The time dilation of Alice in Bob's frame is
$$ \rm \frac{d t}{d \tau} = 1 \div \sqrt{1-r_s/r} \ \div \sqrt{1-v^2/c^2}$$
while the time dilation of Bob in Alice's frame is
$$ \rm \frac{d \tau}{d t} = \sqrt{1-r_s/r} \ \div \sqrt{1-v^2/c^2}$$
where t is the time elapsed on Bob's clock, τ on Alice's and v the local velocity of Alice relative to a stationary probe. Bob is far away from the black hole in this scenario. The gravitative component of the time dilation is absolute, while the kinematic component is relative, see here.
Here we are assuming Schwarzschild geometry, in the vicinity of rotating or charged black holes the split of the components is little bit more complicated but following the same principle.
So if Alice would be free falling from infinity (then v would be the negative escape velocity) both components would cancel in her frame, and Bob would age with the same rate as herself. In Bob's frame Alice then would age with the squared rate of locally stationary observers at the same height above the black hole.
If you are not only asking about the time dilation but also what one sees with his eyes, you also have to take the regular Doppler shift into account, which makes the received signals redder when they move away from each other, and bluer if they move towards each other. Which effect is the strongest depends not only on her exact position but also on her velocity and vector relative to Bob.
For example, if Alice falls from rest position at r=5rs, and Bob stays fixed at this position, Alice reaches the horizon after a proper time of τ=33.7GM/c³. If she watches Bob, the time she sees on his clock in the moment when she crosses the horizon would be t=27.328GM/c³, so Bob's signal would visually appear redshifted in her frame.
Since her free fall velocity is smaller than the escape velocity (because she fell from rest at finite height) Bob's clock would still be ticking faster in her frame though, but in this scenario the effect of the regular Doppler shift would trump the time dilation visually.