X = Objective2's maximum VRMS input signal before clipping = 7/A
A = Gain (Vout/Vin)
ODAC output = 2.0VRMS = X
2.0 = 7/A
A = 7/2 = 3.5
Setting high gain to 3.5 gives maximum amplification of 2*3.5 = 7.0VRMS. This is slightly below O2's maximum capability of ~7.5VRMS, but ensures no clipping will occur.
Most listening with dynamic headphones is (or rather, should be) conducted at low gain, since the noise floor is superior. Assuming a 1.0/3.5x gain configuration for O2+ODAC, unity gain of 1.0 gives output ranging from 0.0 to 2.0VRMS, and high gain gives a range of 0.0 to 7.0VRMS.
At a proposed high gain of 3.0, output would be limited: 0.0 to 6.0VRMS. This is a loss of 1VRMS compared to 3.5x. Why unnecessarily limit the amplifier?