M3GA2 Added Step: Would This Work?

[Neil]

Hello everyone!

I have an idea to run by you. However, I'm not sure if it's going to create too much work for our members to submit games, or if it's a natural thing to add.

M3GA is designed to really put a game through its 3D paces by encouraging users to adjust the convergence controls. You often need to do this to spot anomalies and problems that can easily be overlooked and take away from the visual fun of the game.

In Part II, M3GA asks for a list of eye candy settings that need to be turned off to reduce or remove anomalies. It works best when the user is using convergence settings to really test things out. However, if a game is able to play with depth-only and keep these eye candy settings on, I'm wondering if M3GA should take that into account.

After users select the required game setting reductions, would it be too much work to ask them if these eye candy reductions are needed in a depth-only situation?

For example, let's say you have to turn HDR and shadows off. At the bottom of the screen, M3GA could ask you, of the setting reductions listed (M3GA would only list the settings you picked earlier), which can remain active in a depth-only situation? Would this extra step be a natural addition for M3GA users, or just cause extra headaches?

Regards,
Neil

[BlackShark]

I have had the exact opposite situation in some games where stuff would work correctly in pop-out situation and break up in depth (fog effects in UT3, textures being clipped in Mirror's edge,...). Or in other variations of the same issue : having lots of bugs with a particular elements proportional to the separation, but if you put this particular element at screen depth, the issue is no longer visible (you have to use a particular convergence setting that is neither all in or all out).

I think the convergence issue should be treated separately from graphics bugs otherwise there will just be too many situations to count and you'll end up with too many questions or too much information required to fill up that will discourage users to fill in reports.
I would suggest you add a section dedicated for convergence issues which would allow stating there are graphics glitches but without going into details.
Here is what I'd see in the convergence reports :

Does the game provide you with the following controls over the 3D effect :
- Separation (adjusts the strength of the 3D effect) y/n
- Convergence (adjusts the balance between in-screen and out-of-screen effects) y/n

What convergence setting do you use ?
-all inside
-a lot inside with a little out
-balanced
-a little inside with a lot out
-all outside

Do you like this 3D setting ?
-yes
-no : the game does not allow tweaking or not enough to fit my favourite 3D setting
-no : bugs appear when tweaking the 3D settings

[Neil]

For me, I'm thinking of problems like lights getting disconnected from light sources, textures sliding off objects, or HUDs that split in two. Really obvious problems that shouldn't be there in any case. Thinking about it some more, maybe the solution is to keep it as is because the games that are convergence locked won't get the errors, and the games that aren't convergence locked will have the means to test for the errors that should be corrected in any case so gamers can get the precise 3D experience they are looking for. I think it will come out in the wash as it has to date.

Separate from this, I think your convergence idea can have a place under the subjective section of the process - we can try it.

Regards,
Neil

[cybereality]

First off all, you have to recognize that all the drivers do different things and have different naming conventions. Most notably, what IZ3D calls "convergence" is what Nvidia calls "depth" and what iz3d calls "separation" is what Nvidia calls "convergence". DDD has the more accurate naming, with "depth" being the range of values and "convergence" being the focal point. But DDD offers even more adjustments, which all inter-effect each other. It is usually the separation between the virtual cameras that causes issues, which equates more with having aggressive settings more than what the particular vendor may call it. Its not a simple matter of the thing called "convergence" make stuff pop out the screen.

Honestly, I don't think there needs to be any mention of convergence, out-of-screen effects, or anything else like that. The users should use the settings *that are right for them* and report what anomalies they see. If a particular game has blatant issues (for example, shadows only showing on one eye) then most users will probably see it. If most users don't see it, *with the settings they would choose*, then the problem is thus less important. You can't tell users to use some jacked-up un-natural settings simply to highlight bugs in the driver. The grade should be based on real-world settings, not some arbitrary idea of what a 3D game is supposed to look like. With enough entries into the database, it will all even out. Some users may go for super high settings, others are more conservative. And you can see anomalies in either situation. Its not that anomalies only happen with one set of settings as BlackShark noted.

Really, I agree totally with Tom's Hardware. There should be no mention or questions relating out-of-screen effects. It should not have a bearing on whether a game is graded well. It should only be about the bugs seen and whether turning off settings fixes them. I understand most gamers do enjoy out-of-screen effects, but enjoyment is a subjective experience, and does not have a place in an objective scoring system. And beyond that, even masterpieces like Avatar (the movie) barely had any out-of-screen effects AT ALL and look how that was received. Its not all about those gimmick shots that are probably why people hate 3D.

[Neil]

Hi Cybereality,

I think the first step is to ensure that the naming conventions for M3GA are determined by MTBS. Separation and Convergence are pretty standard, and have consistently been described in our S-3D settings guide. The new M3GA version will include sample videos and pictures so gamers will know what to try when doing their testing.

However, M3GA was never intended to define what a good 3D game is for everyone. There is a QA component to make sure there is a certain amount of error-free compatibility, but the best choice 3D settings should best be left to the individual gamer.

Instead, M3GA was designed to give an accurate measure of what visual flexibility can be achieved with games, and what if any trade-offs are needed to attain that flexibility. So if I get a Platinum graded game, I should be able to have out of screen effects to my heart's content without being worried about turning down settings or things sliding off the screen. A Bronze graded game would limit me to depth only, and I may or may not have anomalies to contend with. I may love my Bronze game - but I may want Platinum level flexibility if I want it.

That said, I think we are faced with a new opportunity. We could have a score that indicates the overall QA value, and a secondary score or marker that indicates the flexibility offered in the game. The current grading may be throwing people off with all the colors or what people think the colors represent. Fortunately, the data is there, we just have to present it in a meaningful way. If we can accomplish this, games would not get dinged for being credible depth-only experiences, and gamers would still know what 3D flexibility they can expect. I think this would be a fair mix. Let me think about it and ask around.

Regards,
Neil