https://www.dropbox.com/s/uj2qqw6a8oofcns/miditest.cwb
 
 
Here's a simple test project I created in Sonar, consisting of a few midi tracks
and one instance of Battery 4 (vst). I rendered the audio of this project within Sonar.
 
I then exported the midi parts, saved the Battery kit, recreated exactly this simple project in
Studio One. I rendered the audio of this cloned project within Studio One.
 
I then loaded both rendered files in Sonar.
 
Both Files are starting at 0'00''.
 
The files are sample-aligned.
Ok, let's try to null those babies... they don't null.
 
I have to pull down Sonar's rendition by .2 or .3 db to get close to some kind of null but without ever reaching it.
 
Even more interesting is that at .1, .2 .3 and .4, different elements of the drums get nulled, but never
the whole drum part. For example at -.2 I get almost a null on the clap, and at -.3 I get almost a null on the kick and snare.
 
Is it timing? is it sound? Is it the way each software handles the mixing of transients that hit at the same time?
is it velocity interpretation by both daw's?
 
Musically speaking now, my gut feeling is shared by many here at the studio. In Sonar's, there's almost to many information, like if the different elements weren't talking to one another... like if every elements was separated from the others. it's a weird feeling, not very musical and almost ''cold''. In S1, it's almost like there's something not SONICALLY but MUSICALLY correct about the way
the drum feels
 
p.s :
If I do the same experience but with an audio part or an audio loop instead of a midipart/softsampler project I get a complete null. 
 
So this issue is only related to the way sonar handles midi playback and softsynths.
 
p.s2:
we've tested with live and logic as well and found similar ''musically correct'' attributes.
 
 
 
 

You're still not going to get much feedback on what you're talking about unless you post links to AUDIO comparing the two. Why not export wav's from both projects and post them on Soundcloud so we can hear? 

There's an aspect to quantizing that I don't think has been mentioned yet...velocity. Groove quantize solves most of my "humanization/mechanical play" issues, but velocity is where the shape of a performance comes into play.
 
I produced an album for a piano player. He sent me MIDI files, and then I "fixed" his performance. My entire job was finding the right piano (I actually settled on TruePianos (full version) Saphhire. Careful EQ shaping and reverb to give it a live feel was one thing...but without shaping the velocity (staying as true as possible to the performance) really made an incredible difference. And the album sold very well...nice Christmas album (the one on the right, here: )
 
It was an important lesson for me: performance isn't JUST about timing. :)

sharke
You're still not going to get much feedback on what you're talking about unless you post links to AUDIO comparing the two. Why not export wav's from both projects and post them on Soundcloud so we can hear? 

mmm... I've done even better and created a cwb file, the link is at the begining of my last reply... I take it you don't see it?
 
here it is again : https://www.dropbox.com/s/uj2qqw6a8oofcns/miditest.cwb
 

aschwabe
There's an aspect to quantizing that I don't think has been mentioned yet...velocity.

See the second paragraph in my post #19:
 
http://forum.cakewalk.com/FindPost/2889615
 
I agree. Velocity is key, and also duration - both more important than start time.

I will add that it also depend on what instruments are being used. Softsynths, Romplers?? Perhaps it SEEMS more musical as those other DAW's are less accurate and have slightly different audio latency responses, hence a "perception" that it is looser. If a MID piece is quantised to 0 beat @ 100% then it is impossible for any DAW to "interpret" it any other way than that, so any perceptions you are getting are most definitely being caused by something else. Assuming all MIDI settings are exactly the same for Sonar and the other DAWs. Have you checked rigorously that this is so, there are no "unseen" settings happing in the other DAW's.
 
MIDI quantising is pure mathematics, there is no interpretation as in the real world of maths.

Kewl Hendagang
https://www.dropbox.com/s/uj2qqw6a8oofcns/miditest.cwb
 
 
Here's a simple test project I created in Sonar, consisting of a few midi tracks
and one instance of Battery 4 (vst). I rendered the audio of this project within Sonar.
 
I then exported the midi parts, saved the Battery kit, recreated exactly this simple project in
Studio One. I rendered the audio of this cloned project within Studio One.
 
I then loaded both rendered files in Sonar.
 
Both Files are starting at 0'00''.
 
The files are sample-aligned.
Ok, let's try to null those babies... they don't null.
 
I have to pull down Sonar's rendition by .2 or .3 db to get close to some kind of null but without ever reaching it.
 
Even more interesting is that at .1, .2 .3 and .4, different elements of the drums get nulled, but never
the whole drum part. For example at -.2 I get almost a null on the clap, and at -.3 I get almost a null on the kick and snare.
 
Is it timing? is it sound? Is it the way each software handles the mixing of transients that hit at the same time?
is it velocity interpretation by both daw's?
 
Musically speaking now, my gut feeling is shared by many here at the studio. In Sonar's, there's almost to many information, like if the different elements weren't talking to one another... like if every elements was separated from the others. it's a weird feeling, not very musical and almost ''cold''. In S1, it's almost like there's something not SONICALLY but MUSICALLY correct about the way
the drum feels
 
p.s :
If I do the same experience but with an audio part or an audio loop instead of a midipart/softsampler project I get a complete null. 
 
So this issue is only related to the way sonar handles midi playback and softsynths.
 
p.s2:
we've tested with live and logic as well and found similar ''musically correct'' attributes.
 

 
I checked this out. There are amplitude rendering differences that might be attributable to SONAR's Double Precision 64-bit mix engine or other mix engine differences. But the timing of transients seems pretty much identical except for the two clap transients that differ by 1 sample, and do not null well as a result. But they are both well off the nearest tick in the timeline so it's impossible to say one is more accurate than the other. I find this a little weird, because SONAR will render events precisely on the tick in my environment unless the synth itself has wonky rendering response as some do (TTS-1 for example).
 
I would be interested in understanding the amplitude differences better, and seeing the different parts rendered separately to investigate phase differences, but I see no cause to conclude that one is more accurate or more "musical" than the other at this point. The timing is not different enough for a human to detect outside of the nulling context, and the amplitude variations are no more than you would get with a live drummer hitting two drums with slightly different velocity and phase from one measure to the next.
 
EDIT: One critical thing that needs to be checked before anything else is whether either SONAR or Studio One will render the output the same way twice. As mentioned earlier, many synths randomize samples or have unsynced FX that won't render the same way twice in succession. I just played one track by itself, and noticed that the track has a lot of reverb on it. The first thing I would do is turn off any FX that BFD is adding. I don't have BFD so I can't check it myself, but you need to verify the it will render the same output every time. I can tell you, for example, that Superior Drummer will not.

brundlefly
.....but I see no cause to conclude that one is more accurate or more "musical" than the other at this point. The timing is not different enough for a human to detect outside of the nulling context,....

I observed the same.
 
I have a method that I use to blind test something.
 
Set the solo button on two tracks to the same group (red), but opposite value. When one is on, the other off. Place the mouse on one of the Solo buttons. Get in a position that you can click rapidly and not move the mouse off the solo button. Now listen to what you want while clicking the Solo button. At some point, rapidly click the mouse button while paying attention to what your listening too. You will see that you can loose which one is really playing. And thus, a blind test of sorts!

Grem
Now listen to what you want while clicking the Solo button. At some point, rapidly click the mouse button while paying attention to what your listening too. You will see that you can llose which one is really playing. And thus, a blind test of sorts!

I routinely do my own "blind" testing in this way, too. Works very well. You'll only be able to consistently identify which is which if there really is an audible difference. In this case, the amplitude differences could conceivably sway someone's preference, but I doubt it.
 
BTW, I found that normalizing the two tracks to full scale was the quickest way to roughly match the amplitudes.

brundlefly
 
EDIT: One critical thing that needs to be checked before anything else is whether either SONAR or Studio One will render the output the same way twice. As mentioned earlier, many synths randomize samples or have unsynced FX that won't render the same way twice in succession. I just played one track by itself, and noticed that the track has a lot of reverb on it. The first thing I would do is turn off any FX that BFD is adding. I don't have BFD so I can't check it myself, but you need to verify the it will render the same output every time. I can tell you, for example, that Superior Drummer will not.

Very good point.
 
There is no doubt that something is happening after listening to the test. That's for sure. But I still can't "sense" that one track is more "musical" than the other. And that's his whole point.
 
 And in my opinion, in regards to Sonar not playing Midi info "musically," is something perhaps, "subjective."