The issue first showed in the post / video / link that you made dyopp21 (http://www.vpforums....ing#entry203203) was a somewhat different problem than flipper tearing as it was more confined to the flasher overhang and not as random as flipper tearing. It does look like what ball stutter could look like in slow motion or a "morphing" for other graphic drawing anomalies. However, you are right that flipper tearing and this ball issue, whether exactly the same as flipper tearing or not, can be fixed by allowing Aero / not disabling Desktop Composition on the VP executable. This is why I started to use Windows 7 a long time ago because Windows XP does not have this feature. I wrote about this aspect several times in different topics over the last few years - you didn't by chance try searching on "flipper tearing" and did you 
I know this started with concerns for VPX and that historical posts may not have seem relevant but as the discussion got more into VP 9x and 9PhysMOD that would have yielded you the same result in general at least about the specific flipper tearing. This is a link to one of the times I raised the topic and the resolution and was back from VP9x before any PhysMOD or VPX which meant for any video syncing option to work you had to use Aero / DWM because the internal syncing options still tied to physics frame rate and would cause problems with physics and odd flipper shots if utilized alone:
http://www.vpforums....ing#entry203203
Now, the thing that's changed is that since VP9 PhysMOD and a working vsync option, the need for having to allow Aero to remain on / not Disable Desktop Composition should have been eliminated in theory as the screen syncing capability of DWM / Aero / Desktop Composition should not have been required with software based vsync if functioning fully. However, from the more recent testing I've done, no method of disabling Desktop Composition with any combination of VP vsync or other parameters can yield a stutter-free result. There are a number of inputs I would have that may or may not be relevant on this issue and too much to get into here / at the moment but I would like to discuss the topic still further for a couple reasons. I've been too busy with other aspects of VPX testing / QA to dive more in or raise the somewhat difficult to explain nuances of the issue yet with fuzzel or Toxie (I do have a test table and some setup specifics that may help make it viewable and easier to troubleshoot and I hope to continue with them in this regard). One thing seems to be that the non true full screen could have a bearing on the change in behaviour or inability to use the software based vsync on it's own. Also, the variety of results with different video settings just seem to be about the more or less demanding the options the quicker or easier the problem is produced. I've seen test screens of 640x480 without the issue but I'm killing the FPS if I were to disable the sync and the massive margin for operation between what can be drawn and what is seems to play a role at times as well and make troubleshooting difficult and misleading.
Lastly and most importantly, the reason that it is desirable to get actual syncing and stutter free working without allowing Desktop Composition / Aero is that there is a noticeable improvement in flipper lag and responsiveness for shot timing. This has been quite useful with the somewhat weaker backhand / close to base flipper shots that have been a slight issue with PhysMOD or VPX physics especially if trying to keep the flippers power reasonable and also utilize the coil-up for more realistic flipper behaviour (in general / in other regards).
So at this point, there's a choice for everyone, use Aero / DWM / Desktop Composition and sacrifice some input lag but have totally smooth play or have faster flipper response / more precise shooting but some degree of tearing and / or ball stutter (the FPS between the two options nowadays is essentially equal on the systems I've tested so not really part of the decision). The best result for the least amount of moments of ball stutter with Desktop Composition / Aero disabled is by setting the FPS to your screen refresh rate manually (I've used 60 and 120 as I have a 120 hz screen for my mini / HV pincab). If trying to use the "1" to allow vsync but with Desktop Composition disabled, you will get terrible stutter and when trying to look at the FPS it seems that what it reads out is a likely cause why as it's struggling at frame rates much less than what the monitor is at (28 - low 30s and such for a regular monitor and 80s - 90s for 120). Basically, it is not syncing in that mode. Like I say though, setting a fixed rate does not solve the problem and it will look pretty good and smooth but there will be moments and still areas that seem on the screen to be consistent where the ball morphs / skips / stretches.
Sorry for the long post but this issue and all the nuances around it make it difficult to provide anything meaningful without the allowance of verbosity.
EDIT: One thing to correct is that the odd FPS numbers in the 2nd last paragraph above when vsync is used / set to 1 within VP was not necessarily when using standard resolutions or refresh rates and is not as simple to reproduce at 1080p at 60z as with trying 1024x768 at 75 or a 1440p resolution then trying with Desktop Composition Disabled and VP vsync set to 1. There is also a little further complication in that sometimes switching resolutions / refresh rates and Windows and VP seem to get out of sync (I think I've seen Windows itself get out of sync though as well judging by the mouse pointer) and this complicates again the scenario. However, a fresh reboot normally solves this for using Aero / Desktop Composition but does not if trying say 1440p or 1080p / 720p at 120hz and the built-in vsync
Edited by jimmyfingers, 27 October 2015 - 11:49 PM.
Top
Gotta try this asap 
are all trademarks of VPFORUMS.