Extract (Base64)PDF from large >10MB HL7 messages

I'm processing HL7 messages that include extraction of multiple PDF's stored as base64 and write these to file(s). Problem is that mirth throws an heap space error for larger HL7 files (Usually larger then a few Megabytes). I've already increased Java workspace in mirth configuration but without luck. I'm using mirth 3.0.2.



Any suggestions how to process these kind of larger messages ?

Extract (Base64)PDF from large >10MB HL7 messages