The vocoder is not an effect for the microphone, per se. It is the AD Input being used as a modifier for a synth Voice effect. The synth Voice is the carrier. It is designed for use with a microphone but no one says that is what you have to use as the modifier. Anything you can plug into the AD INPUT could theoretically be used as the modifier in this situation (your mileage will vary… as I said it is designed for use with a microphone as the input). The word Vocoder comes from Vocal Encoder (developed originally to encrypt battlefield communication).
So the Vocoder is actually an effect applied to a synthesizer sound. The input device is the modifier. The carrier (what you hear) is provided by the synthesizer sound that you select. The richer that sound is in harmonics the better it is as a carrier.
There are four Preset Voices already setup for ‘vocoding’.
Preset 8: 125(H13) Vocoder with Delay
Preset 8: 126(H14) Funky Vocoder
Preset 8: 127(H15) Vocoder Pad 1
Preset 8: 128(H16) Vocoder Pad 2
Motif XF facts:
_You can sample the AD Input. This we know.
_You can route the AD INPUT to the Large Insertion effect of a Voice in PART 1 for ‘live’ vocoding.
_The sequencer of the Motif XF is designed to record MIDI data (it, in fact, never records audio… You create a USER SAMPLE VOICE that is triggered by a MIDI note-on event).
_The only way to sample an internal synth sound (as audio) is to use the Resample feature. (So, we will use the XF’s SDRAM to record the Vocoder to the sequencer).
• Press [UTILITY]
• Press [F2] I/O
• Set MIC/LINE = MIC
• Set “FW Monitor Setup” to “with PC (DirectMonitor)”
This will allow you to monitor the FireWire assigned audio channels mixed together with the main output (L&R). Next route your recorded tracks to the assignable FW outputs… Anything routed to the “L&R” output will be ‘resampled’ (recorded). So by sending the pre-recorded tracks to a numbered pair of FW outputs we can hear them, but they do not get sampled.
Route the pre-recorded PARTS to the FW1&2 Outputs.
In order to keep from recording (resampling) the internal Parts along with your vocoded Part, route them to the assignable outputs, FW1&2, as follows:
• Press [MIXING]
• Press [EDIT]
• Press Track Select  -  in turn to route each PART to FW 1& 2
• Press [F1] VOICE
• Press [SF2] OUTPUT
• Highlight the OUTPUT SELECT parameter and press [SF6] LIST to view a list of potential outputs (Shown below)
• Set the OUTPUT SELECT = FW1&2 for all playback PARTS 2-16, as necessary.
Leave PART 1 assigned to the main L&R output (it is already set there by default).
Route the AD INPUT to the Large Insertion of PART 01:
This sets the microphone as the modifier for the Vocoder From [MIXING] > [EDIT] mode:
• Press [COMMON EDIT]
• Press [F4] AUDIO IN
• Press [SF1] OUTPUT
• Set the OUTPUT SELECT for the AD INPUT = “INS L”
Edit PART 1 of your Mixing setup so that it is set to use the Vocoder:
Or use a Voice already setup for the Vocoder.
• Press [MIXING]
• Select PART 1 by pressing Track button 
• Press [F6] VCE EDIT to drop into full Voice Edit
• Press [COMMON EDIT]
• Press [F6] EFFECT
• Press [SF1] CONNECT
Setup the “Insertion Connect” parameter: Set this parameter found in between the INSERTION A and INSERTION B boxes (Insertion Connect) so that it reads “INS L” This will allow you to select the “Vocoder” as the INSERTION L (Large Insertion) TYPE for the synth sound in PART 1. Where the AD INPUT will now be the modifier.
Or you can use the Voices: PRE 8 125-128 (H13-H16); which are already setup to use the Vocoder.
Setup to resample
In this manner the only Part being re-sampled is the synth Part on PART 01 (which includes your Vocoder signal). The other parts will not be recorded (Sampled) because they are routed to the assignable Outputs, which takes them out of the mix being sent to the sampler.
• Press [INTEGRATED SAMPLING]
• Press [F6] REC
• Set the SOURCE and DESTINATION parameters
• Make sure RECORD TYPE = SAMPLE+NOTE – this will ensure your sample will be automatically triggered on playback
• Set INPUT SOURCE = RESAMPLE
• Set your Destination TRACK 1, KEYBANK C3
• Press [F6] to move to STANDBY - check your levels on the Meter; set your TRIGGER MODE - for the type of sampling you are doing (use MEASURE and set Punch-In and Punch Out points… etc.)
• Press [F5] START to arm the sampler. “Waiting” will appear on the screen… the sampler is waiting for you to start playback of the transport. (If your vocal begins at measure 1 you may want to setup the SAMPLE COUNT-IN function found in [SEQ SETUP] > [F1] CLICK. You can have the metronome count-in 1-8 measures prior to starting the sampling).
In order to monitor the other musical parts you simply need to route the Motif XF outputs to the assignable FireWire outputs (FW1&2) this works whether or not you actually have the FW16E2 installed – we only need to hear them and the FW MONITOR SETUP parameter will allow you to monitor anything assigned to the assignable outputs for the purpose of overdubbing. You have the main L&R outputs (your synth/vocoder sound) and the FW1&2 (the rest of your PARTS) routed so that we can monitor – by setting the FW Monitor to “with PC (DirectMonitor)” we can hear both the PARTS we assigned to the assignable outputs and the internal L&R. Use the headphone jack on the XF; so you can turn your speakers completely OFF, for isolation. Now you can balance how loud your backing tracks are compared to your ‘vocoded’ synth sound by using the CS on each channel.
When you complete the re-sampling session press the [SF1] AUDITION button to hear your recording. If it is a take you want to keep, press [ENTER] to keep it.
• Remember to set the OUTPUT SELECT for your PARTS 2-16 back to the main L&R output for playback
• [STORE] your Sequence and Mix.
• When you have completed working with Resampling the Vocoder don’t forget to return the “FW Monitor Setup” parameter back to “stand alone”.
[UTILITY] > [F2] I/O > FW Monitor Setup = stand alone