yeah this is easy to do.
first, map voices 1-4 to desired outputs. if you just want to use module outputs they are already pre-mapped - voice 1 to output pair 1, voice 2 to output pair 2 etc. if you want to map them to just friends / telexo / er-301 you need to map that manually and enable i2c.
next enable only 1st voice for pattern 1 (pattern voices are set in the 3rd column from the left on the voice assignment page, see the diagram above), only 2nd voice for pattern 2 etc.
now you can program each voice independently by switching between patterns 1-4.