I'm glad people really are leaving this time. Every time this site has suffered one of these catastrophic "new and exciting upgrade!!!" disasters there have been lots of people threatening to leave, but few actually do. The pattern has been the same for years: Big ugly makeover, community uproar, admins do the "be patient, we're still ironing out the little wrinkles" routine with hugzzz and smileyfaces, and management just hunkers down and waits for it all to blow over. And it does - eventually the complaining stops when people realize nothing will ever improve, and eventually one forgets that everything used to be better. But it occurred to me that perhaps some of the ongoing squabbling in the Poser forum might be due to a growing level of general frustration and discomfort in this shitty site environment, like kids in a hot car.

If we don't learn from history we are doomed to repeat it, I finally got tired of repeating this particular history lesson. I shut my store and deleted my gallery, this was a home for 12 years but I'll find another. I'll find a car parked in the shade with the windows open.

If management appear to be attempting to un-break a few things, perhaps it's only because enough people really did leave this time. But for me it's too little, way too late.

PS: I had to type this twice because when I came to post it I was taken to the LOGIN page (even though I couldn't have typed anything in a community forum thread if I wasn't already logged in. Duh!) This is the sort of crap that adds up and adds up and eventually makes this place intolerable.

Well, I was an old time member, joined in 2003, but with this latest downward lurch, I'm done. As of a few hours ago I'm no longer a vendor here and probably won't stick around much as a visitor. Enough is enough.

You're saying this mess actually went through a beta? Surely you jest!

I experienced this same thing after the last forum 'upgrade' and it took me months to figure out a fix. I had to go to my forum preferences and delete the values for "threads per page" and "posts per thread" or whatever the options are. Saving the change reset them to their default values (30 for both) and that fixed the screwy last-page-is-not-the-last-page thing. Any choices other than 30 and 30 broke it again.

Unfortunately, that fix doesn't work this time, Pinky and Perky in the control room have broken it more than it was already broken. Am I surprised? Sadly, no.

Nyghtfall said: "What I find particularly noticeable about BB's Superfly render is the lack of shadow artifacts where the ceiling meets the wall above the windows. Unless SM has finally solved that problem with Firefly, they may have finally provided a solution with Superfly."

What you're seeing here are unbiased renders - a totally different animal. The shadow artifacts you're referring to are a common problem with interior scenes rendered with a biased renderer (like Firefly). They are a result of the way biased rendering works: to save time it takes a predetermined number of samples from the scene, measures the light at those points, and interpolates the light in between them, essentially guessing. Usually the guessing works well enough and the time savings are huge, but in tricky areas that interpolation process can yield screwy results which show up as shadowy and over-bright splotches in the render. More samples closer together can obviously help, but that takes longer, so the better biased renderers try to find more efficient ways of taking the samples and more sophisticated ways of processing them.

With an unbiased renderer (like Cycles) it doesn't interpolate anything, it just keeps firing rays into the scene until, if given enough time, it has sampled every single pixel. If you stop it before the process is complete (either by hitting the stop button or by limiting the total number of rays it should fire) you get a grainy image - the grain is the pixels it hasn't had a chance to do properly yet.

The trade-off is time. Guessing is quicker than firing lots of rays. By way of example, the unbiased MODO render I posted above took 35 minutes and it was still grainy (that was using modo's preview render set to full size). The same scene using modo's normal biased renderer took just 50-something seconds, but it has some splotchies around the edge of the ceiling. It typically takes quite a bit of skill, experience, guile, cunning and trickery to get a biased renderer to deliver the result you want (to help it guess correctly, if you like), whereas an unbiased renderer doesn't get involved in all that, it just needs fast hardware and time.

STILL can't drag text to rearrange it in the editor window - while you drag you get sort of "Drop your image here" icon popping up and your text editing is ignored. They were too busy fiddling with important stuff like avatar shapes again. Un-____ing-Believable.

 Having said that, all of them are still grainy and if experienced users can't avoid that then I'm certainly doomed.

You're not doomed at all, with a progressive renderer just giving it more time to cook should get rid of the grain. I just stopped mine when I thought the grain looked about the same as BB's. Compare the ceilings or the white ball on the left - they're not that far off - not so far that a comparison cannot be drawn, as shvrdvid is implying.

BB posted the time that render took, so I assumed the information was relevant. Hope I'm wrong, 'cause it sounds like socks.

And here's what MOD801's not-particularly-speedy preview render did on that scene, I stopped it after 35 minutes. I' be interested to see what blender's cycles makes of it, if somebody wants to try this scene in blender-Cycles, drop me a PM and I'll send you the obj.

 I have never come close to something like this in Firefly - impossible.

Render time was 3 hours 4 minutes.

I think it's really great that it's finally possible to do stuff that firefly simply can't do, and to easily do stuff that firefly could only fake by means of very complex node setups.
But... 3 hours is... glacially slow for that render.

Here's a rough facsimile of that scene rendered with Twilight, a plugin for sketchup - I stopped it after 11 minutes.

 You wanna be the one to tell them their stuff doesnt work anymore? Be my guest. Suggestion: keep your asbestos suit handy. You're gonna need it.

It will always work in firefly.

She seems to have disproportionately large irises. Not sure if it's texture, UV or model related, but I've seen it in all her promos, so I assume it's the latter. This particular texture set accentuates it with an extra band around the iris.

Here she is compared to an actual, you know, non-digital amazed person.

And she could also really use some hand cream.

I am sure this will come off as odd, but if you are going to do thru all of that, why not just use Genesis in Studio and be done with it?

A valid question, but I think one needs to keep in mind that quite a number of Poser users just can't or won't deal with Studio's interface for whatever reason, so the only way they are ever going to use Genesis is in Poser, half-crippled though she may be. If you regularly switch between complex applications it can be difficult to understand this reaction to Studio's rather innocuous interface, but I suspect that for some of those people, Poser is the only complicated program they know, and because DS is something strange and different, it appears doubly complicated. > There seems to be some misunderstanding that Genesis is for the Daz Studio environment only. The good news is there are multiple ways of using some of the Genesis figures inside of the POSER environment. I thought I would share some info and useful

Some misunderstanding, about Genesis, in this forum?? Surely you jest!

And now back to scheduled programming:

"So anyway, I heard that when DAZ takes over the world they are going to make Poser illegal"

"Well I heard that Genesis is going to eat our children"

;o)

If models that are 100% complete in every detail exist at all, they'd be so large as to be quite unusable in a large scene.That's if they exist. Generally I'd imagine that detailed models fall into 2 categories - those made by professionals with a particular purpose in mind (advertising/marketing, CAD/CAM design and manufacture etc) and those made by hobbyists just for the sake of making a cool very detailed model. Both groups tend to focus on what interests them and leave out stuff that doesn't. The hobbyist who models the inside of a gearbox including the the little spring under the lip of the oil seal (because he has one open on his workbench for reference) isn't going to bother to make a car around it at the same level of detail because life is too short. The car manufacturer who produces detailed models of every part isn't going to bother to model every crimped-on terminal on the wiring harness because they simply don't need to see those. Both types of model will have too much detail in some areas and not enough in others.

Both types are made with a particular focused intention, localised detail for its own sake for the hobbyist, particular selective details for the pro, with no consideration given to optimizing the model so that it can be used as part of a large, complex scene (such as a game asset would be made).

Something like the car would probably have been modelled using a NURBS or parametric solids modeller (like Rhino, ProEngineer, Solidworks, Inventor etc) and even in its native format it would be a monster file with a million layers and a million material definitions. Converting that to a polygon model would be just the start of your problems - the process is never straightforward, successful conversion between formats generally requires fluency in both applications and a full understanding of the strengths and weaknesses of each. Almost every nurbs part would require different export tesselation settings to achieve the right balance of enough polygons to define the shape but not so many as to overwhelm the system. Modelled detail like threads on bolts and studs would probably overwhelm anyway, so you'd have to re-work all that as bump or displacement maps in your poly model. The original application probably made extensive use of instancing, clones and proxies for repeated elements like screws and bolts to keep the file size down, you'd lose that optimization during conversion.

How would you see inside it? By means of physical cut-away booleans, or render booleans, or peeling away geometry, or changing opacity? Each of those would require a different approach and the way the model is set up would be quite different for each.

I think models for a project like this need to be made specifically for the project. Optimization is everything, keeping one eye on the bigger picture all the time, adding detail only where you need it. You simply don't need 100% detail on every part because you won't see it all, there's no way to see every detail of every part. You only need enough detail for it to look right in the scene, and that's a lot less detail than you think.

 but some people might ... think "wow, i should just keep my mouth shut no matter what, just to cover my arse." 

Now there's an interesting mental image.