Now I am running out of ideas. Maybe there is an obvious mistake which I just can't see?

Any help is appreciated.

Delete the .txt at the end.

Posers import of lwo files is very iffy.

I always use PoseRay (it's free) to convert the lwo to OBJ.

thank you for the hint, but it did not help. Still the same shading, even though it is now an obj and PoseRay triangulated the mesh.

Attached Link: "Simple Room" Tutorial compliments of Dr Geep Studios

Maybe this tut could help.

cheers,
dr geep
;=]

Poser doesn't like long thin tris and will shade them. For that matter, it doesn't like long, thin quads either.

My suggestion: turn off smoothing for all items (located on the object's properties panel). That should fix it.

Laurie

You can export an obj directly from Lightwave.

Remember that any uv mapping is lost if you dont assign the uv map in lighwave materials.

Poser does not handle thin polys very well. Select the obj and split vertices should fix it. Merging the polys and  splitting them differently is a better solution.

Try turning off smoothing in Poser and see if that helps.

When common vertices are used, Poser tries to "bend" the light (i.e., texture) around the corner.

When the vertices are "split" then Poser will display the faces with a sharp edge.

cheers,
dr geep
;=]

Hi all,

thanks for the tips. Using Crease Angle = 22,5 solved it. I forgot that the window frames are beveled, so 80 was not low enough. But even then, I thought that switching off smoothing avoids this "surface smoothing" at all. When I made the screen shots smoothing was off, but I still got the unwanted shading. So last question is, how to switch off "surface smoothing" completely? Is it possible at all, or is the only solution to split the edges where required?

you can turn off smoothing in 2 places. globally in the render settings, or on a mesh level per object.

I turned it off on the Object. (it's along with the crease angle settings.)

madno2 is referring to the phong interpolation of normals, which is affected by crease angle, and has nothing to do with smoothing.

The smoothing boolen checkbox affects the smooth polygons feature, which is separate from phong interpolation.

Both of these features (polygon smoothing, polygon interpolation) are going to pay attention to crease angle. Since they are both influenced by crease angle, many people get them confused. They are two different things.

You cannot disable phong shading. You can control it with crease angle, or you can split edges - in the latter case no interpolation happens, ever.

http://en.wikipedia.org/wiki/Phong_shading

Phong interpolation of normals is why the low-res sphere appears to be smooth, even though it is low res, and even though the perimeter appears to be not round. If you then enable polygon smoothing, then the perimeter will appear to be round as well.

Phong interpolation of normals is to polygon smoothing as bump mapping is to displacement mapping.

Hello everybody,

thanks again. Got it and know the name now "Phong Shading". This Mr. Bui Tuong Phong was a clever guy. He gave us a smooth surfaces without the need of to much geometry.

Probably after the fact, particularly after BB's very knowledgable explaination, but I was wrestling with this problem with sharp edged models also, particularly artefacts from bevels. With strictly hard, straight edged objects where no smoothing is required you can completely disable smoothing by setting both crease angle to zero and smoothing to off. Either one or the other by itself won't work and poser will still smooth the object.

Another method I've found that works if you want some smoothing and still maintain the appearance of hard edges requires that you increase the poly count of the object somewhat by creating a series of micro bevels around the edges of the objects. I like this method because you maintain the appearance of hard edges, get the specular effects that bevels produce, and don't have to spend alot of time tweaking crease angles or deal with unwelding. This method is illustrated below by a couple of preview renders. the results are the same and better with full on firefly renders...

...You can also assign smoothing by material regions in some modeling apps and UVing software (UVMapper Pro allows you to utilize this most flexible option).

Anyway, your lucky that people here seemed to respond to your problem with some real answers... when I posted the same query a few monthes ago I didn't really get any kind of concrete answer other than unweld the object. Research and experimentation on my own was how I discovered the solutions.

In LW unweld all points and export as (wavefront) object. Here is a render after I imported. No settings were changed on the object in Poser.

I'm sorry, jerr3d. Did you read any of the above posts, it's been long established in the thread that unwelding is one method of approaching the problem.

Quote - Anyway, your lucky that people here seemed to respond to your problem with some real answers... when I posted the same query a few monthes ago I didn't really get any kind of concrete answer other than unweld the object. 

I missed your query, if I had seen it I would have referred you to an old image I made in response to the same query a long time ago:

A single bevel on the edge doesn't fix it because the large surfaces are still involved in the smoothing - large faces + smoothing = smoothing artifacts = ugly

Splitting verts or turning off smoothing/overriding crease angle are all going to give you a sharp edged model which probably won't look realistic, sharp edges don't occur very often in real life. 

Adding a billion polys to your model fixes it but now you have a billion polys in your model.

Placing another edge near the corner allows the smoothing to happen on the corner but it doesn't involve the large faces, in my opinion this gives the best looking model for the lowest polycount. Different modeling software gives different names to this operation, some call it inset, some extrude, all that matters is that there's a narrow poly near the edge in the same plane as the large surface (not bevelled) . The width of that narrow poly will determine the radius of the rounded edge after smoothing.

Here's the image and the original post below:

As you can see splitting the edges (or turning off smoothing in Poser if it's possible) fixes it but gives you a 'dead' looking model. 

'bevelling' the edges helps a little but doesn't doesn't cure it, the smoothing action still goes across one face, round the corner and along the next face (curved arrows).

The best-looking model comes from adding extra edges just in from the corners. The smoothing still happens but it is confined to those very narrow faces near the corner, it doesn't affect the big flat faces, giving that little edge highlight. 

Sure it costs polygons, but you can make savings elsewhere.
You can also leave the bottom faces off completely if they won't be seen, that avoids the problem. So make your table legs as open tubes, no top, to bottom, then you only have to fix up the 4 long edges.

No offense to anyone but this thread is suffering from some serious "post before you read or look" syndrome.

Those ARE called bevels in some programs. Control edges in others.

Laurie

Quote - obm890, please see my second image and you'll observe multiple narrow faces as you are describing above. In my early morning fatigue I mistakenly referred to this as bevels, They're actually insets.

No, those are the ones I was referring to when I said "Adding a billion polys to your model fixes it but now you have a billion polys in your model." You have 6 very narrow polys going around the corner, those are forming a rounded bevel. Some of those polys turn the corner. The original sharp corner is gone, replaced by those 6 little faces making a rounded edge. They are redundant because Poser can create the illusion of a rounded edge, you just have to control where it happens. 

They are not what I was referring to when I said "all that matters is that there's a narrow poly near the edge in the same plane as the large surface (not bevelled)". In the same plane. The original sharp corner of the object is still there, you just add a control edge very near the corner to limit the extent of the smoothing at render time.

Maybe you started off by insetting the faces in Wings (which is what I'm talking about), but then you hit the subdivide button to get the result you show? 

I'm not saying your solution doesn't work, it does, it's just inefficient. I downloaded your Messiaaahg prop to take a closer look at the modelling and the part you show in the above image is 11700 polys. It's just a stick, perhaps the simplest part of your prop (Antonia-Lo is only 9500 and she's a human figure with eyelashes and fingernails and teeth). The whole Messiaaahg prop is 338000 polys, that's probably a few more than you need to define those forms. If Stonemason made his props with that sort of mesh density his scenes would be unusable.

I think it's an application specific terminology confusion, LaurieA. In wings (which I know you're familiar with amongst others) the function I used was Face:Contour. In later versions of wings Contour became Inset which obm890 refers to. So yeah, a bevel by any other name is still a bevel (or is it?). Works for me.

Anyway, hope I didn't come off as overly hostile, obm890. Someone in a different thread raised my blood pressure a little right before I was getting ready to leave for my hideous job... I'm afraid it may have spilled over into this one. You've always been very helpful and amiable.

also, otherwise what do you think of the model for someone who has been doing this for basically less than a year?, knowing that you took the time and effort to download so that you may analyze my modest modelling skills... and considering this don't you think it's a little unfair to compare me, a newbie, to someone like stonemason?

see image above? any clues?

And yes, you're right, it started as a single inset control edge... having discovered that the model at a lower resolution was crawling with these bizarre artifacts at rendertime I subdivided.

I wasn't trying to be preachy...lol. Just trying to lessen the confusion (and 3D sure is confusing...lol).

Laurie

didn't thougt my initial question would cause such an informative discussion.

Tested some more.

Poser has "Smoothing Groups" functionality (those groups can be created in the group editor). Documentation says polys that belong to different Smoothing Groups will be shaded indepentend from each other. Means, the smoothing by the phong shading does not cross the borders between those groups. That means one can keep merged edges, have a higher crease angle and still get sharp edges.

I tried it out and come to the conclusion that it works, but it is tricky as the ceometry within each group is relevant too.

I made two versions of the "room" now. One with "triangle like quads" and with "quad like quads". Everthing else (mats, smoothing groups, crease angle) is the same for both.

I've heard that smoothing groups are tricky, a suggestion that I read said that it should really only be used as a last resort... I'm interested to see what others have to say on the subject.

Following.

Just for fun I deleted all the groups I have created previously. To my suprise the smoothing IDs seam to remain internally. Maybe Poser deletes the groups but not additional information assigned to them.

Quote - ...in poser 8 for some reason I keep getting these artifacts where it looks like the mesh is showing as little black lines... 

see image above? any clues?

I'm not seeing them on the prop I downloaded, maybe you can post a bit of mesh which is showing the artifacts and I'll take a look at it.

Quote - ... and considering this don't you think it's a little unfair to compare me, a newbie, to someone like stonemason?

I should have qualified that reference to Stonemason. Every time the subject of economical mesh comes up there are posters who say "polygon count doesn't matter, computers are fast now, who cares about a few extra polygons?".  I was just preempting them, they'd better hope people like Stonemason continue to do what they can to limit polygon bloat in a scene with lots and lots of objects, or that fast new computer isn't going to seem quite so frisky.

It's a bit like saying "Conserving energy is stupid, it's cheap and there's plenty for everyone". Riiiight. But if every person in India and China and Africa takes the same view and installs aircon and buys a Hummer, things will change fast, for everyone.

Really hoping to find a viable solution to this...

I'm not even sure that these "Mesh X-Ray" artifacts will appear in someone elses installation of poser but I'd rather not risk it. I'm pretty sure I read a discussion between Bagginsbill, PJZ, and Corodan that briefly mentioned the problem but I can't find it now and it was all rather vague as if they couldn't figure it out either.

Try deleting the tongue on messiaaahg! and rendering the interior of the mouth, If I remember correctly they were showing up on the final model in the mouth interior (in my renders at least). they're very minor, but present.

I'm using P7 and PP2012 and I'm not seeing any artifacts, even inside the mouth. If you get them again on another mesh post it and I'll see if I can reproduce them.

Great news!, obm890... now I don't have to make all kind of concessions in my modeling to accomodate additional levels of subdivision. It's the strangest phenomena, I'll have a model at a state where I like it (imagine much of Messiaaahg! at one or even 2 lower levels of subdivision, particularly in the simpler forms), I go to render and there they are. All I know is that it's not caused by non-planar faces (I'm careful to avoid that) and that it's really a gigantic PITA.

I'll start a thread with some images and an example upload sometime during the coming weekend. Hopefully this is really only a problem with my copy of poser(8) or that the cause can be tracked down and squashed.

Thanks.

I'm thinking it has something to do with shadow samples settings, which I just recently changed to a proper value to match shadow blur radius (i.e. If blur is 0.4, set samples to 40 or thereabouts)... I just tried to reproduce the artifacts in the mouth and I'm not getting them any more. Prior I had my sample settings way too low. Another example of Poser inexperience. Need to do some more testing but I've got to go to work, I'll post an update of my tests later tonight.

I've seen this... or at least I think it is the same thing... when importing some models I've made into Poser Pro 2012.

Seemed to only happen with certain shader set ups... but yeah... or so I thought...

I thought it was either the shader or to do with shadow min bias... and it may have been, in my case I guess.

Nonetheless that nugget about the shadow blur relative to the samples is very interesting - thanks Primorge.

Haven't spotted the issue on Messiaaahg as yet, when I've been playing with that... but then I've not ventured down its throat. Yet ;-)

I didn't read all so I'm sorry if I'm repeating an answer.

Those meshy artifacts are self-shadowing, the result of smoothing, which also happens with displacement. There ends up being two versions of the mesh - the original unsmoothed which is used for raytracing, and the smoothed (geometry altered) for actual rendering of geometry. These two are at odds with each other and produce inconsistencies in shadowing (raytraced). It's part of the baggage of being a hybridized raytrace-REYES renderer.

If you increase shadow min bias, it goes away.

And I think I saw a complaint that nobody ever explained it. I will counter complain that I've explained it at least 20 times and I was not read.

Shadow min bias is the cause of a lot of different kinds of grief. I searched for some posts I've made. If you don't understand this parameter, you don't understand Firefly.

http://www.renderosity.com/mod/forumpro/showthread.php?message_id=3902894&ebot_calc_page#message_3902894

http://www.renderosity.com/mod/forumpro/showthread.php?message_id=3797767&ebot_calc_page#message_3797767

http://www.renderosity.com/mod/forumpro/showthread.php?message_id=3936336&ebot_calc_page#message_3936336

http://www.renderosity.com/mod/forumpro/showthread.php?message_id=3919633&ebot_calc_page#message_3919633

http://www.renderosity.com/mod/forumpro/showthread.php?thread_id=2831427&page=2#message_3820239

You're right bagginsbill, increasing shadow min bias was one way to lessen the artifacts (with a corresponding loss of shadow quality)... another way that I've found to eliminate them entirely was to disable smoothing by setting all smoothing (in properties settings) to none and setting crease angle to zero. this, of course isn't an acceptable way to render an organic object so it's not really an option.

What's peculiar is that my shadow min bias for raytraced shadows is usually set around 0.3500 which is a happy medium for me because lowering the value generally produces the polygonal shadowing that you speak of and raising the value results in a loss of shadow detail and granular looking shadows. I recently did a render with 0.350000 min bias, 5 blur radius, 50 shadow samples, shadow strength of 8 and the artifacts did not appear. If, however I use my old settings which were 0.350000 min bias, 5 blur radius and 20 shadow samples, (same object) I get the mesh artifacts. I don't know what combination of settings is really causing this but I plan on tracking it down. I was convinced that maybe it was some problem with Poser 8 as I don't recall this phenomena in poser 7.

Most of my knowledge about raytracing with Firefly comes from Blackhearted's "Firefly Render Settings and Tips" tutorial pdf (all of his suggestions seem pretty sound in practice to me) and the information in the Poser Reference Manual.

Thanks for the links, which would have been difficult to locate as I'm not really quite sure what the problem was (is) to begin with. Hence, I wouldn't have known where to search. I'll make sure I read all of your observations on Firefly.

Thanks for taking the time to address my confusion, BB!

update... Definitely Min Bias settings as the culprit. The question is, I know that many experienced poser users utilize lower min bias settings than I do. When I share one of my models will these artifacts appear and if so is it then the job of the person rendering the object to increase the min bias or is this a fault, somehow, of the modeler and his work?