It should be more straightforward than separating harmine from harmaline; on the basis of its molecular structure THH should require a higher pH to precipitate than harmaline does. There should be a thread somewhere around here where this separation has been tested out but I haven't searched for it yet.
One thing I'm not fully certain about is the practicalities of the post-reaction separation of the alkaloids from the magnesium salts. Magnesium may not redissolve in ammonia the way zinc salts do.
An outline idea might be to crash everything with sodium hydroxide, which should yield freebase mixed harmalas plus magnesium hydroxide, followed by a quick rinse in distilled water and then the
Ethyl Acetate Approach for Harmala minus the need to add any lime. Hopefully the freebase alkaloids will separate nicely from magnesium hydroxide paste that way.
Another approach would be to try precipitating the magnesium as the practically insoluble magnesium ammonium phosphate.
What was your plan with regard to this issue, Nachooo? Perhaps if you were to control the pH exceedingly carefully you might see a precipitation of harmine first, followed by magnesium hydroxide and then the THH - but it might turn out rather messier than that (in all likelihood!)