Tachikoma said:
Nuvendil said:
I would just say this part is what usually gets me because alot of the "downgrade" issues people reference were present from the start. Mech's have never been super clean in terms of textures and polygons. You can see the modest textures and polygon issues as early as the 2013 trailer and the interior textures being low res or suffering from heavily delayed popin in issue we saw quite plainly in the 2014 trailer. Character model polygons being high or low relative to the importance of said characters is also something we've seen frequently throughout footage, though this didn't become nearly as apparent as it is now until recent vids that zoomed in more. Now some things have changed for sure, like the LOD models from the early footage to now have seen optimization of polygons in favor of more distinguishable texture detail, as you pointed out before. And some changes are not actually changes at all, but assumptions. Some have said the character models have seen downgrades, but when you compare characters we have seen multiple times across multiple videos, detail is the same. And the armor we saw from 2013 is still present in the same level of detail. This isn't to say there aren't any changes at all, the game *is* in development. But I would classify on the whole the changes as optimization as opposed to a Watch Dogs-esque "downgrading." In other words, conscessions made for performance that strive to maintain the visual impact as much as possible as opposed to rampant degredation of detail on all fronts with little regard to effects to the visual experience.
|
What i bring to the table with this is a brief tenure working on the game myself, while with Nintendo EAD, o while I will be referencing earlier builds/footage, the tech analysis isn't going to be a black and white everything-depends-on-the-difference analysis.
More a broad look into the design choices, benefits, downsides and limitations met while designing the game.
|
Sounds good. That's actually the way it should be done because counting polygons and texture pixels really isn't helpful. All games are downgraded along the way from a letter-of-the-definition perspective; that's a major part of optimization. The question should be how effectively the optimizations preserve the visual integrity of the experience. If they reduced one texture set by 75% but you never see it clear enough to notice, then it was not nearly as bad as reducing a texture by 50% that is constantly in your face and noticeable. To make a rather basic comparison.
Also, one design choice I don't agree with right off is the resistance to any kind of prerendering of cutscenes. I don't mean ridiculous high-detail models, but perhaps the high-poly bases for the in game models with higher textures, just to give the scenes a cleaner look. As it stands, you go into cutsenes and often have your face pressed up against textures that were clearly never meant to be seen at such close distances, the dolls being the biggest offender.
