Joe Meek's studio was based in a flat (apartment) above a leather-goods store in Islington, London. He bounced tracks between a pair of reel-to-reel machines to build up the song, commonly adding additional percussion at the final stage to brighten the mix and compensate for the unavoidable frequency loss.
 
He produced such songs as 'Johnny Remember Me' (John Leyton), 'Have I the Right?' (The Honeycombs) and 'Telstar' (The Tornados). It's widely reported that Phil Spector's 'Wall Of Sound' style was based on Meek's signature sound. During a heated telephone conversation, Meek went as far as accusing Spector of stealing his production style.
 
Just imagine what he'd have done with a decent laptop and a copy of Sonar.

Because we digitize tapes we get a lot of this original tapes.
We hear the stories of the makers that played a guitar-solo 20 times on the 24-track-machine to make it sound fat. Then the complete tape was played and synchronized back to the main 24-track-tape as a stereo track with a lot of flanger and echo effects.
It's nearly impossible to get the original tracks. And they did the same with vocals.
They all came with perfect composed songs to the studio and recorded them.

The performers were called artists...

I think most people don't get the answer to the original question that started this thread. I'll try to explain again but a shorter version this time. Back in the analog days, the mix was 80 - 95% complete before you had even finished tracking. You would be committed to the sound you were printing to tape due to the constraints of equipment i.e. no reusable plugins on every track. Basically, most of your inserts on a channel would be committed to tape along with the channel source. Tasks like comping vocal takes would be done as soon as the vocal recording was complete. All takes would be redone unless absolutely there or at least had a really great bit at 2:03:56. There was no auto tune or the like. The take had to be right both from the performer's point of view and from the sound engineer's point of view. You would bounce the comp to another track, each take having the same EQ and compression/de-essing applied. Also, a major point that I didn't mention before is that the mix engineer was almost always the same person as the recording engineer. So you were mixing as you tracked, hence the mix was 80 - 90% complete by the time you got to the mixing stage. That's why if you shove an old multitrack tape on a machine, it sounds pretty much mixed. At this stage, all you have to do is balance levels a bit, add some gates and erase noise, maybe recreate busses for drums, reverb & other effects sends etc. and add some compression and EQ, get the tape opp to insert some head and tail leader tape, then mix down to a two track (often muting/unmuting channels and riding faders as you went) to send to the mastering engineer. Your job as the recording (and mix!) engineer is now done and you can pop to the pub for a well-earned pint before returning to the studio to set up for the next session.

THambrecht
They all came with perfect composed songs to the studio and recorded them.

That's my experience too.  I did horn tracks added to beds that bands had already pud down.  Those bands always came to the studio knowing the material they wanted to record.  And I always had the horn arrangements fully written out.  The studio wasn't looked upon as a place for creative experimentation.  Just too expensive for that, at least for small time working bands.
 
But the economics are completely different now, and fr many people the process is 180 degrees different, with the studio time being the creative experimentation time.
 
Another way to put this is that in the 70s / 80s the songwriting  phase was distinct from the studio tracking & mixing phase. And that was completely separated from the mastering process.  Today they all seem to blend together, yet there can be a lot of wisdom in keeping the discipline of separate stages.
 
The studio I worked at most had a 4-track and 8-track, so the tape drives definitely determined the work flow.

It has absolutely nothing to do with perfectly composed songs. The OP wasn't referring to the composition but to the quality of the pre mix coming off tape which has more to do with a perfectly composed engineer!

There seems to be a consensus emerging on this thread that "people used to come in and record complete songs, but now they write while they're tracking". I wonder: how many of the people saying this work as recording engineers, recording people other than themselves? Because I do, and honestly, that's not my experience. In fact, quite the opposite; the days of having the budget to tinker around when you're paying by the hour for studio time are long gone for most bands.
 
Self-recording at home tends to involve a lot of tinkering around with the material, but that's not due to anything technical. It's not like people using Tascam 4-track tape machines in the 80s and 90s were paragons of process efficiency.
 
Myself, I don't think people are mixing up the writing stage and the recording stage; they're perhaps mistaking the demo stage for the recording stage.

John T
 
Myself, I don't think people are mixing up the writing stage and the recording stage; they're perhaps mistaking the demo stage for the recording stage.

That is probably the best assessment. My personal experience is similar in that ideas can be captured readily (and just as easily multiply and lost), and it is easy to fall into the pitfall of trying to salvage those into "final takes" when they best fit a demo. Perhaps is the belief that editing tools can make demos into final masters, but that concept is often promoted by editing tools. This is not meant to imply these tools are bad, but overuse and dependency is as applicable as elsewhere.

Yeah, exactly. Just yesterday, I had to edit a keyboard take because the keyboard player wasn't available to re-track. Some mistakes and some iffy timing. And it's nice to be able to do that to the standard we can these days. The raw material was good enough to get away with it.
 
The thing is, though, it took me about an hour, and the song is three minutes long. If the player had been available, we could have done 10-20 takes in that time. More realistically, I'd expect to get it nailed within two or three takes.
 
So once you're in the habit of trying to edit experimental noodling into finished tracks, well, you'll get some accidental gold, to be sure, but your overall efficiency will have gone through the floor.
 
Write the song > rehearse the song > tweak the arrangement in rehearsal > play it live a few times if you possibly can > track it. Never fails. Other methods can produce great stuff, of course, but they're all less optimal, at least in terms of music performed by a band.

So in the old days of vinyl & tape, how did they cut a song to make the short version (usually the single) from the long version (usually the LP)??