The DE-5000 is probably more then you need for audio work, unless you have a reason for needing accuracy for a particular reason.
200pF range is fine.
For pico farad level readings for 2 wire capacitors, accuracy only makes sense if the test rig is precisely defined, and the capacitance will be different in an actual PCB.
So the Agilent 20pF range is probably more for matching capacitors, or checking the tolerance range in a batch, rather then for measuring the actual capacitance value. If you need capacitors in the pf range, you will probably pick the final value by testing the board, rather then by design.
The only way to get accurate sub-picoFarad readings is to use 3 terminal capacitors where one terminal is a shield that encloses the capacitor. With 3 terminal capacitors, you can measure accurately to attoFarads with the appropriate gear.
In fact the standard for Capacitance is usually based on the LAMPARD-THOMPSON CALCULABLE CAPACITOR (double thumbs up to CSIRO in Australia) which is usually about 0.5pF and I think accuracies have been achieved to better then 1 part in 10-15. That means measuring down to billionths of an attoFarad.
Beat that, Agilent! Back to the old transformer bridges and heaps of volts.
If you are designing audio boards for manufacture, then your design should aim at not needing better then 1% resistors, 5% capacitors, and if possible, no accurate inductors. If for some reason, you need capacitors better then 1%, then you will have other issues to deal with, like the change in capacitance with frequency, the thermal and long term drift of capacitance, and the effects of the internal inductance and resistance of the capacitor.
Richard.