first you should be able to detect where the laser hits the photodiode/ detector/transistor in different frequencies to be able to know where to mask the thing properly. Of course the high signal would hit in a place much higher or lower than the lower frequency signal. After knowing this, you could mask the thing properly so that you could get the proper amount of current in relation to the amount of frequency produced by the input. Since a laser wave lenght is not a variable rather a constant, your idea of using the doppler shift via a moving mirror is a good idea of "immitating" frequency modulation.
any comments, suggestions or violent reactions, just post it here.
good luck..