Quoting ( tanpura thread )
"As the sung intervals move away from the tonic in harmonic space,
>accurate perception of intonation becomes more problematic,
>and often more dependent on melodic context. Tivra Ma, for example,
>is both melodically and harmonically distant from the tonic (some
>people construe it as a 7/5 interval, but the usual procedure would
>be to intone a clear, non-beating 5/4 (major third) above the 9/8
>(major second) above the tonic. That is, "the shuddh Ga of shuddh
>Re." Because of this complexity, the tivra Ma relationship is
>both harder to intone accurately and harder to diagnose when
>mistuned.
>
>Komal Re is melodically quite close to Sa, but very
>distant in harmonic space. The most effective intonational path
>in my experience is to "mentally hear" the natural fourth, and then
>intone the Re a pure 5/4 major third DOWN from it, that is, such that
>the 5th partial of the Re equals the 4th partial of the Ma. Another
>method, useful when the Re is to be rendered with andolan, is
>to imagine the note undulating freely in the exact center of the Pa-Pa
>octave space. Even though komal Re will not be equidistant from
>the two Panchams, the analogy is useful; when I tell my students
>to "think of the two Pa-s as trees, and the komal Re as reclining in
>a hammock strung between them," their vocal intonation improves
>perceptibly. "

I think the kom re comes naturally when one has the habit of hearing the vedic triswari ( `n S r ) that is used in Shlokas , stotrams ...
It evokes Bairagi , Basant Mukhari and similar raags .