Environment mapping is a way to make fake reflections. You can make natural reflections with raytracing. The problem is that must exist something to be reflected. A dull room with nothing will not work. You need objects in the room, extended area light sources or windows acting as a light source.

Part II

With metals and conductors the subject complicates. Conductors have a not zero conductivity and a perfect conductor has an infinite conductivity. An object with infinite conductivity or perfect conductor doesn't allow light to enter the object, so 100% of the light is reflected.

With metals the conductivity is very high, but not infinite, unless it is a superconductor, so light is allowed a little to enter the object and what is relected is less of 100%. The light that enters the conductor is very quickly absorbed.

The refraction index now has no sense because there is no refraction, but we can still use Frsnel equations to know how much light is reflected, but in this case the refraction index turns into a complex number because it depend on the dielectric constant and conductivity too.

I our case of making images to make thing easy we can use a fake refraction index, as metal almost reflect all incoming light we can use a big number for refraction index. The value of this number we can find empirically the one that produce the images which we want.

Until now we have ignored the magnetic permeability, that is very near one in many cases. But we with magnetic materials we cannot ignore it and so, we must forget Snell's law, refraction index and Fresnel equations and use instead the characteristic impendance of the medium and its related equations.

For transparent metals things are even more complicated. Conductors don't allow light to penetrate, but if the layer of metal is thin enough the shielding effect of the conductor decrease and some light is allowed to pass throgh. More thiner the layer more light pass and we have a transparent conductor.

The conductivity also depend on the wavelenght, the other two parameters also do. A very thin layer on our microwave window allows us to see inside the microwave owen, but this same thin metallic layer is opaque to microwave wavelenght reflecting all the microwaves and don't allowing to reach us.

When light hits an object a part of light is reflected and a part enter the object. The part that is reflect is reflected and with the part that enter the object can happen two things: Or it goes nowhere and is absorbed by the object or it pass through the object, in this case we have a transparent object. When light pass through a transparent it change its direction bending in one direction, bending in oposite direction bending nothing depending on the properties of the media, this is called refraction.

Refraction only has meaning for transparent objects. With opaque objects what happens with the light then enter the object can happen two things again: or is absorbed or it bounces several times with the atoms or molecules and goes back to outside as diffuse light.

How much light is reflected and how much light eneter the object is determined by the properties of the media ouside the object and the object itself and also depend on the angle of incidence of the incoming light.

The properties of the media are defined by three parameters: dielectric constant, magnetic permeability and conductivity. For dielectric materials (not conductors) the conductivity is zero so this parameter has no effect. For most non magnetic materials the magnetic permeability is very near to one and so we can ignore this parameter and only remains the dielectric constant and then we have a number called refraction index and so we can use Snell's law for refraction and Fresnel equation to determine how much is reflect and how much is refracted in case of transparent objects or only how much is reflected.

Quote - Hm, but you normally cannot the determine the speed of the processor by measuring the speed of a single instruction.

Yes you can, first you can know the intruction speed and latencies for each instruction in the data manuals for your processor.

You also can measure the execution speed in CPU clocks for an instruction or group of instructions, this is the way how codes are optimized.

Quote - If 64- and 32-bit processors were the same in every aspect, except for the number of their registers, the 64 would still be faster on average (for any real program).

64 bit and 32 bit instructions run at the same speed in most cases. 64 bit can be a bit slower because require an extra byte prefix for the code, also multiplication and division is slower.

In most of a 64 bit code you use 32 bit instructions, you win nothing using 64 bits. If you want to count from one to ten you don't need a 64 bit register, even a 8 bit register is enough. Much more, in most part of any 32 or 64 bit code 16 bits are just enough, you only use 32 bit because the processor is optimized for 32 bits and run a bit faster than using 16 bit registers.

You only use and muste use 64 bit instruction for pointers to memory access, it work with the same speed as 32 bit pointers but you are able to access memory in any memory area and not restricted to the 2/3 GB range.

64 bit code have some advantages that can make run faster the app if you optimize the code because you have eight 64 bit registers extra and eight 128 bit SSE registers extra, you can use this registers for processing instead of accessing memory.

Quote - Likely because of the cache-size; past AMD processors had much less cache than Intel processors and the less cache you had, the more any comparison you did between 32 and 64 became a comparison of RAM-speed (which is essentiallly the same for 32 and 64).

The comparison that I did was at CPU level, the execution speed of the instruction assuming all the data you need is in level 1 cache.

Cache size importance depend on the application. If all your code and data is able to fit in the cache level 1 your app will run at CPU speed. If all your code and adat is able to fit in cache level 2 the app will run a little slower.

The problem is that the caches are very small, 1MB or less, so a simple nude Vicky4 is unable to fit in any cache. Accessing Vicky4 data will always be at memory bus speed and the caches in this case have an adverse effect because the overhead of cache miss and useless cache filling.

CPUs were not designed to run 3d applications, it were designed to play mp3, watch some video or play with Microsoft Office and so, caches work pretty well for this kind of application.

Quote - Interesting, Kawecki,  as I'm contemplating building a new machine this Fall.  Is the difference because Intel does 64-bit better, or because AMD does 32-bit better?

It is very difficult to say which is better in 64 bits, Intel and AMD have taken a different approach, in some features one is better in other features the other is better and it also depend on the CPU version.

The problem is the competition, AMD and Intel refuse to follow the path of the other and who pay the fight are we. Intel introduced the MMX extensions in its Pentium and some time later AMD created its K6 with MMX and 3dNow  that would be very useful, but Intel refused to do what AMD did and never did the 3dNow extension. Some time later introduced with Pentium III the SSE instructions, that were nothing more than a better 3dNow instruction, but not compatible. Only later AMD introduced SSE with Athlon, then came SSE2, SSE3, SSE4 changed to SSSE3, SSE4a, SEE4,1 and all a mess. Today what is common to both is SSE, SSE2 and SSE3, the other are not compatible and so, a programmer cannot use them unless he want to make his software run only in Intel or make two versions of the software.

AMD was the first to introduce the 64 bit CPU and won the 64 bit battle, Intel was forced to use the AMD 64 bit instructions, but refused to do it in the same way as AMD did, the consequence was the poor performance running 32 bit code and the need to have two different versions of Windows, one for AMd64 and one for Intel64, probably your Windows CD have both Windows, if not you must have the right Window for your processor, if not it will not install. (My XP64 is only AMD)

AMD has a long tradition of making memories, processors, integrated circuits. Intel has a long CPU tradition. The problem with Intel is that always wanted to impose the standarts and be the one that set the rules. Intel ignored the PC. Intel always tried to dominate the microprocessor market dedicated to industrial applications and negelcted the PC that was the most important consumer of their products. Intel lost the war to Motorola and was forced to turn to the only market that had, the PC. Only then created the first decent CPU for the PC, the 386, before this the PC was a mess. For some time had the monopoly of the PC market until appeared competitors. Intel won the battle until entered the scenario AMD and all has changed.

Both AMD and Intel are making good CPUs, but are different and following a different path. AMD is now associated with ATI or ATI is making part of AMD so, the integration of CPU with GPU is very near, Intel seems to be fighting a lost war.

The problem is not the mesh itself, the problem is to setup the joint zones that are a headache!

I did a quick test with the bodice, that is made of triangles. The only problem that I found was with the shoulders movement. The collar present a little problem and chest and abdomen works well.

What I found with your mesh has nothing to do if is made of quads or triangles. First problem is that the mesh is not welded and so, it can sepparate the partes with some bendings. The joints for the collar need a small adjustment and the joints for the shoulders are wrong, the bending of the shoulder have no effect on the collar and maybe chest too and it must have in this cloth.

Beside this, is a nice cloth

I shall see which is the problem with your Vicky dress.

A conforming cloth is nothing more than a posable figure. A figure can be made with quads or triangles and it doesn't make any difference. I even subdivided Vicky4 breast to have better breast morphs, so her breast is made with triangles now.

Maybe I am Don Quijote, but I don't see any reason why conforming clothes cannot be made of triangles

Quote - ha ha,  talk about sense of proportions! Nothing is easier than to correct wrong normals

You cannot correct the normals of a degenate or not planar quad, three normals can be good but the fourth is always bad, you only can select which one will be the bad one.

Quote - Edit: But the most compelling argument is that there is not a single cloth model in the marketplace that is not made in quads, not one.

Yes I know and almost all the clothes that I have are made by quads, but just go to the cloth room, pose a dynamic cloth and see all the abominations that happen with the quads in some parts of the cloth. Of course in most of the cases the abominations are not visible in the rendering with the camera angle you use. And when are visible you must change the camera angle, use magnets to hide or correct  them or use Photoshop. With conforming clothes happens the same at the bending zones.

The most important factor is memory speed and not CPU speed. Mememory speed is much slower than CPU. As Poser consumes and process a huge amount of memory, this amount doesn't fit in the CPU cache and always create a cache miss slowing all the process.

Several cores do help, but the problem continue to be memory access speed, all the cores share the same memory, so the bottle neck is memory. When a core is accessing memory the other 49 cores will need to wait doing nothing, if they need to access memory.

Newer processors and its respective mother board will run faster because they have faster memory bus speed, better management of memory and faster memory that use, but the chipsets of the motherborad have a great importance. Depending on the combination CPU, motherboard and memory you even can have case where a slower CPU rendering much faster that a newer and faster CPU.

A 64 bit rendering software will run much faster because it can use many GB of memory and load all into memory without using virtual memory that is very much slower than memory, but it also depend on how is written this software.

With Intel processors in 64 bit mode a 64 bit software runs faster than the same 32 bit software. With AMD processors it has no importance, 32 and 64 bit softwares run with the same speed.

Quote - If these conditions cannot be met, then you're not "rendering" somewhere else... it seems to me that you are doing the job twice... setting it up once in Poser, then doing it again elsewhere.

You are not doing twice the same thing. First at all you need to load your characters and props, pose them apply morphs, dress with clothes conforming or dynamic. This only can be done in Poser, the only alternative that I know is DazStudio

The second step is set the lights, camera, set the ambient effects and maybe change the materials. This you can do it in Poser or other app. If you want to render in Poser do it in Poser. If you want to render in other app then do it in other app.

LuxRender is very limited, it only render the scene nothing more. Kerkythea is more friendly and you can setup all there.

Quote - But still, the triangular mesh looks more like natural cloth. So why can't you have both?

Quad mesh is easier to model and consume less memory, but triangle mesh are better for conforming and dynamic clothes and any deformable object because its normal are always correct after the deformation. A quad  after a deformation can produce a degenerate quad and so have wrong normals and can produce undesired illumination or black spots in the rendered image.