipped highlights: they were white but not zingy, electronic white, similar to what I see on the Canon 5D Mk. III.