Try this

Input:

This line Transcribed, For David Jones #
This line Arranged, For Mike Johnson (great)
Terrible weather today
[download]

Program:

#!/bin/perl
my $file='regex.txt';
my $work = do {
   local $/;
   open my $fh, "<", $file
    or die "could not open $file: $!";
   <$fh>;
};
while ($work =~ /(Transcribed|Arranged),\s+For\s+([\w\s]+)/g) {
  print "Match: $1 Person: $2\n";
}
[download]

Output:

Match: Transcribed Person: David Jones
Match: Arranged Person: Mike Johnson
[download]

Comments:

The \b (word boundary) matches have been removed as they don't really do anything. Note that it looks for (Transcribed|Arranged) as suggested by other The code looks for normal alpha characters and spaces as a name, terminating on the first that doesn't match, but this can be easily changed

In Section Seekers of Perl Wisdom