Even if 32MB of esram provided more bandiwdth(not saying it does) than the bandwidth provided by 6GB of gddr5, besides bandwidth you need render power, ps4 gpu seems to be about 50% more powerful than the xbox one, and kind of makes sense that many xbox one games are 900p wich is about 50% less than the 1080p. The wii u edram has enoug bandwidth for the 1080p with double buffering with just 16MB but that doenst mean the gpu can keep up with the framebuffer and shader power for graphics also must be accounted.

As for the 4k-8k textures, they havent mentioned anything about them recently, but if their were not going to sue them anymore then they would have tell us by now.  Its only a bet right now, but i think they could have broken down the etxtures into tiny tiles so they could fit them in texture memory and updating them in texture memory while the rest would wait on system ram, the technqiue is known as texture paging, the reason i beleive this is that the terrain in the first image looks a bit tiled

https://www.opengl.org/archives/resources/code/samples/advanced/advanced97/notes/node29.html

@

Paging Textures

Imagine trying to draw an object which is covered by a portion of an arbitrary large 2D texture. This type of problem typically occurs when rendering terrain or attempting to pan over very large images. If the texture is arbitrarily large it will not entirely fit into texture memory unless it is dramatically reduced in size. Rather than suffer the degradation in image quality by using the smaller texture, it might be possible to only use the subregion of the texture that is currently visible. This is somewhat similar to the texture tiling problem discussed earlier, but rather than sequence through all of the tiles for each frame only the set of tiles necessary to draw the image need to be used [41].

There are two different approaches that could be used to address the problem. The first is to subdivide the texture into fixed sized tiles and selectively draw the portion of the geometry that intersects each tile as that tile is loaded. As discussed previously, this is difficult for GL_LINEAR filters since the locations where the geometry crosses tile boundaries need to be resampled properly. The problem could be addressed by clipping the geometry so that the texture coordinates are kept within the [0.0, 1.0] range and then use borders to handle the edges, or a single large texture consisting of all of the tiles could be used and the clipping step could be avoided.

@

File size is a consideration here; the whole game is less than 1GB.

You're also not really going to see this texture tiling while playing as the camera angle will be lower.

Apperently so:

@MaxWill37699872 Yeah, streets took a lot of resolution. Tho biggest textures are used for the shadows.

— Shin'en Multimedia (@ShinenGames) September 24, 2014

Consoles aren't THAT bad ... 

At least with the PS4 and X1, both are capable of doing quarter decent Global illumination such as radiosity and cascaded voxel cone tracing plus with a little more work I bet we can get realtime photon mapping on the PS4 too compared to being stuck with awful SSAO from xbox 360, PS3, and WII U.

On the brightside high end PC gamers can now enjoy the increased baseline for better graphical fidelity ... 

thans, but was not refering to the 3d scans jusy the 10MB textures shinen mentioned they applied BC1 comrpession, since 10MB texture still cannot fit in neither texture memory nor edram is possible that they broke down the etxtures into tiles to fit them in the texture memory and the technique is known as Paing Textures

i will leave the 3d scans for later once i get more info

for Pemalite

"

No offense Megafenix, but for the love of god, stop posting ancient, irrelevant crap.

ATI's "Smartshader" doesn't apply to GPU's today, that crap was released 13 years ago and ended 7 years ago, why you would think that Direct X 9 shader operations apply to a Direct X 11 world beats me, AMD doesn't even use VLIW anymore, let alone seperate Vertex and Pixel shader entities in the hardware.

As for your "6 textures in a single pass". - That's called single pass multi-texturing, that technology has been around for decades, shall I school you on it? Hint: It has nothing to do with the multi-pass that everyone else is talking about.

Another issue in what you copy/pasted is you claim the Playstation 3 has "8" SPU's and only 7 are available for games, well... No. The Playstation 3 has 8, one is disabled in order to increase yields and another is reserved for other tasks, that makes 6.

"

No thanks, i appreciate the offer but i have known it for quite a while. actually have known it since the gamecube era, dont uinderstand what exactly want to teacxh me besides the name of the technique that i didnt bother to mention. And well, eventhough deffered rendering requires multipass, its only 2 passes while with forward rendering could require a lot more depending on the complexity of material and light combinations, and shader power suage is also a concern., The concept is simple, the less passes you use the less strain you put on the pipeline

They can just store them in main RAM. 10MB isn't going to be a burden on 1GB of memory.

Yea, i am sure that storing 10MB textures in RAM of 1GB is easy, but since they cannot fit the textures in the texture memory directly they would lose a lot performance, thats why breaking down the high resolution textures into tiles to fit them in texture memory would be ideal, and since the first image has a terrain that looks tiled i suspect they have used  the Paging Textures technique

ftp://ftp.sgi.com/opengl/contrib/blythe/advanced99/notes/node68.html

"

Paging Textures

As applications simulate higher levels of realism, the amount of texture memory they require can increase dramatically. Texture memory is a limited, expensive resource, so loading high resolution images as textures is not always feasible. Applications are often forced to resample their images at a lower resolution to make them fit in texture memory, with a corresponding loss of realism and image quality. If an application must view the entire textured image at high resolution, there may be no alternative to this approach.

But many applications have texture requirements that can be structured so that only a small area of large texture has to be shown at full resolution. For example when textures are used to produce a realistic flight simulation environment, only the textured terrain close to the viewer has to show fine detail; terrain far from the viewer is textured using low resolution texture levels, since a pixel corresponding to these areas covers many texels at once. For many applications that use large texture maps, the maximum amount of texture memory in use for any given viewpoint is bounded.

Applications can take advantage of this phenomena through texture paging. Rather than loading complete levels of a large image, only the portion of the image closest to the viewer is kept in texture memory. The rest of the image is stored in system memory, or on disk. As the viewer moves, the contents of texture memory are updated to keep the closest portion of the image loaded.

There are two different approaches that could be used to address the problem. The first is to subdivide the texture image into fixed sized tiles and selectively draw the geometry that corresponds to each image tile, one at a time, reloading texture memory for each new tile. This approach is difficult to implement. Tile boundaries are a problem for GL_LINEAR filters since the locations where the geometry crosses tile boundaries need to be resampled properly.

"

Is just a thought so I will not go to further on that, maybe i could somehow get the answer from them, hope they are in the mood to answer this question though,hehe

What "texture memory"? Wii U doesn't have a dedicated bank of RAM for textures.

txture memory can be the texture caches on the GPU(L1 and L2 texture caches) or maybe part of the edram used as a cache(in gamecube they used 1MB os embedded 1tsram for texture cache). All modern gpus(lets start from hd400 to newer ones) have texture caches with high amount of banwidth but with limited memory storage made of sram pools

...no offense mate, but talking to you is giving me a headache.