brconflict
Gilmour had the right idea. Done this many, many times, myself. The creative part is the fade-ins/fade-out of each phrase.

Miles Davis took this concept to an extreme with In a Silent Way. If you listen carefully you can hear all the splices on the two-track master that joined different improvised sections together...it wasn't just individual parts, but the entire mix. It's done so seamlessly you're not even aware of it unless you're listening really carefully.

In my view, the more the general public are becoming aware of all the "tricks" used in "producing" rather than creating a record, the more interested they are becoming in wholistic recorded performances, away from all the studio fixups and modifications. Again jmo, but how a singer/instrument soloist can possibly maintain an overall ambience for a song when it is done in a myriad of takes is beyond me. If the vocal is so bad that even with a gazzilion takes they need autotune, then they should not be singing. Period! This and other aspects could be reasons why many of todays recordings sound the same with the same levels as energy as the next song and lack a certain...umm spontaneity. It is why I will always try and get a live concert recording (if available) rather than the produced studio version. Far more interesting. Unless of course the performers are using backing tracks but THAT'S a whole other story
 
 

Anderton

Grem
So lets be clear: Great records of the past had tuning done to vocals before Autotune came along.

 
I used to sample off tape into an Emulator, use the pitch bend wheel to correct pitch, then record back on tape 
 

Great records of the past had guitar and vocals leads that were comped long before digital recording came along.

Punching was also a highly developed art. I've sort of re-discovered punching with Sonar, sometimes it's simpler/faster than setting up to do comping.

I'm not sure which is easier for me. If it's just me, comping is easier with punching in, but it's a terrible process to go back and navigate all the takes. I would love to see the Lanes auto-minimize each time a new one is created.
 
If it's me recording another artist, it's still days faster and more efficient to just manually operate the transport than to comp or punch in. The only thing I miss from Sonar that the old AW4416 has was the ability to hit record while the transport was playing, manually punching in the artist with zero delay. I'm an old-school guy in this way and sometimes, a mechanical-feeling transport allows the engineer to "feel" the punch in vs. having to research where to punch in, if that makes sense.

This is what I've been saying and I do it all the time. Hit play and then R when you want the recording to start. You just have to arm the track beforehand.

Its funny but I recall seeing a group known for the tightness on their albums. Live they sounded awful. I wont say who they were but they were hot in the 60s.  

I think there's a lot of misconception about the creative process and production. Personally I prefer being spontaneous however client might not have enough of that for the talent button. I've been known to spend 8 hours to get a vocalist to sing one song properly till they want to strangle me or their vocal chords fall out, and that's their own fault for not singing right in the first place.
 
The best gigs are the one's where they bring their own talent button with them, and then the issue is what to do with that much genius. Making genius sound like genius is supposed to is hard to do.
 
There are some genres though that require pure skill and repetition rather than being spontaneous. James Brown for instance would fine a band member for playing the wrong note...
 
It just depends on your approach, and who you are working with. And that generally alters all the time...
 
Sometimes I comp, sometimes I don't. Sometimes I purposely overwrite performances so I don't get too distracted. Sometimes I listen to the artist and give them everything they want. Sometimes I give them what they want but they just don't realise it yet...

brconflict
The only thing I miss from Sonar that the old AW4416 has was the ability to hit record while the transport was playing, manually punching in the artist with zero delay.

To add to what Sanderexpander said, as long as the track is armed before you start the transport, you can also click on the Transport record button as well as type R. Works like a charm.

Anderton

brconflict
The only thing I miss from Sonar that the old AW4416 has was the ability to hit record while the transport was playing, manually punching in the artist with zero delay.

To add to what Sanderexpander said, as long as the track is armed before you start the transport, you can also click on the Transport record button as well as type R. Works like a charm.

Does that work well with sample buffering? I'd question the latency of the click to the actual recording takes place.

Sanderxpander

brconflict

Sanderxpander
Brconflict, if the record mode is set to comping, you will hear the previous take as long as you haven't actually hit "record". I usually press play a bit in advance for a good lead in and then hit R a second or so before the actual line for sort of a punch in. I'm pretty sure you could also use sound on sound mode if you go line by line, and you can simply create two lanes in the same track to work with.

Just giving options, if you're happy with your workflow ignore me please :)

Right. The problem being that, when a vocalist is connecting run-on lines together (which we do, typically to save the vocalist's breath), hearing the previous line while tracking is paramount. The only way I've found this possible is to toggle takes between two different tracks. I'd love to see a quick little toggle button to allow JUST the last recorded Lane play, but mute the rest. Would help tremendously with manual punch-ins.

Unless I'm misunderstanding you, this is the way it works in comping mode. Only the last recorded lane plays. Although yes, it will make an automatic comping of sorts, taking the last recorded lane in any place. That could be take six in one spot and take three in another. But since hitting rec mutes all takes so you can only hear the current performance, that should work fine, I'd think...?

I should clarify. The previous lane is what I meant to say. For example, Lane 4 (and only Lane 4) would play unmuted, while I'm tracking the vocalist in Lane 5. That way, the vocal line in Lane 4 could be heard in the cans while the singer continues the flow in the armed and recording Lane 5. Once I press stop the Sonar could (wishfully) open and arm a new Lane (Lane 6), mute Lane 4, and unmute Lane 5. We would continue on from there. It's manual comping at its best! If it were possible, that is.

Yes this is exactly the way comping mode record works... Did you not try this yet or am I still misunderstanding you? :)
Perhaps you've been manually arming and muting lanes? That would bypass the default behavior which in my experience is exactly as you describe it.

To be clear, to get what you want, you have to PLAY the armed track (which will by default solo the previous take lane) and hit R as manual punch in when you want to start the recording. I do this all the time.