Following closely, hope they can address this soon!

I'm seeing MUCH better results than that.
 
Here's my example:
 
I've cloned my kick drum track. In the track you're getting snare and tom bleed, which you can clearly see. Both tracks are in view, the top one is the one I'll apply the gate to, the bottom is untouched.
 
First, this is how I set up my gate:
 

 
EDIT: Link to full image here so you can see the gate settings better: http://i8.photobucket.com...ample1_zpsetnoatxt.png
 
Now bear in mind that this effected track sounds TERRIBLE. But it's only there for the detection. Essentially you're just getting pops where every kick hit is. 
 
After you apply the effect, it looks like this:
 

 
When I enable Audiosnap on the tracks, just leaving it as the default threshold, I get this result:
 

 
A closer look:
 

 
Full image here: http://i8.photobucket.com/albums/a3/lordtim/SONAR/as_example4_zpsrfejr3r3.png
 
As you can see, the bottom un-gated track is terrible. Early detections, ghost notes... you name it. Even playing with the threshold won't get you much better than that, and you'll need to do a heap of manual work to make sure everything is where it should be.
 
On the contrary, the top track is dead on the start of each transient where it should be.
 
As far as saving times goes, there's not a lot you can do about that. If you're making sure your multi-mic kit is all phase coherent, you'll need to apply the transients across all of the tracks at the same places anyway, so one extra guide track is negligible as far as save time goes.
 
Now don't get me wrong, I'm not defending Audiosnap - this NEEDS to be fixed. It's a tool I've used on literally hundreds of commercial tracks over the last year or two and the less screwing around I need to do to just make it work, the better! But the gating thing REALLY helps for me and dramatically speeds my workflow up.

brOOtal explanation Lord Tim, very useful

Cheers, man! 
 
I think the other thing that needs addressing is the manual or the tutorials in using this. Cake's Dan Gonzalez did some excellent tutorials on multitrack drum stuff including a bit of Audiosnap things, and his methods - while effective - were a little confusing, and the manual made my brain hurt trying to understand the method they prefer you to use.
 
I don't think this stuff is as hard as it appears to be, and fighting with the tools is just making the matter worse...!

I´m still intrigued by how efficiently Pro Tools & Reaper manage clips compared to Sonar that begins to slows down as soon as you start cutting & moving

It sounds like there are really two issues here:
 
1) Audiosnap's transient detection could be a lot better.
2) Sonar's clip management needs to be more robust and able to handle lots of clips without falling over or slowing down.
 
I'm in...
 
Regards,
Dan
 

Absolutely!
 
If you think about it when you use the Split at Transients method, you might have maybe 100 splits in a given song? 200? Then combine that over 12 tracks of drums. You now have a LOT of tiny clips to deal with, and if they all are Audiosnap enabled, that's making SONAR go "er, what?" at you. So some kind of clip caching for that kind of thing sounds really necessary.
 
This is another reason I prefer to use the transient stretching rather than splitting because I find there's a huge performance gain, even aside from having a lot more control over fine tuning the transients/quantize after it's done and before you bounce to new clips using the higher quality offline render algorithm.
 
But for me, just adding a gate into the pre-detection routine bypasses a lot of the bad detection right away. There's no question the detection algorithm needs work, but I think it'd need far less work if there was a pre-gate (with EQ like the Sonitus Gate has) built into Audiosnap before it hits the detection routine.

I too do this many times and agree there needs to be a lot of work done on this workflow. I also agree with Lord Tim about gating the guide track and also thank him for the detailed info. 
 
I thought I might also add a few random tips that I sometimes use that might help someone. 
 
- I find that quantizing without splitting works pretty well if you have very minor adjustments to make. If you have larger or more difficult problems it causes problems for me, so splitting becomes necessary.
 
- If you split and have lots of splits, yes saving and working with the data becomes really slow. As you put things back together they are not as slow any more. It would be nice to be able to select a section and work on it, but AS ONLY works on the entire track. So to work on a section, you have to split it somewhere (say at the end of the 1st verse) and copy that to another track. Then you can split/work on that section and "bounce to clips" to put them back together. Then copy that back to the original track, select the next section and do the same.
 
- I also sometimes use melodyne to quantize. Even Melo essential does a much better job of this than sonar. However, since you don't have multitrack, you have to align everything in a track to the grid, then do the same for other tracks to maintain alignment. If you have a pretty accurate track, this actually works better for me than doing it with audiosnap.
 
- Sometimes I'll use both. Split a group of tracks at beats 2 and 4 ONLY and align those to the grid on 2/4. This gets everything pretty close, but the beats in between are some times a bit off. Now bring those into melodyne one track at a time and quantize. Render those and bounce to  create clean tracks.
 
- Drum replacer can also act like a gate. Use DR to replace a kick track and it eliminates all the bleed just like a gate does.
 
- If you're not going to keep the track in the final, then don't worry about pops/clicks. Sometimes the final intent of my work is to create midi tracks from the part and trigger AD2. If I'm doing this, say for a kick track, then I can work on the original kick and not worry about the artifacts associated with splitting or moving around hits as they won't wind up triggering a note anyway.
 
- Another thing I've posted about that I wish would be changed is the behavior of grouping when you use AS to split. If you have all your tracks in a group, split them with AS, the result is that all the clips are still in one large group. What we really need is for each vertical slice of clips to be in a separate group as that's the way we want to work with them. If you split manually that's what happens. So if I'm going to only split on the 2/4 snare hits and maybe a few other place, then I sometimes create a group of the tracks and split them manually with the split tool. It's time consuming, but more accurate and everything winds up in the correct grouping structure. 
 
 
That's a short list of the top of my head. I'm also considering purchasing the studio version of melodyne 4 which does multitrack. Their method of working with "blobs" is better than the idea of working with transients. However, at the moment we would still be using it stand alone to edit multitrack drums and exporting/importing the finished tracks to sonar.
 
gabo
 

@Lord Tim, thx a lot for this very deailed and useful explanation of a possible workaround. As a matter of fact, if the detection was the only problem with Audio Snap it would be somehow ok and acceptable. But from then it is still a nightmare to edit multi track recordings. If you have ever used a software like PT or Reaper you knows there are much easier and faster ways. For instance the fact that you must copy (merge) "all" markers "before" you can start edit all tracks together. A double click should not only select all adjacent markers, it should create them on the tracks without detected markers.
The other already mentioned problem is the performance, which maybe has to do with the hard disc driver structure. If our drummer records 16 tracks a 7 takes, you have about over 100 clips. Editing and Comping with a SSD hard disc works fine, but as soon I switch to my place or the studio (with just normal HDs) it becomes horrible slow and unstable. I could either buy a SSD or transfer the project to PT (100+ tracks is easy going there, also with moderate hardware).

Once you do have the markers all quantized together, you *can* double click one and it selects all of the tracks with the markers that fall on the same beat... well, in theory. That works about 80% of the time, and other times it either selects none, partial, or grabs markers around the ones you're trying to select. Pretty annoying!
 
One good thing about the transient smart tool is it's easy to left-mouse drag around the markers you want to select and select them all. If you've quantized the track, just turn your project snap on and you'll be able to drag them to the nearest beat quite easily.
 
But I do agree, if you've got everything locked together, one marker should move the others by default when you drag it around.
 
Since I don't use the split method, I don't find any slowdown at all. In fact, if you check out the specs in my signature below, you'll see I'm running a pretty low powered laptop for my recording computer, and even worse, my audio drive is a USB2 external drive. But I'm getting very acceptable performance out of it, even on really big projects.  I'd expect that to change really fast if I decided to split at each transient, though! That would likely bring my machine to its knees!