Topic on Talk:Mass Effect Legendary Edition

I have personally not noticed this. I have had certain sections of ME1 drop to mid 40 FPS, if it was truly double buffered, it would drop to 30. And then the source for this is merely an assumption, and then provides an incorrect "fix" as a workaround.

Vote to change the citation to "citation needed".

The way to question references is by using the {{dubious|reason=}} template. Glancing at this particular reference, it seems to be reliant on one post. One comment vaguely confirms the report but it's up for interpretation. If it were just one, we could just remove the reference and the claim without thinking, as per the References guidelines.

It's complicated, but basically the game makes use of two back buffers (this is a requirement for flip model which the game makes use of). On top of this I believe in some scenarios there might also be a third front buffer.

My current hypothesis is that this depends on what presentation mode the OS is able to grant the game, but I need more info from Kaldaien to say for sure.

Basically flip models requires two back buffers, but as far as I understand it when engaged in the best optimization mode where the DWM of Windows steps back there wouldn't necessarily be a third front buffer -- a direct scanout would occur straight from one of the two back buffers of the game.

This would, I guess, result in the noticeable double buffered v-sync behavior where frame rate drops from (Refresh Rate / 1) to (Refresh Rate / 2).

However, there is also composed situations to consider... If the DWM is involved, and a direct scanout cannot be achieved from the back buffers, a third front buffer would be involved where the contents of one of the back buffers would be composed along with contents from other buffers on the system.

That would, I once more guess, set up a scenario where the FPS can render unrestricted in the two back buffers without dropping down to (Refresh Rate / 2) when the FPS drops below (Refresh Rate).

So what we're seeing here might simply be the result of different presentation models engaged in Windows -- the guy who experienced the double buffered v-sync behavior might have Windows engage a more optimized presentation model than those that do not experience that behavior.

Aren't games complicated?! :D

This is just my current hypothesis that would explain everything -- I haven't actually confirmed it myself yet. I've also pinged Kaldaien on the Special K Discord server to check with him if this sounds like a reasonable assumption/explanation.