Melodyne Efficiency in Sonar vs. Other DAWS? (Comparison stats inside)

Quick Summary: A comparison of Sonar and Studio One showed a massive speed difference opening clips in Melodyne (Sonar much slower).  
 
Quick background:  I used Cakewalk/Sonar for many years but switched to Studio One for a while and then returned for X3 and am up-to-date with 2017.10.  Also update to date with Melodyne for this comparison and same for S1 (which I keep for the mastering page and CD burning mostly)
 
I was using S1 when it announced the first pairing of Melodyne and ARA, and though I didn't use it a lot I had some familiarity with the speed of opening clips in Melodyne when I switched back to Sonar.  Over the past year-plus I got the impression Melodyne was slower in Sonar but didn't take the time to check until last night.
 
This comparison used a 7MB, five-minute mp3 file (192kbps) of a commercial track (I'll skip the details for now but in short I wanted to create a live-feel tempo map, which is a really cool thing to do for use with Superior Drummer); when I had tried using a WAV file in Sonar I gave up after 27 minutes of the machine chugging away and the progress circle was barely past halfway.  As my specs show, this is a pretty powerful machine, and if there is a setting somewhere in my machine I'd love to get to the bottom of it, but the S1 stats show that might not be it.
 
Anyway, I converted the WAV to mp3 and used Melodyne in both Polyphonic Sustain and Polyphonic Decay mode.  Here are the average comparison times from multiple tests:
 
Sonar:   Sustain:  Approximately 18:30 to open a five-minute file
             Decay:    Stopped after 19:00 minutes both times
 
S1:        Sustain:  3:00
             Decay:    4:10
 
These figures match the perception that Sonar has always seemed slower at creating Melodyne clips but I'm interested to hear what others have experienced.
 
For what it's worth, Melodyne reached the halfway point in its progress circle rather quickly most of the tests, but then really delayed during the second half, particularly the last quarter.
 
All insights greatly appreciated,
 
Steve

I've never used Melodyne with MP3 (didn't know you could). I've also never used S1, so I can't compare that. But it only takes a minute and fifteen seconds to render a five minute and twenty -four second clip (mono clean electric guitar or stereo mixdown, both took the same time) using either polyphonic mode. I used to think that was taking to long.
 
I am using Melodyne Studio in SPlat 2017/10.

SONAR doesn't read MP3 files directly, it imports and converts everything to WAV files and uses only those.  So I'm not really sure what exactly your MP3 to WAV process actually was.  Perhaps you could give a step by step for both SONAR and Studio One.  Also I'd be interested to know how long the stand alone version of Melodyne takes to do the same operation.  I'd think that is the better comparison.

After 50%, Melodyne start actively write "separations" files.
 
It seems like it always does that, even when RAM is available.
 
The locations of these files defined in Melodyne preferences, "Recording". So make sure:
a) the folder is on fast storage
b) if you have an antivirus or other real-time scanner, exclude that folder and the process from scanning (in this case Sonar).
 
On my system (see signature), there is a huge difference in Melodyne timing between with and without (b).

I work all 4 weeks with Melodyne. Sometimes StandAlone, sometimes as VST-plugin, and I have also testet the ARA integration. But I see no difference in timing between StandAlone, VST or ARA. I use Melodyne Studio. The computer has 3 SSD drives. Opening 5 minutes of a vocal track (polyphon) needs about 2 minutes if its ready to work with.

Thanks to everyone who responded.  I've been away all day and have the following answers to questions:
 
1.  Re: system specs, I'd have to look but it's pretty fast CPU, 16GB RAM
2.  MP3 to WAV conversion occurs very fast when importing
2a.  The first time I tried a straight WAV file it seemed Sonar had locked up since it had been 27 minutes with not even 75% conversion, but the computer light was running; the second time the WAV file took the same amount of time as using the mp3 conversion.
3.  The process in both Sonar and S1 is to either right-click in the clip and choose Melodyne, or CTL+M (no difference in speed of Melodyne)
4.  The test file was on the C Drive (Desktop) of a 7200 RPM high-quality drive
5.  Antivirus is on the machine (we 99% of the time is not connected to the 'net) -- azslow can you explain a little better how to exclude it for the purposes of using Melodyne, and I can then compare again?
 
Thanks again everyone.  
 
Steve

If I compare Melodyne integration  in Sonar and in S1, I get the impression the approach the S1 guys took is not as tight. To enable this higher level of integration, the bakers may have needed to take steps which resulted in more processing in the background.
 
IIRC, Sonar would reliably crash if a clip hadn't been bounced (is this fixed?), whereas with the S1 approach, this concept of pre-bounced clips simply did not exist, meaning the issue never arose.
 
Edit: The audio quality of pre-bounced clips, e.g. with Audiosnap, is not as good as when bounced. This was however a great feature back when PCs were a lot less powerful. S1 is newer and probably the need for such an intermediate step never arose.

I've used melodyne a lot recently because I'm trying to really learn the fine details of things it can do. I've used it stand alone and integrated with sonar with single tracks, multi tracks, with all algorithms, and done just about everything you can do with it.
 
I have not noticed any difference in speed stand alone vs in sonar. I have a quad core I7-4810 processor, crucial m200 SSD, 32 gig ram, win 10 with all the latest updates. I have done the simple tests similar to above, doing tasks and timing the outcomes. Doing the same things stand alone vs sonar the results are always withing a few seconds of each other. 
 
I do not have S1, so all my work is stand alone vs sonar. 
 
gabo

I have made typos in my reply, and after editing the forum software block it...
Here is the text I have tried to post: http://www.azslow.com/index.php/topic,392.0.html