Hi Res Texture Pack

[Shioban]

I'm up in the air for those suggesting Tessallation though

Tessellation requires too much processing time to even bother putting it into an MMO. For the likes of a first-person shooter you have a smaller FoV (Field of Vision), allowing you to occlude anything not in the FoV.
Where-as in a 3rd person MMO with hundreds of mesh objects in the scene it's simply a waste.

The detail can be easily acquired from the use of normal/height/bump maps.

[CrAZYVIC]

Tessellation requires too much processing time to even bother putting it into an MMO. For the likes of a first-person shooter you have a smaller FoV (Field of Vision), allowing you to occlude anything not in the FoV.
Where-as in a 3rd person MMO with hundreds of mesh objects in the scene it's simply a waste.

The detail can be easily acquired from the use of normal/height/bump maps.

Depend of your rig, the resolution and the amount of tesselation applied.

The Tesselation option would be "optional" with levels.

Low x16
Medium 32x
High 64x

In my gaming experiencies the tesselation is a game changer. This effect in moderate amounts is always noticeable.

I cant play Crysis 3 or tomb Raider without that.

Yes is a demanding effect but nothing a 780 TI or a R9 290x or CF/SLI cant handle.

[Shioban]

Depend of your rig, the resolution and the amount of tesselation applied.
I cant play Crysis 3 or tomb Raider without that.

Yes is a demanding effect but nothing a 780 TI or a R9 290x or CF/SLI cant handle.

That's the thing, the average user simply doesn't have that kind of hardware, it's actually a very very small percentage of people, it would be considered a huge waste of time at the moment to focus on something only a relatively small amount of people will be able to use.

You'll notice that any game that actually uses tessellation, does so very sparingly.

Even a simple grid when tessellated can almost quadruple the amount of triangles to be rendered, this will always cause a huge performance hit as there's no way to avoid it, you're essentially adding in more triangles to be calculated per frame.

Tessellation is nice to go in and add detail where you simple didn't before. Let's say a cobbled stone road, after developing the area you feel more detail could be added, tessellation+displacement mapping would be perfect if you didn't have the time to go back and adjust the original level layout, however even on a low factor Tessellation can still cause performance and intended implementation issues.

Another big issue is artifacting with low sub divisions, usually forcing developers to use a higher tesselation factor per edge. It's an automated process, where something hand-designed will appear correct to the designer (A brick wall for example, then when a low factor tessellation is applied the algorithm won't understand that "This is meant to be flat, this is meant to be rocky", applying the tessellation at a low sub division to avoid an artefact appearing (A lump/bump where one should not exist) is a nightmare of a task. It's no easy task to go about fixing this issue either.

Tessellation is essentially subdividing and displacing the new triangles (and or not displacing), as with all mesh models the more triangles being rendered the more processing time is required to render the item, this still to this day requires a beefy GPU as you already noted. Until these 'currently beefy' cards are the standard I can't see Square Enix taking the leap to use it.

The ARR team don't seem to want to add anything that the average user can't use (And for the most part Tessellation would be ignored or turned off due to its requirements). So as many people would argue "Well why not add X and make it optional, I bought a $2,000 PC and I want to use it!", the team don't want to seem to add anything optional that only a small handful of people can currently use.

Overall however, if your 3D artists are capable, this should be at all costs avoided as it's an unnecessary "Oh crap this probably needs more topology" addition.

Personally I think the more options the better, but most likely they'd wish to invest their time in something the majority can actually use an enjoy.

[Vasher]

Although tessellation would be nice as an option, this thread is focused on the textures, since Yoshi P stated he has no plans to increase them for DX11.

[Tanama]

I support this proposition for Hi-Res Texture.

Signed and liked.

[Fumijj_Gah]

I support this proposition for Hi-Res Texture.

Signed and liked.

Same. Signed.

[Namasu]

Probably going to need 1000 or more likes to get it taken seriously. Truth is I think yoshiP is right and most of their target audience don't care about textures sadly. :/

You are most likely right that a good number of players doesn't care about textures. They probably didn't know that the textures were downgraded from the 1.0 version. If they know this fact, I'd think most of them would want the hi res option.

I think this is a much needed feature for FF14 if we want to see it age well in the next couple of years. Does this have to be a top priority request for the dev? Most certainly not, but at least acknowledge that it's a legitimate request.

If they come out with this and it keeps happy or gets X people; but in the end they lose money on it or waste valuable resources...<snip>...don't be surprised if they say no or just say "make your own".

I'm highly sure that they still have the original HD texture assets from 1.0 and a little over a year ago, Yoshi boasted that they wrote some awesome conversion tool to push 1.0 assets to the 2.0 engines. This might imply that the 2.0 engine only support low res texture, but that should only be an issue for PS3 with memory limitation. For PC and PS4, I don't see how they could have just throw away the HD textures after the conversion.

At the very least, I think they should re-release the HD textures in increments if they don't have the manpower and resource to do it all at once. Start with HD character models so people can stop making threads about flatbutts. Then move onto gears, then monsters, then background environment. It's not a lot to ask but will please fans for years to come.

[Icehunter]

They are probably looking at this as well:

http://store.steampowered.com/hwsurvey/?platform=pc

There's more than just pleasing a customer that is taken into consideration when developing something. Here are some examples:

Internal/External Customer Satisfaction
Initial Cost
Technical Debt (how much this will cost after it's done to maintain)
Market Desire
Current/Future Roadmaps
Device Support

If they come out with this and it keeps happy or gets X people; but in the end they lose money on it or waste valuable resources. They might reconsider. More than a few hundred million has already been pumped into this game and even more with the upcoming expansions and what-not.

I'm sure they are considering it; but don't be surprised if they say no or just say "make your own".

[Shioban]

I'm highly sure that they still have the original HD texture assets from 1.0 and a little over a year ago, Yoshi boasted that they wrote some awesome conversion tool to push 1.0 assets to the 2.0 engines. This might imply that the 2.0 engine only support low res texture, but that should only be an issue for PS3 with memory limitation. For PC and PS4, I don't see how they could have just throw away the HD textures after the conversion.

At the very least, I think they should re-release the HD textures in increments if they don't have the manpower and resource to do it all at once. Start with HD character models so people can stop making threads about flatbutts. Then move onto gears, then monsters, then background environment. It's not a lot to ask but will please fans for years to come.

The original baked maps for each model (as well as the original Hi-Poly model) still exists, yes.

Some examples of some of the original hi-poly models are here;




The conversion tool they're speaking of it probably based on elements. Where-in you can select a "portion" of an object and then lower the resolution. This would be a very fast way of taking something from it's original detail down as far as it can go whilst retaining a balance of detail and the overall memory impact.

Essentially what you would do is, let's say for this body piece;



There are multiple segments to it, and therefore multiple textures that were applied.
(1.23 had the great idea of taking one base piece of armour and then adding on pieces.)

So Linen>Felt>Wool Bliauds all have the same base model, but with more added ontop of them as the level increases (accessories/extra cloth.

Therefore, the multiple textures across the segments had to be adjusted.

As i've pointed out some, (and you can probably apply this to any gear in-game) is what they would have used this tool for.
"Okay, this segment will be 64x64 pixels, where-as this area willl be 256x256. A very streamline process to spead up the 1.23>2.0 conversion".

It would be time consuming but a possible task to go back and compress them less-so than before, so that not everything is sitting at 4096x4096 textures, but so that there's more detail retained than what we currently have at least.

A great example of how new items in ARR are made and how compression is affecting them badly is the 'un-segmented' gear we're getting.
The Crystal Tower/Titan Weapon/Gear pieces seem to be textured differently now in that thing are in much larger segments meaning the overall compression can be horrific at times.
(This is more apparent with the weapons than anything else.)

Example of bad compression:

Example of fantastic compression:

[Mardel]

Signed.

I'm for it, it is sad to say but a good portion of CT gear looks like crap because all I see are right triangle jaggies in my armor. (Maxed out settings)

SE states they want to provide the best experience and want to push the limitations of devices. Well I'm on a PC and the game tickles my graphics card -.-;