ok so a little history then.

before the DAW takeover, sample accurate midi timing was non-existent.

When you were hard quantizing on a machine sequencer (any machine sequencer, or any software sequencer triggering external modules, like Logic on Atari shooting midi data trough cables & interfaces to a Roland JV1080 for example), the feel was super super super tight but remained musical and usable. I won't go into details but back then the technical limitations gave sequences some life, and the midi data was distributed through cables, in a serial way, one event by one event, no 2 events landing at the same time, ever. It just added that magic feel to an event list that, if you were going through it by sight, would look pretty robotic.

In DAW's like Sonar (or Cubase for instance!) when you're shooting hard quantized midi data  to VSTi's, the feel is rigid and non musical.

In Daw's like Logic, or Reason, when you'Re doind the same, the feel is pretty musical, and even if the result is much tighter than ''the good old days'', it remains musical, and usable.
 
so, why!
 
why.
 
Anyone?
 
 
 
 

Are the DAWs in question using the same computer, same MIDI interface (if one is involved), and the same size MIDI buffer?  Is the same software running (or not running)?  By that last question, I mean make sure you don't have virus software running with Logic and not with Sonar.  Lots of things interrupt DAWs and their timing.
 
I cannot imagine how the software could make a difference.  If you quantize to a certain value, that is where the notes will fall.
 
A rigid sound is also not just a function of timing: it is also a function of velocity.  To make things equal, use a Find/Change command to set all the velocities the same in all programs.  If one DAW is playing back more velocity variation than another, it will sound less rigid.  (By the way, velocity variation is a big deal in making a piano sound authentic, even if it is quanitized.)
 
 
This may not apply to the original poster, but to test this, you also have to look carefully at what options are checked or not checked in the Quantize dialog.  For example, the default in Sonar is 50/50 (no swing), quantize start times only (not durations), and quantize to 1/16.  (I usually change those values.)  If another DAW has different defaults, you may think you are doing strict quanitzation when you are not.  Sonar's default is strict start times but not strict end times.

quantizing is just mathematics, correct or incorrect, there are no "shades of grey". I would suggest the reason you are hearing what you are hearing is you have slightly different settings happening in the two DAW's. As was mentioned above, ensure ALL settings in both are 100% identical right down to the TPQN, and then quantize the same raw midi data in each DAW, not transfer midi data already quantized. 
 
Then if you still feel you have a grievance against Sonars' operation, then by all means let us know.

What I have found is that some sythns have a greater latitude in what sound they will play depending on how they work with velocity layers. A good example is doing a drum track using SD3 and then BFD2. The BFD2 drums will sound more real because it has a lot more data to work with and the algorithms to do the job, how it decides which sample to play. It has nothing to do with the DAW used. When it sees a lot of snare drum notes played quickly it may choose to play a roll instead of discrete notes. SD3 can't do that.   

If I record myself, a guitarist by trade, not a keyboardist, playing the keyboard, I will have to quantize the midi data.  :) It might not sound "musical". But it will not annoy people like my un-edited version would. 
 
But If I record a professional keyboard player, I dare not quantize. It is not required and might trash his performance, no matter which settings you use.
If a hi quality digital piano ( or ??)  was used as a controller you can also record the analog output in a parallel track. The midi track if  sent back to the piano will sound exactly like the audio track. There is no perceivable change in the "musical" performance. It was the original performance that made the "musical' component.  
The same can be said possibly of digital drums. 
 
Quantizising was invented to repair the damage done by people like me who only play a keyboard or drums as a means to make music with a computer on hand. If you desire a "musical" performance in your recording, then input music. or learn to cheat like hell...

berlymahn
Hey, do us a favor.... let's do a blind sound test.... do a 20 sec sample in Sonar and 20 sec sample in your other software.  Same VST, same output parameters, no extra effects or processing added.  THE SAME.

Better yet, export the track to audio from both SONAR and the other DAW.  Import both resultant wav files into either DAW, line them up and examine them visually.  Since it's drums, the transients will be obvious and getting the two tracks starting at exactly the same position should be easy.  This will show you right away whether it's a playback-timing issue or something else entirely.

Seems to me that if certain DAW manufacturers were coding a "vintage feel" into their MIDI timing then they would be advertising the fact.

could it be a hw vs sw 'placebo'?

I'll try to keep this short:
 
I have done endless simultaneous Audio/MIDI re-record tests with various versions of SONAR over the years. Based on this experience, I can tell you a hard-quantized MIDI track rendered by a soft synth that does not itself introduce timing/sample/amplitude variations will play back with sample accurate timing and consistency from one playback to the next. As others have said, this is what most of us want. If any "magic feel" needs to be added to a part, I'll take care of it. I don't want my DAW or my synths doing anything but what I tell them as far as MIDI timing goes. If the synth is randomizing samples or otherwise adding dynamic or timbral variation, that's different. I'll also make a special exception for layered sounds where some random phase variation between layers can help enhance the "liveliness" of a sound without being detectable as a timing variation.
 
In my experience, the natural variation of note velocity and duration you get from recording a live performance is way more important to the musical feel of a MIDI part than variation in start times. Although I don't, usually, I can hard-quantize anything I've recorded live from a keyboard/controller, and it won't make the part sound nearly as stiff as you might imagine. So long as the synth has a smooth response to velocity with both dynamic and timbre variation, and the durations are not quantized, it will continue to sound quite natural. Conversely, quantizing durations and/o or flattening velocity variation will instantly squash the life out of almost any MIDI track, regardles of how "groovy" the timing is. Of course, the duration element does not apply to one-shot drum samples, but the dynamic and timbral responses to velocity are crucial.
 
Only a blind test with as many possible variables eliminated as possible  - like the ones suggested - can rule out subjective factors in what you're hearing, and identify real differences. This isn't "making it about the person"; this is just how it has to be done when you're trying to answer questions about how "musical" a performance "feels".
 
And just for the record, at least three of the contributors to this thread don't need any MIDI history lessons; we were more or less present for the birth, as it were. 

That's right, we  I still think the Atari had  a better MIDI "feel" . ha! Good post  ++1