When taking DARM transfer functions at low frequency, one has to be careful to inject sine waves with low amplitude that don't break the lock or saturate the PCAL optical follower servo, while still having high enough amplitude to achieve coherence above the noise.
These limitations have resulted in calibration TF sweep templates that take on the order of an hour to run.

Broadband injections are fast, but often cause locklosses or PCAL saturations as gaussian noise causes strong DARM actuation in a sensitive part of the spectrum.

One possible time-saver would be to inject multiple sine waves at once.  
In March Dan Brown and I wrote a class in ISC_library.py called SineMultiple, which is an awg.Excitation class which allows the user to define the amplitudes, frequencies, and phases of the sine waves.
After running some test injections in the IMC, there is some under-the-hood phase optimization going on in awg and awgbase which automatically finds the optimal relative phases of the sine waves to minimize the "crest factor" equal to the ratio of the signal peak to rms.  That is awesome.

I ran a three sine wave injection at 4, 12, and 20 Hz simultaneously into IMC-L_EXC and measured the OLG.  I then compared to a broadband injection.  Seems to work.

EDIT: I made some plots illustrating crest factor.  The second pdf shows three sine waves with different phases.  The third shows the crest factors vs phases for different frequencies.  Seems that the 3 x the original frequency gives the lowest overall crest factor.