The way I hear it, it's a simple 16notes 4/4 at an average of 71.8bpm.
The thing is, it is very dynamic and it's been played loosely.
One thing you should do is record the drum first and play in time with the drum not the other way around. It never really worked for me to reference to a guitar track when playing drums. Where is the time, I ask ? When the first note of the chord is touched ? or the last note ?.. There lies the time differential.

Don't try to swing it... Just implement the right dynamics... They weren't making time aligned records back then ;)
(my humble opinion anyway)

pcrecord, post: 449178, member: 46460 wrote: The way I hear it, it's a simple 16notes 4/4 at an average of 71.8bpm.

Marco is right, it's a pretty straight 4/4 pattern, it's not like it's 5/4 or 7/8 - BUT - there are a few sections where things like fills or pickup notes might be off, (or crash cymbals) if your loop is set for "straight" 4 or 8 measures (depending on the grid/loop pattern)...

DogsoverLava, post: 449170, member: 48175 wrote: If I record to a click I make it through fine but I add a drum loop and there are sections it just doesn't fit (the pattern get's out of sync before realigning).

Yeah, the loop would be off here and there because they're not doing the "typical" 4 meas or 8 meas drum pattern, nor are the other instruments - including the vocal phrasing - following a typical 4 or 8 measure pattern... so you have phrases and measures starting at places where you wouldn't expect them to ( which IMO makes the song cool ;)).
As Marco said, the time sig is a straight 4/4, but if you inserted a standard 4 meas ( or 8 measures) loop(s), with the fills ( or crash cymbals) either playing the standard 4th (or 8th) full measure - or even a part of those 4th or 8th measures to segue into the next section - it'll sound "off" because your loop is pre programmed to play the common phrasing and beat count of a 4 or 8 measure pattern with a "predictable" turn-around, so you'll have fills or crashes maybe happening in the middle of odd places, beats and measures ( like a fill on measure 3, or a crash on measure 2, or 6, etc.).

You're best bet is to ditch the drum loop workflow, and program your own pattern for this. I wouldn't rely on pre -programmed loops for the exact reasons mentioned above.
Doing your own will also let you program in cool dynamic sections, too.

FWIW
-d.

Addendum:

If you decide to do your own drum programming, don't quantize this too much. As Marco pointed out:

pcrecord, post: 449178, member: 46460 wrote: The thing is, it is very dynamic and it's been played loosely.

He's right... this does have a very "relaxed" kind of groove to it - it's played in time, but it's also played with a laid-back feel. If you quantize the parts too tightly ( too precise to the beats), it's gonna sound very stiff and robotic, and one of the reasons this song is so nice is because of its relaxed feel.

IMHO.

Maybe stick the original song in the project and just paste hits in beat by beat. Some songs just don't fit in a grid. You may be able to construct some patterns you can reuse, copy and paste, throughout the song.

DogsoverLava, post: 449170, member: 48175 wrote: If I play a simple drum pattern live with my fingers I can make it work but I'm struggling to translate that into programming on the grid (as opposed to playing it because I have no skill at finger drumming).

Transient detection software

I've been doing this sort of thing for years. The software to sync to just about anything has really advanced in the last few years.
Here is a perfect and very timely question that will be asked time after time as we move towards online collaborating with other musicians and studios around the world. The category is mixing: Learning and perfecting the art of pieced together music.

There are methods to "sync and sonically glue" any timeline using the transient peaks of a performance to follow a loose performance without time correction. You can do it inside example: Samplitude. You can also pass any track over to a drummer (other musicians), have them play live to something and use sound replacement software to make seamless integration to all or any section of a timeline.

Simply put, its the art of learning how to use transients to trigger and follow. Practice this and I guarantee it can be done without notice.

Rob, you might want to upgrade to Samplitude. It does this kind of thing with flying colours.

This is what I use but it is much improved. It now includes real time transient detention that can sync anything to anything.
This video was done years ago.

[GALLERY=media, 42]Samplitude - Audio to Midi Transient detection - YouTube by audiokid posted Mar 23, 2015 at 8:48 PM[/GALLERY]

NOTE: Time correction and "quantizing" is the focus here but you use this same process to sync "out of sync" tracks as well.
The key is to find the transients you want to follow in a session and use that as the trigger to keep everything lined up to that as the main grid. You have endless options as how much you want a relaxed feel. And if you really want to be true to the exact measures... you can take it all down and do things bar to bar detail.

It takes some practice, and not everyone can think like a "natural" musician who can program to sound organic, but if you study this stuff the rewards and possibilities are endless. It really comes down to knowing it can be done and learning how to use things to do other things for you.

Which is why I am so excited about spectralayers next. It translates audio into colour detection technology which can trigger or remove millions of things that go way beyond the transients.

Thanks guys:

It was that loose feel where the session drummer was able to follow the feel of the song compared to my grid aligned drumming I was trying to do that was confusing me. Because of the idiosyncratic structure of the song (not being in a 4 or 8 measure pattern as Donny pointed out) looped drum patterns would suddenly feel or sound wrong, then sync up right again - and as the feel and phrasing of the song changed I had this disconnect between performance and time. I was worried maybe there was something unique and deeper going on that I was missing.

I hadn't thought about beat mapping this but that's a great idea. I have done this before - that just may be exactly what I need here to set up my matrix. I'll consider beat mapping based off of my scratch tracks and/or the original song. Thanks Chris and boulder for that suggestion. I'm at the "I can do this" stage of this project and to move forward I want to get my drums down and locked it (I figure that will take be at least a month to 2 or 3) while I work out the actual arrangement for the orchestration etc. But I'm totally in love with this version of the song and the George Martin orchestration. I had a very good friend work for George in the early 90's when I was resident in the UK so I've started the process of trying to track down the original score he used for the orchestration (if it exists). It's a treasure hunt for sure but you never know.

Thanks again guys - there are a few decision trees here but you've helped confirm a few things and cleared up some doubts I had.

I've been producing loads of tracks in the past six months -as faithful copies as I can of original, mostly 60s material from one band. Up until then, the most tricky tempo mapping I'd done was (of all things) Carpenters songs. Cubase = that I've used forever is pretty good with tempo mapping, but it just can't cope with the Carpenters - The male and female voices plus the piano are all over the place. The tempo map goes wildy up and down, yet listening to it you don't get a real sense the tempo is changing so much. So bad that any kind of cut and paste edits just didn't work. Once the tempo map was done, it did mean I could use the drum editor, but even then a few beats needed shifting - even with the tempo map done as accurately as I could.

However - many of the 60s recording had pretty elastic tempo, sometimes even with what sounded like decent timekeeping, but also clearly varispeeded too as the tuning was also up and down. I'm quite sure many songs were actually played in C, but slowed down to B. Again, Cubase couldn't cope.

Trying to make a grid fit some recordings is just impossible.

The reason that so many of these older songs don't fit a precise tempo, is because using a click track at that time was pretty rare. Accordingly, the reason that they don't sound out of tempo when we listen to the songs, is because in most cases, the drummers set the tempo and the other session cats followed the drummer, so that any "push or pull" that was happening was going on as an ensemble.... everyone was pulling or pushing at the same time, so on the surface, listening to the song as a whole, they sound tight and well-knit rhythmically. It's not until you place the songs against a metronome track that you can hear the drift, the push and pull. And to be honest, I'm GLAD so much of that stuff wasn't played to a "precise" click, because having that human element involved infused a sense of feel and groove and pocket in so many of those songs, that they would have not otherwise had, had they played to a click.
The metronomic reference becoming a standard in recording didn't gain wide spread popularity and acceptance until disco came onto the scene, mid 70's, much of which was using early sequencers and drum machines. Donna Summer's "I Feel Love" was one of the early ones I remember having a sequencer driving a synthetic kick drum throughout the track. There were other songs at that time too, where early versions of monophonic sequencers were being used... There was kind of an early techno instrumental track called "Popcorn" that used a sequencer. By 1980 they were standard studio tools, and could be heard on songs like M's "Pop Music". The floodgates opened from there when the Linn Drum was released.

Hire a drummer.

bouldersound, post: 449220, member: 38959 wrote: Hire a drummer.

YA - there's always that.

For me this is a great project because I'm really digging into workflow with respect to how I record, and setup for recording and how I will deal with the question of time and programming drums. I'll write up something on exactly what I'm experimenting with but I'm experimenting with this idea right now:

1. Record scratch guitar tracks and vocals to a click.
2. While playing the recorded tracks record a live-take midi beat to mark the 1, 2, 3, 4 (sans click)
3. Beat-Map a grid based on the midi beats
4. Using this new grid program my drums
5. Record keeper guitar tracks and vocals with new drum tracks
   What I'm hoping this does is allow me to escape the oppression of the time exact grid and program-in the "feel" of my own musicality....

DogsoverLava, post: 449284, member: 48175 wrote: YA - there's always that.

For me this is a great project because I'm really digging into workflow with respect to how I record, and setup for recording and how I will deal with the question of time and programming drums. I'll write up something on exactly what I'm experimenting with but I'm experimenting with this idea right now:

1. Record scratch guitar tracks and vocals to a click.
2. While playing the recorded tracks record a live-take midi beat to mark the 1, 2, 3, 4 (sans click)
3. Beat-Map a grid based on the midi beats
4. Using this new grid program my drums
5. Record keeper guitar tracks and vocals with new drum tracks
   What I'm hoping this does is allow me to escape the oppression of the time exact grid and program-in the "feel" of my own musicality....

Sounds like a good plan.. Having the drum to follow when other instruments are recorded is the best way to make it all work.
Drum programming is tedious when done on the staff or a grid. Having the chance to be a drummer, when I have to create a midi track, I prefer sitting down at an electronic drum.