It all depends on what you want the demo to be...

Do you want to capture your 'live' sound, warts and all? A simple two track recorder will do that fine.

Want to edit the performance to get rid of any flubs and clams, and then sing after it is recorded? You can still do that with a two-track. Record the arranger performance to the built-in sequencer. Edit there 'til the way you want it, then record to the two track and sing over the top of it at the same time. At least if you screw up the singing, the arranger parts are recorded, so you don't have to get both perfect at the same time..!

But if you want to get up to commercial release standards, even if using an arranger, a multi-track software can help put the 'pro' touches on you might need. In this scenario, you still record the arranger performance to the built-in sequencer. But then you transfer the SMF to your computer, and open it in the DAW program. You can now solo each individual track (after editing into shape) and record each arranger part separately. Even each drum sound or subgroup individually if you prefer. You can now do far more extensive EQ, compression, transient design, effects, and automation to make it more polished.

Then you can record the vocal, and use far better comps and eq's than are generally built-in to arrangers, plus a touch of Auto-tune if you need it. Be sparing, but it can do wonders without getting all Cher-like if you are careful! Now's a good time to add REAL backing vocals rather than the harmonizer, too, if you want that extra polish... (plus a real guitar part or horn can make it sound less arranger-like). You could even edit the sequence to make all the fills just a LITTLE bit different, add a little bit of variety to the repeated nature of the arranger.

How much you do is up to you, but amazing results can be done, with as much care as you'd use on a real band recording. But, unless you want a very simple 'this is me, live!', I would always recommend, at the very least, using the arranger's sequencer to make the initial 'capture', because it is the easiest to edit at this point. and then overdub the vocal afterwards (or record it at the same time you are playing back the edited sequence).

Most of us fall into the trap of having our own played parts and vocals louder than they need to be if you listen objectively, after the recording, and there's nothing you can do about this if you do a simple one take stereo recording. This two-step method at least stays close to being live, but allows easy fixing of these basic problems, as well as more complicated ones before the actual audio recording takes place...

Hope this helps.