What's wrong with my renders - DS 4.10 to 4.11

jmlazgar

I load the new Hikaru for G8 character, and dranted, the lighting is different. But my 4.10 looks about right, except for something going on around the mouth, but my 4.11 version has "plastic" skin. I believe I used IRAY Skins for both. I understand the lighting on 4.11 is different from 4.10 .... I Set render settings to default in both versions. I haven't notced this in other 4.11 renders. Any ideas? 4.10 version is on left, 4.11 is on right
Thanks

Jay

jmlazgar

So,

ditched the 4.11 render and am working on the 4.10. Can't figure out why ist's giving me a green tracing around the lips. I have set up entirely new scene and lighting, tried both Iray and non  iray skin. Same result.

Please help

nicstt

I'm not seeing any differences in appearance; 4.11 is slower though, particularly when doing skin as the added 4.11 features don't help so far.

The lighting is different you state, which is true, find out what is different; is the cameral light on? It is best kept off - as a rule.

Edit:

looking at the two images, and the reflections in the eyes, you have two different light set-ups. It is also possible there are different shaders on the the characters. Save a scene in 4.10 you like, open it in 4.11; it should be the same. The 4.11 iamge has other items in the scene; this always makes some difference. Light reacts to objects - just like real-life.

Richard Haseltine

nicstt said:

I'm not seeing any differences in appearance; 4.11 is slower though, particularly when doing skin as the added 4.11 features don't help so far.

The speed is because of extra checking, in handling normals on translucent surfaces, which older versions did not (and which Iray does not currently do itself).

http://docs.daz3d.com/doku.php/public/software/dazstudio/4/change_log#4_11_0_91

nicstt

That's interesting Richard; any idea what affects we might expect those checks to result in?

Richard Haseltine

nicstt said:

That's interesting Richard; any idea what affects we might expect those checks to result in?

You mean in renders? Presumbaly it's avoiding issues which could coem up before, but I've no idea which content (if any from the store) showed the issue to do a before and after comparison.

Richard Haseltine

Richard Haseltine said:

nicstt said:

That's interesting Richard; any idea what affects we might expect those checks to result in?

You mean in renders? Presumbaly it's avoiding issues which could coem up before, but I've no idea which content (if any from the store) showed the issue to do a before and after comparison.

Apparently if you have normals on for the default Genesis 8 skins then the direction of the surfaces will be switched in 4.10 - what should dent in will pop out and vice versa - while being correct in 4.11. Thick surfaces with translucency.

nicstt

Richard Haseltine said:

Richard Haseltine said:

nicstt said:

That's interesting Richard; any idea what affects we might expect those checks to result in?

You mean in renders? Presumbaly it's avoiding issues which could coem up before, but I've no idea which content (if any from the store) showed the issue to do a before and after comparison.

Apparently if you have normals on for the default Genesis 8 skins then the direction of the surfaces will be switched in 4.10 - what should dent in will pop out and vice versa - while being correct in 4.11. Thick surfaces with translucency.

Ahh, I use normals a lot. Good to know, cheers.

Richard Haseltine

nicstt said:

Richard Haseltine said:

Richard Haseltine said:

nicstt said:

That's interesting Richard; any idea what affects we might expect those checks to result in?

You mean in renders? Presumbaly it's avoiding issues which could coem up before, but I've no idea which content (if any from the store) showed the issue to do a before and after comparison.

Apparently if you have normals on for the default Genesis 8 skins then the direction of the surfaces will be switched in 4.10 - what should dent in will pop out and vice versa - while being correct in 4.11. Thick surfaces with translucency.

Ahh, I use normals a lot. Good to know, cheers.

It's not, as I understand it, all normals but normals with surfaces that have non-zero translucency and which are "thick".

nicstt

Sounds like it isn't jpeg normal files, but more likely Tiffs?

Richard Haseltine

nicstt said:

Sounds like it isn't jpeg normal files, but more likely Tiffs?

No, it's the interpretation of the map relative to the geometry I think - not a format issue.

SpottedKitty

nicstt said:

Sounds like it isn't jpeg normal files, but more likely Tiffs?

They'd better be .tiff files. I've come across a number of sources over the years stating that the lossy nature of the .jpg format makes it not really suitable to use for normal maps. The exact value of the normal direction is dependent on the exact values of the separated colour channels of each pixel in the normal map — that's something you get as default in a .tiff but due to its lossy compression the .jpg format doesn't allow for that level of colour fidelity. IIRC most of the normal maps used in DAZ products (especially Genesis characters) in recent years, and in old PC+ items that have been updated to use D|S files and Iray materials, use .tiff format.

nicstt

SpottedKitty said:

nicstt said:

Sounds like it isn't jpeg normal files, but more likely Tiffs?

They'd better be .tiff files. I've come across a number of sources over the years stating that the lossy nature of the .jpg format makes it not really suitable to use for normal maps. The exact value of the normal direction is dependent on the exact values of the separated colour channels of each pixel in the normal map — that's something you get as default in a .tiff but due to its lossy compression the .jpg format doesn't allow for that level of colour fidelity. IIRC most of the normal maps used in DAZ products (especially Genesis characters) in recent years, and in old PC+ items that have been updated to use D|S files and Iray materials, use .tiff format.

Most normal maps are jpegs, sadly - although there is some decent detail there in the ones saved as higher quality. Although, I have noticed some starting using tiffs, which is good.

Edit

I've been experimenting with 4.11 again.

Actually, I've been getting better results than I initially thought; this is due to the fact, that I like a little noise, it can simulate the affect of grain in photos, which results imo in a more pleasing image. Too perfect looks odd.

nicstt

Richard Haseltine said:

nicstt said:

Sounds like it isn't jpeg normal files, but more likely Tiffs?

No, it's the interpretation of the map relative to the geometry I think - not a format issue.

Sounds like we need more info there.

LenioTG

The "jump" bug on 4.10 used to bother me so much and so often that everything else doesn't matter to me! :D

Richard Haseltine

nicstt said:

Richard Haseltine said:

nicstt said:

Sounds like it isn't jpeg normal files, but more likely Tiffs?

No, it's the interpretation of the map relative to the geometry I think - not a format issue.

Sounds like we need more info there.

As I understand it it potentially affects any surface with a non-zero translucence and normals, and "thick" - I'm not sure if that just means Thin Walled off or a closed volume or is actually dependent on the thickness of the emsh or size of the scattering distance.

nicstt

That probably explains the slow down I'm seeing as I usually use translucency, normals and possibly other factors too.

Richard Haseltine

"Thick" is not, I'm told, any particular setting or mesh characteristic but a matter of how the material is intended to be used. You can toggle the change off for a surface with the Imvert Transmission Normal property on the surface settings.