The cables

 Here is the necessary connections: basically it's like doing a frequency response measurement, but the amplifier output is connected to the filter input and the filter output is connected to the driver and to the DATA channel of the sound card (LINE IN right); for the Reference channel (LINE IN left) we can use the sound card Line Out or the amplifier Output: don't use more than 500 mV from the amplifier Output, otherwise you can burn the sound card. Using the amplifier Out as Reference, gives some advantages: since SW measures the voltage difference between the Data and Reference channel, the amplifier gain will not be taken into account, therefore the chart will show 0 dB of filter attenuation in all the range where the filter is not working! Moreover if we sum the NTF chart to the driver unfiltered response chart, we obtain the filtered driver response.

 With the Caioli's method we measure the Network Transfer Function with the real driver impedance: thus we can verify if our filter is right as we wanted, since it is very common that, in complex filters, the components tolerance brings to different than needed filters.

Measuring a parallel filter

 Once the cables are ready, assuming you have already performed the sound card calibration procedure, put the amplifier volume to its minimum, and connect a voltmeter to its output; then create a new signal, RESOURCE/NEW/SIGNAL, or use one you already have; double click on it and then MEASURE/ FREQUENCY RESPONSE and increase the ampli volume till you read a 300-500 mV AC on the voltmeter. Repeat the measurement till you get the desired volt reading, and pay attention not to fry the sound card, especially if you are using the amplifier output as Reference: as a precaution, you can disconect the LINE IN mini-jack, remembering to connect it once the volume have been set. When the volume level is OK, repeat the measurement and open the newly created file, that is the Network Transfer Function: in the next chart I compared the NTF of the same filter measured using, as reference, the Amplifier Output and the Line Out; as you can see, the only difference is in the amplitude level. The filter is a second order pass bass, with 2 parallel RLC in series with the Focal woofer.

 Now let's take a look at the phases: here the limit of the measurement made with the LINE OUT as reference comes out; in fact at the band extremes, the phase is wrong.

 To check how reliable the measurement is, I compared  it with the NTF calculated by a simulation software using the driver impedance and the same filter used. The result below speaks for itself:

 If we have used the Ampli Output as reference, we can sum the NTF and the unfiltered driver response to get the filtered driver response: open the NTF chart (the one measured using the amplifier as reference) and click on CALCULATE/ COMBINE; a window will appear and choose MULTIPLY and then click on   to select the unfiltered measured driver response file:

 After clicking on  we will see, in green, the driver frequency response chart, obtained through the sum of the network transfer function and the unfiltered response of the Focal (it is the gated far field measurement, therefore it's not useful below 300Hz); to verify the method I added, in violet, the measured frequency response of the filtered Focal (gated far field):

 I repeated the procedure for the ATD tweeter, that uses a third order filter, and once again the comparison chart shows the reliability of this method:

Measuring a series filter

 It is possible to use Speaker Workshop and the Caioli's method also to measure the NTF of a series filter, using the symmetry of this kind of filter. The measuremet procedure is just like the one used in the parallel filter: what changes is the connection and the need to use the amplifier output as reference. As you can see below, is sufficient to change the position of the woofer-capacitor block with the tweeter-inductor one; remember that both the drivers need to be connected, otherwise the NTF will be wrong.