The Role of Computers in Prog

A PC can play basically any sound a digital synth can generate. In a PC workstation you can upload libraries of sounds (which can be sampled from real sounds into digital or generated directly digitally, so you can have whatever you want, Moogs, Hammonds, clarinets, bells, you name it (plucked string instruments such as guitars are however extremely difficult to render realistically). You then assign which sound you want each track to play. For the notes to be played you can either input them using a keyboard or other instrument with digital interface, or you can write them directly with the computer keyboard. You can also introduce all the subtle effects you can do on a physical keyboard synth such as pitch-bends, key velocity, modulation, vibrato etc, but programming all these one by one can be very time consuming if you want a result which sounds reasonably natural.

I understand the reasons for people generating their music with a computer (I have done too as shown in my post above), but if you want a good-sounding final product I'm afraid that often that can take much more time than recording from real instruments, because programming all the subtleties that you do naturally when playing an instrument can be very time consuming.

I agree. Many if not most DAW's have functions to partially cope with this, if I remember well in Cubasis the function was called "humanise" and it would slightly alter the perfect quantisation to make it sound more "natural". The quantisation rate can also be adjusted to less than 100% for the same effect. Other tricks are avoiding a tempo which is too steady, at the change of phrases or sections, introduce slight slowing downs or speed-ups. Applying modulation and slight bends at the beginning and end of the phrases, and avoiding flat volumes also help.

But indeed it all depends on what you are trying to do. These comments are related to trying to emulate the sound of real instruments with a computer, but when that is not the case they may become irrelevant. If you want the music to sound like computer music, then it's alright.

Good question. I suspect it is partly because a guitar is not a perfect instrument. Before I expand on that...

...another pointless anecdote:

Years ago I bought a Yamaha Portasound keyboard, these used FM synthesis and sounded great. While I was show my new toy to a friend of mine (who was an excellent piano player and a good guitarist) I pointed out that disappointingly the "guitar" preset didn't sound much like a guitar, and demonstrated that by playing a few bars of Life On Mars, "See" I said, "it sounds like a keyboard". He smiled and said, "that's because you are playing it like one". He took the keyboard from me and played it. Low and behold it sounded a hell of a lot more guitar-like when he played it, not exactly like a guitar, but enough to make me hear that the guitar preset was indeed a guitar sound.

...anyway. A guitar is not a perfect instrument...what do I mean by that? Unlike a concert violinist in an orchestra, a guitarist is not trying to make his guitar sound exactly like every other guitar. While the violinist strives to play the perfect ideal note, the guitarist has no truck with that. A sampled "perfect" guitar note played on a midi-keyboard will be perfect every time, guitars aren't like that because that's not how guitarists play them. As you observed, a sampled acoustic guitar is better, not much, but better and that's partly due to our expectations of how an acoustic guitar should be played and how it should sound, with each note approaching the perfect ideal note.

In my earlier pointless anecdote I explained that I modelled a guitar using 6 modules, one for each string, but there was more to it than that. A guitar is, as we know, polyphonic so it can play more than one note at once, even when playing with a plectrum picking single notes one at a time the sound is polyphonic with each string adding to the total sound. However, each string is monophonic - if I play a C on one string and then an E on another I will get a chord but if I play the C on one string and then an E on the same string I do not. This is a very obvious thing when you read it, it isn't quite so obvious when you play it on a polyphonic keyboard or computer score. When a guitarist solos he uses all strings with some notes still ringing out as the next is played on a different string and some not if they are on the same string, this is something that is quite difficult (or perhaps even impossible) on a keyboard. But on the computer score you can do this quite easily so I programmed my six guitar modules to be monophonic and carefully chose which note to put on which "string", deciding which notes I wanted to harmonise and which I didn't.

Robert Fripp. ;-)

Also. The hypocrisy is that Keith Emerson, sitting in front of a Moog Modular, is "prog". If he plays on a VST Moog, by the argument, it's no longer prog....... and if he plays on a non analogue synth, let's say a Nord, that's not prog, either. ;-)

Well, duh. That's the point. 

Biko by Peter Gabriel. When I saw it performed live, a group of bagpipers proceeded down the aisles of the auditorium.

Great moment.