Good points Bill!

To sidetrack a little, the reason I use midi files (in the keys / drums / vocals duo of which I am half) is that the public - especially the younger public - is expecting us to produce a noise which is a reasonable representation of the original. We started using sequences over 18 years back (we were a trio then) after being involved in a disastrous audition; we were the only totally live act (everyone else had tapes) and were completely and obviously outclassed. An Ensoniq ESQ1 was purchased to provide limited sequence capability, and to beef up the sound.

More recently (say 15 years back) it became obvious that when working with DJ's we were finding that even though we were doing pretty good versions of modern stuff, the younger end of the audience just would not dance when we were on. At this point I got an Ensoniq SD1 (with floppy disc!) to allow much greater flexibility and opportunity to use more sequences (the ESQ1 was memory limited).

And a few years after that I did a couple of years mid-week as a solo act (to help out a hotel who'd been let down). I got a PSR630 for that job and thus got into Midi files.

And now I'm a PSR3000 owner!

These days, even though as two middle aged gentlemen mostly singing songs originally sung by younger acts (with (a) more members and (b) girls) we find that we can shift any age group, because whilst the files are "our version" (and more especially "our key") they still are close enough to the originals that they are INSTANTLY recognisable; this greatly reduces the "I don't know this song I'll sit down" tendency.

To anwser the original question, am I try to get an exact representation?

Sort of.

I try to exactly reproduce drums and bass as these hold the song together. I like to get string and brass pretty close as well. Piano I get exact if slow and melodic and somewhere near for rapid rhythmic stuff; ditto guitar.

When transcribing a song I try to isolate the elements that make the song different and get those precisely right; so I look for the hits, fills and riffs, especially those that are not being played behing vocals. I concentrate less on the finer details of the track behind lead vocals because the audience is listening to the vocals and not the backing at that point.

The mix of the track is also biased to leave space for me to play some keys and esepcially the drummer to play lots of drums; these stops our "version" turning into glorified karaoke.

I find that the more uptempo a track, the more fine detail you can leave out because in a live situation no one can hear it anyway. These are the "go for the right feel" situations.