Now, as to the question, the dyes adsorb at their preferred ratio if you add them to the emulsion. Since you stir, the dye is distributed and used up at the proper amount for each grain size. ...
And, I would add that your method would work if there was an excess of dye, but you would also change the vAg of the emulsion thereby changing its surface adsorption characteristics. It would therefor cause some sort of secondary problem IMHO.
I was wondering about this myself, and thought that post-dye washing might lead to a whole new set of variables. There IS a method to the madness