That chart, if I take into account for purrin's graph responses to typically have that thingy at 2k, and translate it into something that's more along the lines of an Zwislocki coupler response, it looks really similar to the 4.Ai. Perhaps things look a little different due to the custom shell vs. universal/silicone, but it looks really, really similar, honestly.
I also took a look at the pictures Rakan posted in that other forum, and the C4 uses the same stuff as the 4.A/Ai; it's a TWFK and DTEC --- I'm almost absolutely sure. Sure, there are many variants of both the TWFK and the DTEC, but the amplitude of that notch at 3.7k tells me that the TWFK used in the C4 is more similar than not to the 30017 generic version used in the 4.A/Ai, as well as the peak forms of the area from 5-8k.
Knowles has been working hard to improve that 3.7 notch, so no they're not 'building it in' the TWFK; I interpret it as a difficulty in aligning the impedances of the FK and the WBFK without showing the steep drop in output below 4k in the WBFK or introducing a huge phase difference between the two drivers. Its latest versions have much less of a notch and less of a strident response in the highs, but it becomes more sinusoidal in response. Anyhoo, the notch becomes much more apparent when you apply any simple 1st order RC high pass.