I've purchased a number of very beautiful sets from a variety of vendors, but keep running into an all-too-common problem: Wall transparentcy.

It appears that all of the walls are modelled as one piece in many of these sets, and that makes it impossible to hide (or even make transparent!) one or more walls for correct posing. The stock advice from the vendors is "Use a wide angle lens" which is horse####, because of the curvilenear distortion that is introduced.

So here's the REAL question: Is there a way to break these models into pieces... that is, segment them so that the walls can be made invisible one at the time? I suspect it will require a modeling program to do this, in which case, I'm out of luck. The complexity of these programs is beyond me.

I seem to remember at one time someone (Bagginsbill?) had a work-around for this, but a search of the forum comes up empty.

Any and all help appreciated.

 

transparency map will work.

use the uv map as a quide and make a black/white mask to remove a wall or part of it.

apply that mask to transparency and specular and set transparency to 1 with 0 for the two falloff/edge  values.

Bagginsbill I believe had made a prop that acted as a tele-converter.

Use the Grouping Tool to make the walls separate objects and then you can hide a wall as needed.

cheers,
dr geep
;=]

Remember ... "With Poser, all things are possible, and poseable!"

cheers,

dr geep ... :o]

edited 10/5/2019

Quote - I seem to remember at one time someone (Bagginsbill?) had a work-around for this, but a search of the forum comes up empty.

You may be intresed in the following quote from bagginsbill:

Quote - Or you can use a "persepctive zoom lens" prop in front of your camera. What I mean by that is you can keep your camera inside the room, but give the perspective of being somewhere outside the room, perhaps even a thousand miles away.

I did this trick when I built my GenIBL tool. For purposes of creating an IBl for an enclosed room, you have to have the camera and mirror inside the room. But the perspective needs to be infinitely far away, so that the view lines are parallel.

I put a one-sided square in front of my camera, very close, and it is parented to the camera. This square must be exactly perpendicular to the view plane. So you want to do this with the camera pointing in a precise direction. Once parented you can move the camera freely.

Then you put a Refract node on the square, turning it into a lens. (You have to turn Diffuse_Value and Specular_Value off as well.) With an IOR of 1.0, the lens does nothing.  As you increase the IOR, it creates exactly the same perspective change as moving the camera straight backwards. With an IOR of 1 million, the camera is effectively hundreds of miles away.

The above quote comes from this thread:
Is it possible to look through walls?

Yes, that's the thread I was thinking of, but was never able to make it work. Can you put up a screen shot of the material room settings, i.e. what plugs into what?

 

"add refraction" to one-sided square prop, place camera and prop in room.  in versions prior to poser 9, there is issue in re: FFRender and transparency calcs., but is not fatal error in ordinary poser render IMVHO.

Miss Nancy... I'm almost there. Where did you get the node marked "BG Color"? I don't see it in any of the new node options?

That's just a simple color node with a different name.

I tried the BB lens method too but could never get it to work.  However now that there is a screen shot I guess i'm in with a chance.  will try it again.

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

he's been pretty quiet about it when people post questions about this so my guess is it will be a coming feature in poser 9

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

Well, I've got it set up exactly as Miss Nancy show it, and all I get are solid white renders.

you wonder why modellers go to all the trouble to make a detailed scene eg a toilet room, but don't have any provisions for camera angles.

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

I'm not sure if:

1. They don't understand how people will want to use their models

2. Have no idea how to do it

3. Could care less... We don't find out until after they have our money that the model is poorly designed

Still, it doesn't matter. It makes for a faster workflow to deal with their mistakes than to try and design everything from scratch. At least for me.

I like to hope that #2 is the answer.

I think people who spend their time making models have very little time left to use poser for fun things so they probably never realize what we need.

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

You are kinder than I, Esther. One of the things I love about ya, hon!

I have my moments.

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

Quote - I think people who spend their time making models have very little time left to use poser for fun things so they probably never realize what we need.

 

This is why you need good QA.

 

"Dude, I can't see into this room you've made" is a legitimate problem.

WARK!

Thus Spoketh Winterclaw: a blog about a Winterclaw who speaks from time to time.

 

(using Poser Pro 2014 SR3, on 64 bit Win 7, poser units are inches.)

 

     The modeler may understand more than you realize.  There are considerable rendering penalties with some of the solutions suggested:

• transmapped model - if you use this method, nearly every pixel the Poser camera sees is a transparency surface, even if those areas are set to be opaque by your mask.  It will take Firefly longer to render, and considerably longer if IDL is engaged.

• separate pieces - to try to provide for most of the possible camera angles, I may have to break one prop into 150 pieces.  That in itself isn't difficult, since it was originally modeled in pieces, but every piece will have to have its own set of material zones, so the memory overhead increases considerably.  There is also the matter of a novice user being overwhelmed by umpteen pieces with three-four levels of hierarchy.  Changing materials becomes a chore because each piece has to be re-matted individually.

     Adjusting focal length works and is easy.  If you want to avoid the fisheye distortion, use BB's virtual perspective zoom lens.  If you can't manage that, I am amenable to modifying my models if it would help.  As it is, I try to strike a compromise between optimizing for low system resource burden and optimizing for flexibility of use, but no compromise can anticipate every situation.  Sometimes I include both one-piece and exploded-view versions in a package.

(estherau)

Quote - I think people who spend their time making models have very little time left to use poser for fun things ...

     That describes my situation well, but I do set up material test and demo scenes, so I do encounter the same challenges as any other user would.

Poser 12, in feet.  

OSes:  Win7Prox64, Win7Ultx64

Silo Pro 2.5.6 64bit, Vue Infinite 2014.7, Genetica 4.0 Studio, UV Mapper Pro, UV Layout Pro, PhotoImpact X3, GIF Animator 5

Quote - I've purchased a number of very beautiful sets from a variety of vendors, but keep running into an all-too-common problem: Wall transparentcy.

It appears that all of the walls are modelled as one piece in many of these sets, and that makes it impossible to hide (or even make transparent!) one or more walls for correct posing. ...

If it is only for posing (ie seeing what is inside the room in order to place your objects), it is often easiest to use the hither of your camera. Point the camera to the room (from outside) and set the hither (parameters panel of the camera) to approximately the distance from the camera to the object. You have to use one of the perspective cameras for that (dolly or main). It only makes everything nearer than hither invisible, you can still select it, so be careful. It does not affect render, though.

Use the Grouping Tool to make the walls separate objects and then you can hide a wall as needed.

 

This would have been my choice over Xray glasses *

Yep, them dang modelers is evil I tell ya. Pure unadulterated evil.

 

Them nefarious purveyors of manipulated deltas should come to my house and render the image I want for me!

I mean how dare they expect me, a poor oppressed user, to have a modicum of knowledge on how to use the program of choice.

Outrageous!

I know that they all sit around and brainstorm on ways to inconvenience particular users.

It's a conspiracy! I say we do away with the vile louts!

Wait...um...crap! I model.

Dang it! Guess I'll have to go off myself. (Have to stay consistent and all ya know.)

 

 

"Goodbye cruel Poserdom.

I can no longer live with the loathsome being I've become through my addiction to pushing vertices around. I know there can be no redemption for one such as I. Therefore my abandoning this mortal coil is the only recompense I can make for my heinous behavior. To all that I've forced to actually do something for themselves I beg...."

 

Ah Eff it, going to go model something.

Whine on good users.

Quote - separate pieces - to try to provide for most of the possible camera angles, I may have to break one prop into 150 pieces.  That in itself isn't difficult, since it was originally modeled in pieces, but every piece will have to have its own set of material zones, so the memory overhead increases considerably.  There is also the matter of a novice user being overwhelmed by umpteen pieces with three-four levels of hierarchy.  Changing materials becomes a chore because each piece has to be re-matted individually.

I'm not asking that EVERYTHING be given the ability to hide... just the main 4 walls.

Quote -      Adjusting focal length works and is easy.  If you want to avoid the fisheye distortion, use BB's virtual perspective zoom lens.

IF the zoom lens would work! I've entered everything the way the screen shots show it to be set up, and STILL all it renders for me is white. I'd be more than delighted if someone would tell me what I'm doing wrong with what should conceptually be quite simple.

millinghost - What I need is a solution that carries over to the renders. But that was a good idea! Thanks!

As to the attitude from bob1965, I'm sorry. I don't think what I'm asking for is unreasonable.

:lol: Oh that's not even remotely near attitude.

Could be that my response was engendered by a certain derogatory tone toward the people that provide content that has become overwhelmingly prevalent among the Poser communities.

I'm not quiet about the lens because it's a coming feature. I'm quiet because I don't understand how to explain it any better than I already have.

Before you try doing fancy tricks, you need to know how to render with raytracing and refraction. There are many things that can go wrong - too many for me to guess.

If you're getting white, then I'm going to guess that you ignored the bit about turning off the Diffuse_Value and Specular_Value.

Start with the camera rotations zero.

Place the one-sided square in front of the camera so it fills the viewport. This may require scaling it, positioning it, etc. Don't rotate the square or the camera.

Then parent the square to the camera.

Now you can move the camera as usual.

Hide the square to see through it. Or, turn it around so it's facing the other way, by using the Y Rotate set to 180 degrees. When it's facing the other way, it's invisible in preview and won't interfere. But you can't tell if it's filling the viewport that way. 

Once you have the square positioned correctly:

Diffuse_Value = 0

Specular_Value = 0

Reflect_Lite_Mult = OFF (NO MATTER WHAT, ALWAYS DO THIS)

Reflect_Kd_Mult = OFF (NO MATTER WHAT, ALWAYS DO THIS)

Plug a Refract node into Alternate_Diffuse. NOTHING ELSE PLUGGED ANYWHERE - NO SPHERE_MAP - NO COLOR NODES - NOTHING.

Alternate_Diffuse should be white.

Make the lens prop invisible to raytracing.

Make the lens prop not cast shadows.

Enable raytracing in render options.

Make sure you have at least one raytrace bounce in render options. I always have it set to 6, permanently, and decrease it if and ONLY if I have a reason to do so for test rendering speed. If you do not have multiple bouncing reflect/refract objects in your scene, then setting it to 6 costs NOTHING WHATSOEVER. But FORGETTING to set it because you don't think about raytracing means you will get Huh? What? Why doesn't this work?

On the Refract node, first time set the IOR  = 1. Render with and without the lens. If you did it right, you will not be able to tell any difference between the two renders. By "with and without" I mean make the lens prop invisible. Render. That's rendering without it. Look at what you see with the camera as is. Then make the lens visible and render again, with it in front of the camera. Things should look the same.

Then set the IOR = 2. Then 4. Then 10. Then 100. Then 1000. Observe how the lens changes the perspective with each change. This activity is called an etude. You are studying how things work and internalizing what you see = what you expect to see.

Do not do this with a complex scene of 1000 props and figures the first time. You're not going to learn anything and just get frustrated.

Practice using a lens in the wide open, with just a couple primitives to look at, or maybe the Andy bot, because he's much simpler to render than a fully textured V4.

---

Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)

> Quote - ... > > *IF* the zoom lens would work! I've entered everything the way the screen shots show it to be set up, and STILL all it renders for me is white. I'd be more than delighted if someone would tell me what I'm doing wrong with what should conceptually be quite simple....

For starters, try this:

1 - Load a duck into poser

2 - Load a primitive box and scale it up, so that the duck is in it (illustration A)

3 - In the material room, set the material of the box (not the duck) as in illustration B:

   - diffuse off

   - specular off

   - refraction node with fairly high Index_of_refraction (i used 50) into the refract input

4 - disable shadows for the box (in the properties panel)

If you render this (you have to enable raytracing), you should get 3 views of the duck on the sides of the box. If you still get white, something else is wrong (no raytracing, forget to disable the colors or so), otherwise: The images of the duck on the box practically do not depend on the camera distance, so you can move your camera directly in front of one side of the box and render the one side. Then you can of course cut away all other sides of the box leaving only one (by loading a single sided square instead of a box), that one would be the  lens. By twiddling with the Index_of_refraction value you can scale the image up or down.

EDIT: of course you can use other primitives or an Andy-Bot instead of the duck, as bagginsbill suggested, it is not that important, apart from that it is practically the same.

 

bagginsbill,

Question. Is the size of the square primitive important, other than that it should fill the camera view? Would making it double the size needed to fit the camera's angle of view, slow down rendering, or have an effect on quality?

Baginsbill,

First off: THANK YOU It works! I've printed this message out for future reference.

I'm sorry that I require the level of instruction that I do, and you are very kind to offer it in the very simple, step by step fashion that you did (which required no small bit of your time.) I sincerely appreciate it!

The problem I was having had to do with never picking up on the issue that you had to toggle the visible/invisible flags on the square. I had it all set up right up to that point. Now that I see it working, I understand what I was doing wrong.

Conceptually, I never did pick up on the idea that the effect would only be visible AT RENDER TIME. I kept trying to see it in preview mode, and now that I see how it works, I realize that was me having (yet another) Homer Simpson moment.

At any rate, thanks so much for the help (and to all the others who tried to help!) This solves my problem, and I'm ready to get back to what I enjoy: rendering.

Doug

Quote -      These models would become unmanageable trying to break them up to accomodate every conceivable camera angle:

I understand what you are saying. What I am suggesting is that in Miri's cottage, for example, there be at least 2 or 3 of those walls that could be hidden. Not all of them. You may be doing this... I do not own any of your models, and I may be preaching at the choir in your case. I was simply suggesting that modelers give the user SOME sort of option to "get back from" the characters in enclosed spaces.

And yes... BB's lens does solve the problem. Now that there is a difinitive tutorial on how to set it up, the transparent wall problem is solved in the simplist possible way. I'm happy with this solution, and can now get on down the road.

And for the record: I DO thank the modelers for what they do. Without you guys I would not be able to do what I do. I was simply commenting on a problem that, to me, should be easily soluable. I've learned more about the whys (if not the hows) of modeling from this thread than from any other I've read. Thanks to all who contributed meaningful content to it. This is how learning occurs.

I erred in assuming this would be in bill's book.  all that remains is for somebody to actually do a set of test renders with the square and a specific camera with a specific focal length in a specific room.  we've been referring to this method, but apparently only millighost actually did a mini-tute with imgs. other details may be needed, e.g. how see rendering camera's location.

the simplest method to allow view control in Poser for a room is to make the Room in 6 sections.

4 walls

ceiling

floor.

then you parent that all to the Floor and make it a figure. this allows you to hide walls as needed giving you the same flexibilty as a TV studio set. tis the method I use for my own models.

Airport security is a burden we must all shoulder. Do your part, and please grope yourself in advance.

Quote - the simplest method to allow view control in Poser for a room is to make the Room in 6 sections.

4 walls

ceiling

floor.

then you parent that all to the Floor and make it a figure. this allows you to hide walls as needed giving you the same flexibilty as a TV studio set. tis the method I use for my own models.

Yeppers. Dr Geep used that method to teach me about building rooms. Works great.

"

QUOTE "The problem I was having had to do with never picking up on the issue that you had to toggle the visible/invisible flags on the square. I had it all set up right up to that point. Now that I see it working, I understand what I was doing wrong.

Conceptually, I never did pick up on the idea that the effect would only be visible AT RENDER TIME. I kept trying to see it in preview mode, and now that I see how it works, I realize that was me having (yet another) Homer Simpson moment."

 

yep that was my problem too.   HOpefully it will work now.  Haven't tried it yet.

Well magnus, walls that can be switched on and off would be good, but also some of the rooms are furnished with cupboards etc, so the see through camera will be a bonus.

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

Attached Link: Miri's treehouse - 1280x800

     @BasicWiz:  Miri's treehouse has lots of openable -even deleteable- windows, plus two pair of French doors on opposite sides of the house.  I figured that allowed for plenty of camera positioning.  I didn't make the walls/floors/ceilings separate parts, because that would be 35 pieces -and 35 times the number of material zones- plus the issues of shadows and IDL.

     Ponder what happens to shadow and IDL effects when the wall behind the camera is missing.

     So, there are instances where I incorporate hideable walls, but it is not my first choice of a solution to the problem of getting the desired camera position.  Nonetheless, I'll give more consideration to separating pieces on future models.  ^^

Miri's treehouse:

Poser 12, in feet.  

OSes:  Win7Prox64, Win7Ultx64

Silo Pro 2.5.6 64bit, Vue Infinite 2014.7, Genetica 4.0 Studio, UV Mapper Pro, UV Layout Pro, PhotoImpact X3, GIF Animator 5

Has anyone tried reducing the main camera's scale from 100%?  By reducing the focal length to 35mm to minimize distortion, you can cover much more visual space in a close quarter room and still maintain full model visibility.

I guess I'm not doing something right, Hborre...

When I reduce the scale of the camera, I have to move it to get the same framing I had before. I still cannot get any more of the scene in if my camera is "back to the wall" regardless of the scale. Is there something else you are doing that I am not?

 

basic set-up, as per explanatory post above.  square parented to dollycam (maincam gives same results). AOK when square IOR=1; render of andy is same whether square is visible/invisible.  however, incr. IOR causes render to zoom in on andy.  decr. IOR causes zoom-out, but with strong distortion, until refracted part of render gradually shrinks to a small circle toward centre of render, like those peep-hole things they useta put in doors.

Miss Nancy:

I am not at home to experiment, but you plugged the Refract Node into the Refraction_Color channel on the Root Node. BB's instructions say "Plug a Refract node into Alternate_Diffuse. NOTHING ELSE PLUGGED ANYWHERE".

As to what difference that might possibly make, I cannot comment, other than to point out that the Alternate_Diffuse channel does not a refraction value.

lmk

Probably edited for spelling, grammer, punctuation, or typos.

 

o.k., thx fr tip, ken.  same render results with alt_diff as with refr. channel.   zooms in at big IOR, zooms out to peephole at small IOR.  poser 8 (final release).

Hmmm, well, BB has said that many of the Channels on the Root Node actually do the same thing, I guess this illustrates that.

lmk

Probably edited for spelling, grammer, punctuation, or typos.

yes, also does refraction when refr. node plugged into refl. color (white).

 

p.s.: does refraction in alt_spec, ambient/translucence value (greyscale) and partially in a transparency channel.

Quote -  

o.k., thx fr tip, ken.  same render results with alt_diff as with refr. channel.   zooms in at big IOR, zooms out to peephole at small IOR.  poser 8 (final release).

In order to use this as intended, you need to change the focal length of the camera.

As you drop the focal length, you get a wider field of view--"zooming out" basically--but increased distortion (as people should be aware of). Increasing the IOR of the "lens" corrects that distortion of everything in the field of view covered by the square, if you tweak it correctly.

The way this would work in the setting everyone wants to use it in, a small enclosed space (like an aircraft cabin, say), would be:

1. Set up the lens.

2. Position the camera in preview and adjust the focal length to get the framing you want. What you see will be distorted.

3. Tweak the IOR of the lens. The reason for the experimentation BB suggested on a simple scene is so you get a rough idea what IOR you should be looking at.

Then render, and adjust the IOR/size of square (as you decrease the focal length, depending on where the square is it might not cover the entire field of view anymore) as needed.

It might take some work to get it right, but at this point you aren't doing 3D any more, you're doing photography/filming. You're doing the exact same thing a photographer/cinematographer on a set would have to do, just without the immediate feedback they have because they can look through the lens and adjust in real time.

o.k., that seemed to work, keith, for no-distortion zoom-out using maincam/square in room with andy.  first try: F.L.*IOR=20 (in my example).  however, i just tried it again and it's not working out.  will hafta check something.

Quote - Use the Grouping Tool to make the walls separate objects and then you can hide a wall as needed.

cheers,
dr geep
;=]

I prefer this myself as you can drop and raise walls to render from different directions, Now you can even parent furniture to the wall so they don't get in way of the camera.

Sometimes you want additional roof and floor behind the dropped wall to give you more freedom in positioning the camera.

I prefer this myself as you can drop and raise walls to render from different directions, Now you can even parent furniture to the wall so they don't get in way of the camera.

NB: Sometimes you want additional roof and floor behind the dropped wall to give you more freedom in positioning the camera.

keith, on further testing, (F.L. * IOR = constant) maintains the same zoom factor as F.L. decr. and IOR incr., with incr. distortion on groundplane.  hence  I can't easily zoom out unless IOR = constant (or F.L. constant and IOR decr.), or else use some other equation/function to account for IOR = 1 ---> refr. = 0.  distortion may still incr. as IOR incr., even if some additive, power or log func. et al.  may need clamped UV-gradient plugged into IOR to correct for curvilinear distortion, but all this math is already making their eyes glaze over.  beyond the scope of this forum IMVHO.

Here's my method using the Poser Grouping Tool.

1: Select room prop.

2: Start Group editor

3: create two groups named Keep and Delete (Or whatever)

4: Add all polygons to Keep

5: Add Polygons to wall you wish to remove to Delete

6: Remove Delete polygons from Keep

7: Spawn Props

8: delete original prop. You now have the same prop with your chosen wall removed.

 

You can now turn visibility off on the wall you don't want to see, parent it to the part you want to keep and resave the prop with both parts for future use.

Is there anything wrong with making the lens prop absolutely huge so it will always cover the field of view?

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

 

preview image

pilots in cockpits, people driving cars, rooms that are modelled too small without wall bits separate, are exactly the sorts or things I want to do.

My grouping tool hardly works at all - something very wrong with it on my mac in poser pro 2010, it will hardly select anything. I'm lucky to get one little red thingy.

 

so I was really hoping this lens would work. Unfortunately I fiddle and fiddle but the renders always look more distorted than the preview no matter what IOR I set it at.

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

and here's a render through the lens:-

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

and here's the node

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

and here's the node

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

hborre I haven't tried your resize the camera trick yet.

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

hborre - I  just tried it and it is quick and easy and works quite well so far.

Thanks,

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

basicwiz - did you ever get the lens working well in a real scene?

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

Quote - and here's a render through the lens:-

Yes, but your IOR is 1.4. Apparently you didn't do the etude or you would know what is happening here.

---

Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)

well I tried all sorts of IORs including 100000 and lots in between.  I had 2 andies a different distances behind a square for a wall, and it seemed to work okay.  put it into a simple scene and not so.  the really high IORs do zoom in but they distort wall corners like crazy.  At least in my scene.

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

I even etuded negatively which turned the andies upside down.

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

the high IORs distorted my room even more than a small focal lenth did.  the walls actually bent.

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

I'm glad that worked for you. Esther.  I meant to get back with some renders but got sidetracked.  The camera reduction does cover more area although you would need to find the best location for optimal coverage.  I even mounted a 35mm wide angle, which gives the best view without actual distortion.  If you find a good vantage point on a 100% scale, reducing to between 60 & 50% should require no or very minimal relocation of the camera itself.

maybe this thing will work better in P9.  I tried a buttload of permutations and combinations (focal length*IOR, sep. of refr. plane from camera, hither, camera from 625% to 25%, et al.) in P8 with no success, meaning it works better (less distortion) for an interior shot with no refr. plane AFAICT.  :crying:

I wonder if BB has tried it in any actual poser model scenes.  When I tried it in a real scene it didn't work well at all sorts of IORs and focal lengths. I did make the square really big to see if it would help and it didn't.

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

Has anyone who's used the lense successfully considered making a freebie prop with the right material settings saved as a smart prop to a camera?

that would be good, but you know I followed all the instructions exactly.  well perhaps I should have put the square right on the camera with hardly any space.  and I still don't know if making the square huge detracts from anything or if it has to be exactly the size of the field of camera view.

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

it would be a really handy thing if it worked well.  Perhaps if it came with a python script to auto adjust the square size (if that is important) to the field of view.

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

I'd like to see a screenshot of just where the square should be in relationship to the camera itself. I mean, in a real camera shaped like that the front is actually inside those square shutters, but does that have any relationship to the true front of a virtual camera? I don't know.

given spherical distortion seen in renders using bill's flat refr. lens, perhaps spherical refr. lens would decrease said distortion. he was gonna do something in re: spherical surface rendering, but got tied up with the beta testing.  one of y'all could try it.  spherical distortion may vary as camera position varies from sphere centre.

Guys the one-sided square, perfectly aligned with the camera, very close, with the camera very wide, such as 5 mm, and the refract IOR = 1000000, there is ZERO distortion. You're not doing it right.

At such a high IOR, the square will be a photo in itself. It will show whatever is in front of it at actual size, projected on the square. If you want an entire picture of Andy, then the square must be at least as big as Andy.

Then the real camera is just taking a picture of a picture.

---

Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)

can square be bigger than needed?

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

I can confirm it works in orthographic mode, anyway.  can't get it to show the ground plane yet.  the square may need to encompass the entire camera view, to avoid confusing results at first. YMMV.

try rendering inside a proper scene that doesn't have much room

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

I had a pretty good idea how it was supposed to work, but I couldn't seem to produce a test render without huge pincushion distortion for some reason. Lack of patience on my part I think.

I was able to get the scene to render fine once BB posted actual numbers to use. This is with a focal length of 5mm and an IOR of 1,000,000.

Not a realistic scene, but there isn't a lot of room in the primitive cube I used.

I didn't make it obvious but the camera is rotated a bit in the horizontal plane, so any distortion of the tile pattern on the wall would be exaggerated.

well that looks good. I had given up but I might have to give it another try.

Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!

Can a prop be loaded as a smart prop already parented to a camera? If not, can this be accomplished through a Python  script?

Maybe some future version of Poser could include a camera already designed and set up to specifically do this?

lmk

Probably edited for spelling, grammer, punctuation, or typos.

Attached Link: http://sites.google.com/site/bagginsbill/free-stuff/genibl---ibl-generator

I managed to do the smart prop of my GenIBL tool to the dolly camera, including changing the dolly camera focal length. I don't remember how I did it, but if people want to figure it out from the GenIBL tool, feel free.

The GenIBL tool actually had to load two props - a lens, and a reflecting probe, both correctly positioned with the camera. (Looks like an anal probe - no kidding - the shape was not an option. The math determined the shape.)

And I suppose a camera feature could be added that does the same. Send a note to SM.

---

Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)

:)    BINGO!

I have some ideas, and one out of a hundred of my ideas actually (sort of) work.

lmk

Probably edited for spelling, grammer, punctuation, or typos.