Question for you folks who create your own music.

I'm a guitarist, I have Toontrack's drum software as well as the drum software that comes Sonar X3.  My question really is not about the drums, but rather, do you ever create a drum track from which ever synth you use, and then create a song over that? And if so, do you have a formula you have a basic formula from which you start?
 
As in,
 
you have an intro that lasts so many measures
you have a verse that lasts so many measures
A pre chorus that lasts so many measures
A chorus .....
on and on.
 
I'm looking to to create some drum tracks and start from scratch creating some 80's style rock tunes and would like to know if there is a basic formula to follow and if so, what would it be.  Mind you, i'm not trying to create radio worthy music, rather, this would all be an exercise in creativity for me.  Creating a drum track, bass track, guitar track(s)... 
 
Any ideas / suggestions would be very welcome. 
Thanks!

Not only do people approach it differently, the approach can vary with the style of the song.  I've been on the recording end and don't consider myself a musician--at least not at the level of most other people here in the forum.  I don't do much creating songs from scratch, but there is one trick that I stumbled onto that may help.  
 
I made a groove-clip of a section of audio and extended the clip beyond where I thought the song would end.  It then acted like a click track to keep the rhythm consistent.  The metronome tends to distract me more than it helps.  Then, I could edit that track anywhere I wanted to insert a fill or a change and be confident that the beat stayed correct for the rest of the song.  

Somewhere, somebody is writing down those steps and probably won't even really notice steps 10 and above until they get there!  LOL!

Rain
1: Song idea is recorded to metronome. Hit save.
2: Click track to replace metronome - a very basic EZ Drummer MIDI pattern or me playing a simple pattern on the keyboard through EZ or Superior. 
3: Record basic song on top of click track and work out the structure, usually on guitar or key.
4: Work out proper drum track, with most of the fills and breaks and all. 
5: Re-record proper guitar/key track and record other instruments.
6: Spend time away from song.
7: Listen to the song w/ fresh ears.
8: Give the drum tracks some more attention if needed.
9: Bounce/Freeze drums.
 
10: Start mixing the song.
11: Find something's wrong.
12: Find everything's wrong.
13: Self-loathing.
14: Consider writing down last will.
15: Get over oneself and just finish the darn song.
 
No matter what, ALWAYS save the MIDI track, clearly identify versions and write down any bit of info that seems quasi-relevant.
 

THIS^^^^^  (apart from metronome, I always use a percussion loop)
 
Lately I've been laying the song/tune out using piano. Being a guitarist I found this steers me away from getting bogged down in riffs and lead guitar too early, and also throws up some interesting changes because 'muscle memory' is removed and I don't fall into the same old patterns.
 

I think Rain nailed the steps.  However, after laying down some chords (guitar or keys) to a cilck or basic drum loop, I work on the song for a while.  I will try to finish the drums later after I have the rough vocals, guitar and keyboard parts recorded.  I like to hear those parts when creating the final drum part. 
 
Rimshot
 

 I try and use the drums as both a time keeper and for beginning emotive content.
 
  If you're going to make 80's music I would look for drum kits and sounds that represent a good 80's kit. Some drum programs let you mix and match different drum sounds. I would worry more about panning, reverb and loudness of each part after I had worked out the basic song, although sometimes I'll make some tweaks as I go.
 
 If I have a good grasp of the material I am going to be making, then I'll look for a pool of loops that fit that idea. I usually work with midi loops either within the drum program or from an outside source. Midi is great because after you get closer to completion of your material you might want to omit a cymbal here and there or add a snare. With midi you can make those changes, for instance changing the velocity or adding additional percussion for emphasis.
 
 Speaking of loop pools. A lot of people never change the 120 default tempo of their song. If you're playing the melody try and determine what tempo you are playing it in and set your project tempo to that.The tempo makes a huge difference in how your midi loops will sound. In BFD( and many others) there are groupings of loops based on tempo and styles. Sometimes experimenting with a loop made for one tempo in another tempo can make the whole thing come alive.
 
 I don't usually use a synth to make my tempo track, although I have added additional supporting percussion with a keyboard. There is no longer any point in me using a percussion map of my keyboard when I can simply get what I want from my drums program. All the map does is play midi into your keyboard and your keyboard plays the sounds in it back into your software.
 
 You make the call on how long you want your song verses to be or how many choruses you have and how long they are. Load up a drum program, in the case of X3 Addictive Drums or Discrete Drums .Send an output from each track into one track in Sonar. When you load your drum program it will have a midi control track. This is the track you load whatever patterns you want onto. Mix and match, erase, edit. Do whatever you need to do. Don't worry if you added too many measures of something. Anything is changeable until you freeze those tracks.
 
 Even after you freeze those tracks you can further edit in Sonar.

I have different approaches but since I'm not trying to write "tunes" in the sense of working with a vocalist I'm writing "exercises"...that are easy to label "instrumentals".
 
But the starting of everything or anything seems to work best when I find a drum groove that inspires me. If I can find guitar hooks that fit the groove then I have something to work with. Sometimes it just "dies" and I move on but it's the easiest way for me...loop a drum groove that I like and jam waiting for the inspiration and licks...if it doesn't happen I go to a different drum groove. To me if I can't get in the groove then it's not going to happen.
 
It's really not that much different than when I use to jam a lot with other musicians...I just didn't realize until I got to record at home that the drummer was really all I needed. Nice to have others inspire but oh well.
 
 

I usually try and start with a melody, the workings of the song comes later after jelling it all together in the end.
At least an idea of a song that then has to be worked on, and perhaps worked on, and perhaps worked on in the end.
I find drum loops are just that - drum loops. I rather play drums and I have, just takes a little more work. But while you are sitting there, there is music that can be played on utube and move your hands to the drum part.
 
I find it varies but a lot of times the bass part comes first perhaps, or initially, even if changed later.
I suppose it follows from the musicians I use to play with or a jam had in Am.

 
 

paulo
 

 
In my case ... a slight difference
 
15) Think, sod it and start on another song thinking I'll come back to that one and finish it some other time.
 
16) Never finish it.

I have gazillions of those too. ;)

There are a couple of missing steps for me in Rains method.
 
a) after what I deem to be a good mix, burn a disc, go to car and play through car stereo. Realize I've mucked it up so badly I don't know where to start.
b) rinse and repeat until worthless cds have taken over my studio. Spend an hour on google looking for uses for worthless cds.
 
Randy

My workflow hasn't changed all that much from the old days when I was playing in a band, TBH, if not for the fact that back then, we wouldn't really mix things, just create a static balance on the 4 tracks...
 
I'd pretty much always record a dummy guitar track to a click, laying the proper groove, accents and all, so that the drummer could get a sense of that. Sometimes I'd also record a temporary vocal track if he felt it helped him get into it. Then after he'd cut his drum track, we'd record the other instruments and I'd re-do the guitar and vocals.
 
The main difference is that the stuff we recorded had been rehearsed countless times, so it was always just capturing the song rather than experimenting w/ different structures and adding parts. Everything had to be planned ahead.

Do you have "loopitis"? I have it and so do many other makers of computer music.  Here's an interesting article which talks about how to overcome it...
 
http://kimlajoie.wordpress.com/2009/05/08/overcoming-loopitis/
 
Actually one of my main problems is this: I come up with what sounds to me like a great few bars of music. From playing around with these few bars, a related piece of music springs forth. Aha, this will be my "B" section. And then I start playing with that. Pretty soon, it's morphed into something that's so far removed from the "A" section that the two of them no longer connect, and I can't think of a way to make one progress into the other. But I love what I've done to both parts! Deep down, they're still related and so I'm loathe to split them into 2 different songs, but when I play them in succession the contrast is jarring. 
 
Take the doomed project I'm working on now. I started off by playing something vaguely Spanish sounding on guitar, which I then turned into MIDI with Melodyne Editor and sent to a patch in Prism which sounds like a very authentic plucked Eastern instrument (of some sort ). By chance, I happen across a couple of Indian vocal samples in the Cakewalk Content (yes you heard right) which are just MADE for the Prism part. A bit of Melodyne adjustment and some ping pong delay and a bit of percussion and I'm thinking this is the best thing I've ever done. So then I wonder how to progress. I pick up the guitar again and play with the original piece. Pretty soon I've composed something in the same mode, only this time it's in a really weird meter with disjointed phrasing. Turning it into MIDI I realize it sounds great through a Rhodes with lots of chorus. And then I get some heavy hip hop style drums and a phat bass line going. Wow! I'm listening to it loop over and over thinking that THIS is now the best thing I've ever done. 
 
So then I'm curious as to what weird ass chords it's comprised of (I rarely stop to consider things like chord names when I'm playing guitar) and so I put it through Strum Acoustic and realize that it's quite "quartal" and out there harmonically. I play the chords that Strum Acoustic is showing me (which aren't exactly the keyboard part, but close), and come up with a funky guitar part that again is related to the previous part, but a totally different feel. Pretty soon I've got a jazzy outside bass line thumping along, a Reaktor synth playing a very electro style lead and Jamstix going absolutely out of its freaking mind through Addictive Drums. 
 
Now I have 3 sections that I'm absolutely stoked with. And they're all sort of related harmonically, because one led to the other. But I've screwed with them individually so much that I'm confronted with the reality that I have three different song embryos in the same project. And it's going to be really hard to connect them. I guess I could just continue what I'm doing, come up with a few more sections, split them into their own songs and call it a concept album 
 
But boy, I find this whole arrangement thing to be very hard. I always did. 
 
 

sharke
 
Take the doomed project I'm working on now. I started off by playing something vaguely Spanish sounding on guitar, which I then turned into MIDI with Melodyne Editor and sent to a patch in Prism which sounds like a very authentic plucked Eastern instrument (of some sort ). By chance, I happen across a couple of Indian vocal samples in the Cakewalk Content (yes you heard right) which are just MADE for the Prism part. A bit of Melodyne adjustment and some ping pong delay and a bit of percussion and I'm thinking this is the best thing I've ever done. So then I wonder how to progress. I pick up the guitar again and play with the original piece. Pretty soon I've composed something in the same mode, only this time it's in a really weird meter with disjointed phrasing. Turning it into MIDI I realize it sounds great through a Rhodes with lots of chorus. And then I get some heavy hip hop style drums and a phat bass line going. Wow! I'm listening to it loop over and over thinking that THIS is now the best thing I've ever done. 
 

I can relate to this. Done loads of 8 or 12 bar segments which sound killer, and never went anywhere.  The only answer that worked for me is don't overdub anything until you have the whole track laid out with one instrument, then it's got a better chance of hanging together.

Hi,
 
This is strange and weird.
 
I take my cue from folks like Peter Hammill and Roy Harper, people that write forever and their material is always different, and makes you wonder where it all comes from. Lately, for my tastes, Peter is not as good with VdGG.
 
If you're gonna sing the song, for example, it should be more important to detail the voice and its inflections and abilities. You can do it with a piano, and a guitar ... easily. In general, from my stupid way of thinking, using a drum track or bass track first, is weird, specially if you are going to treat them as background anyway ... you lose the ability to ACCENT the voice, or the guitar, and have to make changes later, and you will be ADJUSTING to the drum beat, instead of the strength of the song, ie. its message and voicing ability!
 
That would be weird.
 
Thus, similar to what/how I write, no two situations would ever be the same, or you have no idea what you want your piece of music to say and DO.
 
It would seem to me, rather strange, and I am a person that appreciated and appreciate the empty head, all experiment stuff that had no meaning, but what we don't see is that stuff like that TEACHES you how to work with "meaning", instead of you having to scratch and claw to create something that is not totally a part of the piece, and has a tendency to split it up.
 
This is the case in the CHB, where the tendencies go in 15 different ways, WHICH, is one of its most attractive parts as to how it all came together! Tremendous work making it so, but also a bit harder in the end, and then you endup trying to accent details that were already done and changing them could be more difficult!
 
You write, for what you "see", not JUST for a drum beat, or a bass guitar -- although they can be a part of it. If you are trying to add lyrics, now, to a song you already have with a solo, the chance of it matching is not as good as coming directly from the lyrics. There is one EXCLUSION to this ... Bob Dylan, since you all know that everyone else can do his music better, but the music itself has never been important to him, as the lyrics are! We need to see that! He's not interested in Manfred Mann's solos, or Jimi's solo splash! But it tells you what you can do when you work on the lyrics that DRIVE forward.