Hello everyone!

I ran into an issue recently when recording a choir with KM184's in ORTF. I have to say that I never recorded a choir before.
To the largest part, everything went as desired: good location, the group was not too large (about 40), placed as a kind of semi-circle , female voices placed towards both sides; the stereo imaging was nicely balanced with good depth.

Problem:
Occasionally, female sibilants (esses) from either side blended into the other channel, smearing the stereo localization and shifting the image towards the center.
I had no idea what might have caused this until I saw the 90° off-axis response curve of a KM140 quite recently. I assume that the design of the capules of a KM140 and a KM184 is basically identical?
For that KM140, the 90° off-axis resonse showed a +3dB peak around approx. 9kHz whilst the on-axis response showed a small dip at the same frequency. The 0°and 90° sensitivities matched at this point.
(Microphones of other quality brands showed rather similar characteristics.)
This means to me that under those specific circumstances, for an "ess" with a lot of energy around those 9kHz, arriving from let's say 40°-60° off-axis, there is diminished channel separation. The time difference of ORTF will have no real effect here.

I am not absolutely sure if I am right with this? Any suggestions or comments would be appreciated.
regards
Lutz