"Unacceptable" patch layout

[R Reynolds]

So I do a test render of my new steam locomotive and I notice that something funky is happening on the half inch plate under the cab in the shadows (see image). The patch layout for that plate is legal and all the normals are pointing outward. When I built it, I was trying for the fewest patches possible. I took another swing at it without any regard for patch count and came up with something equivalent with only two more patches and a much nicer render. The only real differences are that on the offending plate there are three point patches and the three patches forming the rectangular edges of the central cutout are formed with four, dead end, two cp splines joining the two faces while on the improved plate they are made by continuing the face splines around the edge.

 

Perhaps the moral of the story is never use dead end splines but has anyone else had a similar problem and the patience to figure out if they cause the granularity?

[robcat2075]

the three patches forming the rectangular edges of the central cutout are formed with four, dead end, two cp splines joining the two faces

 

 

Interesting find. I guess it's a violation of the unwritten "four corners per CP" rule

 

What if those edge patches were independent of the faces and only appeared to share splines with he face patches?

[mtpeak2]

Are there any internal patches?

[R Reynolds]

Are there any internal patches?

Good guess, but no.

 

And here's another interesting glitch related to a hook this time. Once again all normals are aimed correctly. I've run into this sort of thing a number of times on this model but I've learned that just because you can use fewer patches doesn't mean you should. Especially in cases like this where I only saved myself a half dozen. False economy.

[HomeSlice]

Patch issues aside .... you certainly haven't lost your touch Another terrific model.

[robcat2075]

And here's another interesting glitch related to a hook this time.

 

What was the splining difference between the two sides? Or is the problem that there is no difference?

 

Great looking beast.

[agep]

I have seen similar render bugs like your first example using AO a few times. All times I had double layer of patches (accidentally hit E and extruded the model)

 

AWESOME model by the way

[itsjustme]

I have seen similar render bugs like your first example using AO a few times. All times I had double layer of patches (accidentally hit E and extruded the model)

 

AWESOME model by the way

 

Now that Stian mentions it, I'm reminded of a similar thing I had happen when using AO...I'll have to take a second look to see if there is a double layer in there. For some reason, I didn't connect the dots until AO was mentioned.

 

Incredible model, as usual, Rodger!

 

 

 

-----------------------------------

EDIT

-----------------------------------

 

Nosir, mine wasn't the result of a double-layer or internal patches...it could have been the material I had applied (it had some displacement) or just how close I had things nudged. I'll have to do some experiments late tonight to see if I can narrow it down...it only showed up in AO renders.

[robcat2075]

Roger, if you can make a test case, say, with just that one panel that shows the render problem in AO that would be a good AMReport thing.

[R Reynolds]

What was the splining difference between the two sides?

The patch layout to form both sides of the door are mirror images of each other. The algorithm correctly interpolated from the hook around the left corner but not on the right.

 

All times I had double layer of patches

Oh yeah, I've done that more than a few times! Its the first thing I check (grab a single cp and carefully move it out of the way to see if it has an invisible twin). Not this time.

 

that would be a good AMReport thing

You're probably correct but (full disclosure) all these renders are from v13. In 2009 when I tried opening my usual "outdoor sunny day" project in v15h, global illumination was not working as expected and since A:M is being supported by one programmer working for free in his spare time, glitches such as that send my brain into "make do, work around" mode. And since every one of these granular shadow problems can be solved with increased patch count, that's what I do. Yes, I know, I really should try a more recent v15 but I get easily distracted by building stuff.

 

Thanks for the kind words about the model. For the curious few, this page shows a comparison of the locomotive I did at the turn of the century and the more accurate version I'm currently working on.

[robcat2075]

that would be a good AMReport thing

You're probably correct but (full disclosure) all these renders are from v13. In 2009 when I tried opening my usual "outdoor sunny day" project in v15h, global illumination was not working as expected and since A:M is being supported by one programmer working for free in his spare time,

 

give it a quick test in v16 if you're up to date on the subscription and see what happens .

 

Or if you can post the test case I'll try it .

[jakerupert]

Hi Rodger,

 

You really are one terriffic modeller!

 

When I watch your fantastic work I feel that strange feeling creeping in , that like with so many of us solobattlers here,

we pass our time and years with modeling and don`t find the time for animating these fantastic beast

for instance. It`s a pity, but well, may be some day....

[Xtaz]

this happens with all newbie .....

 

... leaving jokes aside, have you tried to render it as an individual model ? Can you make problematic model available for download ?

 

EDIT : I remember now ... I had the same problem when I rendered a model that had one of its cp's hidden

[Shelton]

Rodger

 

Is that a Hudson class?

[R Reynolds]

Can you make problematic model available for download ?

Apparently not! I took just the offending components and some nearby parts to create the same shadows and positioned them identically in the same project with the other models turned off. They render perfectly, no strange granularity anywhere. This suggests there's some very specific interaction within the full locomotive. Since the loco is one of the few models I'm unwilling to share I guess I'll plead temporary insanity, thank everyone for their suggestions and tell you to go back about your business. Nothing to see here!

 

Is that a Hudson class

You're absolutely correct, a 4-6-4 hybrid that never existed in reality; the running gear is modeled after a Milwaukee F6 while the streamlining and cab are similar to the N&W Class J.

 

...we pass our time and years with modeling and don`t find the time for animating these fantastic beast...

Well, you can't run a railroad until you've built a railroad. And like any train trip, half the fun is getting to the destination. At least I have the action for one revolution of the wheels.

[jakerupert]

>This suggests there's some very specific interaction within the full locomotive.

 

 

Maybe worth a check....

[pixelplucker]

Amazing model!

 

I have had some issues in the past where the patches would be funky as you showed. I got in the habit of going pig tail crazy to make sure I have continuous splines in the right direction and then deleting them after. I always marvel at models like yours and others that do such fantastic mechanical models with splines. Models like that even put most poly and nurb guys to shame!

[animas3D]

Rodger,

 

First of all, incredible model on the locomotive. Very, very inspiring.

 

Secondly, I saw the images from the Penn Station model you were working on. Are you still working on it? Are you from New York?

 

I'm a native New Yorker and have lamented the destruction of the original Penn Station, a beautiful beaux-arts palace designed by famed architects McKim, Meade and White since I first learned about its existence. It was torn down, as I am sure you know, in the early 1960s in order to construct the unsightly Madison Square Garden which stands there today.

 

It was the public outcry from this ridiculous decision to destroy the original Penn Station which sparked the New York City architectural preservation movement. There are black and white films from the era showing New Yorkers shedding tears as they watched it being destroyed.

 

Very cool. I applaud you for resurrecting this piece of New York City history for us to see!

 

Cheers!

[R Reynolds]

...incredible model on the locomotive...

Thanks.

I saw the images from the Penn Station model...Are you from New York?

Nope, the closest I've come would be a trip to Newport, Rhode Island. But if you do any research into the glory days of railroading you can't help but run into pictures like the one attached of the stations' waiting room and who wouldn't want to experience such an over-the-top room? But since I can't ever visit such a cool building, the least I can do is build one to use in my own films.

Are you still working on it?

Well, not at the moment (you see there's this locomotive ) but it's still in the WIP stack. In the late summer of '09 I was doing test renders of the station's below street level taxi stand when I felt the need for a 50's style Checker cab to get the lighting right. By the winter I was getting somewhat frustrated trying to flatten the cab's back roof and rear quarter panels to apply the belt line checkerboard pattern decal when I got the final locomotive reference material for Christmas. Since then I've decided to try to concentrate on the star of my films before I get back to the supporting cast and sets. But fear not, my Penn Station set will be finished (eventually).

[John Bigboote]

Rodger- NYC was a great visit for me and Dawn, recently. I recommend seeing it. You can grab an AmTrak from Toledo which will zip you right into Penn Station, and Grand Central is not too far a walk from there... hotels can be had for around $200 a night. The food is a little expensive but quality makes up for it. Subways are confusing but get you where you want to go...cheep. There's SO much to see and do.

 

I think you would LOVE it.