@BlackKnight

It was DigitalFoundry that originally made the comparison of MLAA to 16xMSAA. Folks picked that up and ran with it.

The truth is somewhere inbetween, and it largely depends on the internal resolution of the game. If you are gaming at 1080p 4xMSAA will probably eliminate aliased edges every bit as good as MLAA would. At 720p that isnt the case IMO.

You are correct that the items you described Will look better with MSAA. However, if textures are being affected then its due to a poor implementation by ATI. PS3 games that use the technique don't suffer from that issue.

Bottom line, if you are gaming on a PC you should probably always choose MSAA, unless you have an underpowered card and are gaming at 720p. MLAA is a good alternative for consoles where GPU performance is limited.

MLAA is not useful in applications where traditional msaa is an option. These titles use a forward rendering approach which allows for AA to be applied during the rendering process. Most games use this method, such as new Vegas and mass effect 2.

However, games that use a deferred rendering approach don't really allow for traditional msaa and thus they don't support it. This includes games like dead space, crisis 2 and gta 4. Thus, mlaa is available to use on these games as a post processing effect, which can have results equivalent to 8xaa with the performance hit between 2x and 4x.

Using mlaa in forward renderers has mixed results, but mostly the effect isn't as clean an performance is far less than desirable. If msaa is available, use it. If not, mlaa is a perfectly good alternative. It looks great in dead space!

@Arksine

Yea, like I said, you can't say MLAA= *xMSAA. It just doesn't work.

Would MSAA be better at 720P (lower resolutions) because it is actually grabbing extra samples that aren't present in the untouched 720P image? And I mean in the case of something like the fence or anything that has VERY little pixel data in the original image. I agree that MLAA would probably do better with general edges than 4xMSAA at lower resolutions (building silhouette against a sky).

As for the texturing, how do we know whether or not MLAA affects the textures in PS3 games when we can't turn it off for comparison?

The next Xbox issaid to have 36 GB of memory so the price short from ram should not be as apocalyptic as a 64 GB kit. With the PS6 coming in with 30 GB, the RAM should not be what makes the Xbox cost so much more. Of course without Microsoft subsidizing the console the actual MSRPs may diverge wildly.

32GB of DDR5 still costs in the neighborhood of $250-300 for the super cheap stuff, $450 for the name brand. That's what entire consoles used to cost. That eats up a huge chunk of the budget that was supposed to be paying for the CPU and GPU, which means that the cost of this system will be driven farther north than previously anticipated, and it was already expected to be above $1k. Releasing a console in the middle of an industry-wide RAM shortage is stupid. Even GPU makers are scaling back production because of it, and focusing their remaining stock and production towards selling to datacenters. Some memory manufacturers have dropped consumer products entirely and now only make chips for datacenters. Nvidia is scaling back its consumer GPUs, no longer offering the super series of some GPUs, for instance.

For MS to pull the trigger now means releasing at a very risky price point against a PS5 that is simply on fire, even outselling the Switch 2 in many cases. It's coming at a time where the Xbox brand is at its weakest ever, and consumer confidence in the brand is at absolute rock bottom. Nobody wants to drop $1500 on an Xbox when they can play the same games on their PS5 Pro for half the price already, or even cheaper if using a base PS5. Only a select few enthusiasts will bother to fork out that kind of money... by the time this product reaches a price point where it can have mass-market appeal, the PS6 will be dropping... but by that time, RAM prices will be dropping, so PS6 will now be able to, assuming it does delay until 2029, invest more into upgrading its architecture over the previously released spec, invest in more RAM than the new Xbox will have, a better CPU & GPU, etc.

As for nobody saying PS6 will launch in 2029, nobody said it would come sooner either, not officially, at least. As of now all we have to go by are rumors based on internal information that could easily change at a moment's notice. Even the design of the chip itself could change as it has not yet entered into production. They could easily opt to include a few more CUs, more RAM, more CPU cores, etc. between now and when it does officially enter production. So, MS could drop a new Xbox now, but it wouldn't be wise, at all, for them to do so if they plan on even holding a candle agains the juggernaut that will be PS6. PS5 will most likely mop the floor with it due to its price point alone.

And that's assuming MS even gives the green light to start manufacturing the console to begin with. We'll see in the coming months if production even happens. Microsoft's shareholders damn sure aren't going to be willing to subsidize anything at all after they just dumped $100b into buying game publishers, expecting to see a ROI, and not seeing it anywhere near as fast as they'd hoped, which is why we're now playing Xbox games on PS5.

As for MS sitting on RAM, they are sitting on some, but Xbox is sitting on none. Microsoft knows good and well they will make far more money putting that RAM into datacenters than they ever would putting it into a console that is already at a huge disadvantage before it even launches, and has little hope of generating a lot of sales.