Planetside Software Forums

New Erie Canal scene. This one is set in 1824 during the excavation of the "Deep Cut," through three miles of solid rock.

The rock piles and cranes eventually will extend all the way to the horizon. The dude next to the crane is temporary, just to show scale. He'll be replaced by teams of workers around the cranes and in the bottom of the channel. Horses will power the cranes.

The rock walls need more detail and the boundary between the rock layer and topsoil needs some TLC. Lighting is tricky because of the need to illuminate the bottom of the channel, which is about 9 meters deep.

This is my first experience with extensive lateral displacements, and boy has it been an education. The floor of the channel is a second planet, which made it much easier to manage its displacements (Y and Z) and those of the vertical walls (X).

I've set up a new website to document progress on these projects: https://40x4x28.com. Visitors welcome! :D

Looks good.

Very cool, I think it looks quite good already!

- Oshyan

Already impressive!

Very interesting new venue, and quite a task again (making all the workers especially). This is already looking great. Sky is really good too, I like it very much. Good luck, Steve.

Thanks, everyone! It's a much different environment than the last scene. The heavy overcast was added to open up the deep shadows, but I like the look of the clouds here – they establish the right mood.

Quote(making all the workers especially)

This will be the challenge, for sure. I'll be able to use populations in the distance, but not in the foreground.

Wow, impressive already! Looking forward to the next steps.

Love the lithograph images you found. 1500 workers and 100 cranes in a 3 mile stretch? !! You are making great use of many resources, watching as history is explained. Great documentation and authentication for this build. Interested to see how many workers, horses, cranes and rock piles you will be able to fit in one scene, but love the attempt.  :)

Thanks, Luvs. I am a bit apprehensive about all of the people that will be needed, but am hoping to build on the experience gained in the previous scene.

Started on the research for this one several months ago, with several visits to the site (about an hour's drive from here) and to local history archives. Fascinating story . . . they chipped through three miles of solid rock using hand drills, hammers, picks, and black powder. Difficult and dangerous work.

Time to start working on some navvies . . .

Next iteration. Added more rock piles and a few more cranes. More cranes are needed, but these are placeholders and will be replaced with modified versions to eliminate repetition. Foliage on either side (the canal line ran straight through an old-growth forest). More detail and water on the floor of the cut. Smoke from blasting in the distance.

Adding the water surfaces more than tripled the rendering time, so I may experiment with reflective surface layers there instead.

I've found that I have to use a very high MPD to eliminate noise in the stacked voronoi diffs used for the rock walls. I would like the rock walls to be more blocky, square, and am open to suggestions here.

QuoteI would like the rock walls to be more blocky, square, and am open to suggestions here.

There is an old thread called 'Rectangular Noise' , maybe you'll find some inspiration there.

Or another thread called 'Paved Road'.
Hetzen has  shared a pavement file also, but I can't remember the name of that thread.

Ambitious project again, good luck and keep it up.

Found those, thanks. Looks like some good reading there and I'll be spending some time with them.

Oh wow, this is *really* coming to life already! The overcast sky is nice too. If it looks this good now, I can't wait to see where it ends up. :)

- Oshyan

Very nice update! That water on the bottom really does it, but indeed, that might as well be reflective (+ darkening) for faster rendering.
I would make a mix of voronoi and squarish rock, and that can be done several ways; by Jon's function stacks, or a simple repeted image map (something like René also posted) with some squarish rock mask, or even simple blocks in greys, divided by dark lines, then warped and masked by PF.
The problem with the overcast weather and the line of rock piles is that they flow into eachother like one long heap, but once you added workers that will be broken up I guess.

QuoteVery nice update! That water on the bottom really does it, but indeed, that might as well be reflective (+ darkening) for faster rendering.

Thanks. I tried reflective shaders for the water but was not satisfied with the result. For one thing, the water surfaces need to be truly reflective (not just specular), and using a reflective shader with ray tracing enabled was just as slow. For another, the water needs to be transparent because the depths are so minimal – the underlying surfaces need to show. In the end I optimized the masks for the water shaders and adjusted the bias settings for the two water layers. (Inspired by some very helpful threads in this forum.) That seems to have helped the renderer sort things and has speeded everything up quite a bit.

QuoteI would make a mix of voronoi and squarish rock, and that can be done several ways; by Jon's function stacks, or a simple repeted image map (something like René also posted) with some squarish rock mask, or even simple blocks in greys, divided by dark lines, then warped and masked by PF.

Good suggestions. How would you apply an image map? With a camera? The two vertical walls complicate things a bit. I haven't tried to isolate and deal with them separately, as I imagine you would have to with projected masks. Redirects, for the same reason, don't work – a positive displacement on one side is a negative displacement on the other. (So far I've used lateral displacements with a compute normal.)

QuoteThe problem with the overcast weather and the line of rock piles is that they flow into eachother like one long heap, but once you added workers that will be broken up I guess.

Agreed. I've been gradually moving the sun to the south (the canal line angle is SSW), hoping more backlighting will separate the edges of the rock piles. Mixed success so far.

Since your canal is a straight line (presumably straight off north?), it's easy to separate each side by an SSS, or get position (in texture) + X to scaler. Then you can assign a vector displacement map or redirect to each side separately, displacing X and -X. Same with an image map or mix. As you wrote, but didn't try yet.
Btw, positive on one side and negative on the other doesn't matter if you use perlin or an image map where you can't really tell what's negative or positive, but that would be difficult for cracks.
Alternatively, use a mix of fake stones. But it would take some experimentation to get the right rock on sides and botttom, so indeed you need some clever masking, and without needing more than 1 compute terrain preferably.

Cool stuff! :)

Quote from: undefinedSince your canal is a straight line (presumably straight off north?), it's easy to separate each side by an SSS, or get position (in texture) + X to scaler. Then you can assign a vector displacement map or redirect to each side separately, displacing X and -X. Same with an image map or mix. As you wrote, but didn't try yet.

Yes, I realized last night how to approach this, and your suggestions give me some good ideas to experiment with. Thanks, Ulco.

The Z axis runs right down the center of the channel. The actual canal line runs SSW, so the sun's heading is offset to correct the angle. This keeps things simple when placing objects, and will help with the displacements, too.

Quote from: undefinedCool stuff! :)

Thanks!

Quote from: undefinedOh wow, this is *really* coming to life already! The overcast sky is nice too. If it looks this good now, I can't wait to see where it ends up. :)

Thank you, Oshyan! In the future (if you're so inclined) please check in from time to time with some C&C on this. Your thoughts and suggestions are always spot on.

Latest iteration. Many changes to the underlying structure of the image and other details since my last post. The primary concern here is accuracy. It now depicts the historical width and depth of the excavation and the walls more accurately show the rock structure.

Many technical issues here had to be ironed out. As I worked on it the scene became too unwieldy. There were several spheres for the rock and water layers, and multiple compute terrain nodes, and it took forever to render.

Then I decided to try Ulco's suggestion (in another thread) of building the scene with cubes to avoid lateral displacements. It now consists of three cubes, a plane, and a non-rendering planet to provide the atmosphere. All but one of the compute terrains were eliminated and I was able to reduce the MPD (which had been increased to fix artifacts in the lateral displacements). This solved nearly all of the problems I was having and now the scene renders in a reasonable time.

A displacement image is used for the rock walls but there are several layers of strata. The image is scaled differently for each, the layers are warped slightly, and more displacements are added to reduce repetitions. The texture needs a little more work but it is a pretty good representation of what the actual rock looks like.

The rock pile shapes are created with functions, covered with fake stones, then with a couple of layers of imported rock populations.

It's been an education. But I'm really ready to move on and start the modeling for this scene.

Glad that worked, and it looks fab!

Amazing! Cool, that you found a way to make it render faster.

Very good progress! I like the side walls and their details very much!

Looks good.

Thanks, guys. I appreciate all the help I've received here to get through the tough spots. This one's been a challenge for sure.

The next scene is going to be a simple, pastoral landscape with some mountains and a river . . . looking forward to that one.

First effort shading an object with Substance Painter. The application outputs PBR materials which were applied following Matt's instructions in this thread (https://planetside.co.uk/forums/index.php/topic,26658.0.html). Path traced rendering. This could really change the way I do shading.

Looks good.

Excavation scene is awesome! Congrats on that.

Wheelbarrow shaded superb. That creased inner edge with the grime.....oh my!  :)

That does look really nice. Lots of detail and roughing.of the textures. Though I wonder if the supports would span the whole bucket? What's supporting the pan? That's just my logic getting tripped. lol

Thanks, everyone. 

QuoteThough I wonder if the supports would span the whole bucket? What's supporting the pan? That's just my logic getting tripped.

The two long poles support the bucket. The model is based on a surviving example of the Brainard wheelbarrow (https://dp.la/exhibitions/erie-canal/building-canal/new-tools-designs?item=2240), which was invented during canal construction and was revolutionary - it was the first one with a curved bucket that made it easier to empty by upending the handles.

Great piece of equipment, Steve!

Quote from: sboerner on February 26, 2020, 11:30:25 AMFirst effort shading an object with Substance Painter. The application outputs PBR materials which were applied following Matt's instructions in this thread (https://planetside.co.uk/forums/index.php/topic,26658.0.html). Path traced rendering. This could really change the way I do shading.

Looking like Substance Painter is a program I need...nicely done.

Quote from: undefinedThat creased inner edge with the grime.....oh my!

Maybe I got a little carried away. This was, after all, brand-new in 1824.  :)

Quote from: undefinedGreat piece of equipment, Steve!

Thanks, Ulco. My hope is that practicing with objects will improve my human-shading skills.

Quote from: undefinedLooking like Substance Painter is a program I need...nicely done.

Thanks, Bobby. SB is a very deep and capable application. I can see where it may take some time to master but even beginners like me can use it to make usable textures. Very forgiving, completely non-destructive, fully procedural up to the point of exporting texture files. Lots of good training videos available. Great fun, too.

Quote from: sboerner on February 28, 2020, 10:03:02 AMMaybe I got a little carried away. This was, after all, brand-new in 1824. 

Then again, would they be staining and varnishing such a device to be beat up? First load of wet muddy rock will stain it.

QuoteThen again, would they be staining and varnishing such a device to be beat up? First load of wet muddy rock will stain it.

The only example I've seen, which actually dates from the 1830s, is just raw wood. I don't think they would have bothered to apply a finish, and maybe they weren't able to. This area was wilderness back then. Finished furniture items, for example, wouldn't have been generally available until the canal was available to ship them.

I set up this scene to test some new tool surfaces in a real environment. The textures are custom Substance PBR materials. The tools will be used in the deep cut scene and include a crowbar, sledgehammer, shovel, drill, pickax, and a couple of 25-pound kegs of DuPont blasting powder.

@Matt, this is one fine renderer you are building.

That looks great! Love the detail here. Get some more of that field in the shot with a skyline and you have a great piece of art here alone. A lot for the imagination here.

Exquisite!

Love it! Maybe a little suggestion: I remember my grandfather had loads of old stuff and tools in his shed, that looked very similar to yours. However the handles of those looked somehow a bit more greasy due to frequent use and the skin fat of the hands (I don't know the right words for this). So maybe some well distributed darker color and some subtle reflections might look even more natural?

Thanks, guys. I feel like I'm finally starting to get the hang of this.

QuoteSo maybe some well distributed darker color and some subtle reflections might look even more natural?

Yes, good idea. The handles would be worn smooth and stained from continuous use. Working on it . . .

A little more BST....blood, sweat and tears...great stuff!

Little masterpiece! 

Would you mind to make the barrels' textures a bit dirtier?

QuoteWould you mind to make the barrels' textures a bit dirtier?

Sure, but maybe just a bit. They probably didn't let these kegs sit around very long. They had to keep their powder dry.  :)


In the meantime I've updated the horse model that I made last year. The mesh was tweaked to improve the head and face, and to improve joint deformation. Then it was completely remapped, reshaded, and rerigged. The tail and mane will be added once the pose is finalized. (Hopefully this will give me a good base model, because most of the scenes I'm planning will need a few horses.)

It looks really good, expecially the head. It would be even better if you could add some muscle definition (either in geometry or bump), or a tendency of ribs. But I guess you may have that planned...

Yep, good point. That'll be the next iteration.  :D

Horse is looking really good! Though I feel it's material may be a little rough (physically; like the grooves), it doesn't look velvety enough. Maybe more roughness in specular would help that.

QuoteHorse is looking really good! Though I feel it's material may be a little rough (physically; like the grooves), it doesn't look velvety  enough. Maybe more roughness in specular would help that.

Duly noted, thanks.

Updated rendering of the still life with tools. @Hannes and @DocCharly65, is this a bit closer to what you were suggesting?

I think I even prefer the former state of the tools, as they now have an overall shine. I found these refs. But none is the same. But the handling end could have some more dirt, IMO.

Better indeed, Steve. Ulco's images shows exactly what I was thinking of.

Thanks for digging up those references, Ulco. Should have known better than to work from memory. The ax handle might be the one to go with.

Yes. A little darker towards the end and a little greasy.

OK, I think this is going to have to do it. Instead of using the small axe as a reference, I went with the larger one. (The dark stain on the small one just seemed too extreme.) Because the lighting and shadows in the still life obscure things a bit, separate renderings of the axe and sledgehammer are also attached.

Thanks @Dune, @Hannes, for the critiques. Sometimes I get lazy. :)

They are awesome now! Ready to sell too, I guess. I tend to get sloppy sometimes with my tools, so this is a wake-up call for me too.

Great!!!!!

Wow these are a great improvement, only critique is the sledgehammer may be more uniform due to the hand slide motion.

Edit: Looks even better on my new ultrawide IPS. Wow. So vibrant.

Great Improvement, I am impressed! :)

Looks great.

Great - love the texturing!

Child labor was a fact of life in the 19th century. This young fellow will be in charge of the horse that will be powering one of the cranes. Need to add sweat stains to the shirt and some other details. Suspenders will be added after he's posed.

Great modeling again. The boy really has a 19th century face too. Terrific. Wrinkles by shader?

Quote from: undefinedWrinkles by shader?

Thanks, and yes. That was a great tip you gave me a while back. Once he's posed there will be more folds and wrinkles in the mesh, of course.

I get more and more curious about the whole final :)

Test render of the horse and crane assembly. At this distance the horse could use more detail, but I'll wait to see how he looks in the main scene and will add it then if necessary. (Don't think it will be.) Planning to call this part done so I can move on to the next challenge!

Great! I think, if the horse will be not that close to the camera, it will be absolutely OK.

What a great creation! It will look surely good in the final!

Stunning detail, and I'm sure it will be quite enough. The boy is really good. I have one remark about the horse; it seems a bit elongated... but it could be the pose, I can't tell. Boy will need some dirt on his feet!

Thanks, everyone. This is going to be in the middle distance so I should re-export all of the shader maps at lower resolutions to save load time.

Ulco, the horse may be distorted slightly by the low angle and wide angle lens. (That's my story, anyway. :) )

This model was another test of sorts . . . I'm learning to use Maya's nCloth simulator for clothing. It's a little more work up front but eliminates almost all of the manual sculpting I was doing before. The folds and creases look pretty natural, I think.

Oh yes, they do!

The scene is going to include many human figures and I realized I needed a simple way to try out different poses. Thought about ordering an artist's mannequin but then realized I could make my own virtual version. Turned into a fun project and good way to spend the day yesterday. Now I can use this to pose and position all of the human figures in the scene, then replace them with real models as they're completed.

I'll post a rigged fbx version of this model in the sharing section in case anyone else might find it useful.

Does anyone know where I can find an image sequence of someone swinging a sledgehammer? I've looked but have come up empty-handed.

Something like this (https://www.youtube.com/watch?v=23YzsyXeSec)?

Yes. Perfect, thanks!

Latest iteration and the last one before adding the rest of the workers. I wanted to nail down the terrain before that step, since all the displacements had to be finalized first. Dust clouds are v2 clouds. Being a novice at clouds I started out using v3 but the many layers made the 3d view impossible to navigate and sent rendering times through the roof. Did a few searches on the forum and learned a thing or two. V2 works fine at this scale and actually renders nicer (and *much* more quickly).

By far the lion's share of my time on this scene was spent on the shoulder of the cut. That had to be right but creating the transition from topsoil to rock layer took a lot of experimenting. I ended up adding additional cube primitives for the topsoil. So now there are five cubes (two topsoil, two rock strata, plus the ledge for the towpath), with a plane for the floor. Pretty satisfied with the result.

Dune's grass, here and there, struggling to thrive in this environment.

Great that it all worked out so nicely. I love it. You only see the magnitude of all once you notice th worker and horse, so I'm looking forward to seeing the additional workers.

Two things I want to say; I liked your original clouds better, less patchiness. And I would add a slight color variation across the dug out heaps of stones, be it variation in the stones or dirt attached to them, maybe also a layer of dust across them (slope restricted)?

Thanks. I'm not sure what happened with the sky cloud layers. The cloud settings haven't changed in weeks. They've been moved to allow some direct sunlight through in the foreground, and the camera angle has changed. So maybe we're just seeing things that weren't in the frame before. Any suggestions on how to get rid of the patchiness? The cumulus layer is an Easy Cloud.

Adding a slope-constrained dust layer to my to-do list. Great suggestion.

Unbelievable good! Can't say anything else...

QuoteUnbelievable good! Can't say anything else...

Thanks, Doc!

Here is the first of many laborers who will be working in and along the edges of the cut. Spent the past few days setting up a process to make these, since there will be a lot of them. Planning between 50-60 figures in all, and hoping to keep duplication to a minimum and in the distance. (But we'll see.) This will be my task for the next few weeks.

Quote from: sboerner on April 21, 2020, 01:01:34 PM

Quote from: undefinedUnbelievable good! Can't say anything else...

Thanks, Doc!

Here is the first of many laborers who will be working in and along the edges of the cut. Spent the past few days setting up a process to make these, since there will be a lot of them. Planning between 50-60 figures in all, and hoping to keep duplication to a minimum and in the distance. (But we'll see.) This will be my task for the next few weeks.

He looks great. Clothes are done very well.

Great worker!
50-60? Wow!

Quite fantastic!  I missed the boy and horse, don't ask me how, just great! Some very fine work in this project! :)

Fantastic. You're in for quite a job! In the distance you an always use a few pops of workers.

One thing; isn't the handle of the spade a bit on the short side? His hand is very close to the blade.

Thanks, guys. This one took a couple of days but now that I have the process figured out (and bases for the shaders) the rest should go more quickly. The question is whether the clothes are authentic . . . checking with a local source on that and hoping they respond soon.

I put my mannequin to work and came up with a set of 20 poses. I may not need all of them. The same model can be used for more than one pose with different clothing and hair to mix things up. And models can be mirrored. So maybe 10-12 unique models will suffice. And like Ulco says, populations can be used in the distance to get to the final number of 50-60. The plan is to place the mannequins in the scene and then create a list of what's needed, then replace the mannequins as the models are built. I tend to over-prepare when trying new things.

QuoteOne thing; isn't the handle of the spade a bit on the short side? His hand is very close to the blade.

It does seem a bit short. It's based on one on display at a canal museum, though, so I didn't think I should adjust it. And the (more modern) spade in my garden shed is about the same size, maybe a little shorter.

Great poses, Steve! I'm looking forward to seeing the work progress.

Quote from: Dune on April 23, 2020, 01:26:00 AMI'm looking forward to seeing the work progress.

Me too!! Fascinating!

Image you might enjoy of men working Wabash IN in canal  :)

https://1drv.ms/u/s!Al2V1QPLhkdggcV4Z9mwqKjvYko07Q?e=MAZB08

Thanks, Luvs. That's a really great photo and it will help me figure out clothing and hats. Do you happen to know the source?

The Wabash & Erie was an insane project . . . took many years to finish and became the longest canal in the country. I don't think it was open very long, though, and it never made any money.

Quote from: sboerner on April 24, 2020, 12:04:26 PMThanks, Luvs. That's a really great photo and it will help me figure out clothing and hats. Do you happen to know the source?

The Wabash & Erie was an insane project . . . took many years to finish and became the longest canal in the country. I don't think it was open very long, though, and it never made any money.

That is actually a grab from Wabash County Museum archives.  All this stuff is my back yard. Indiana was so prone to spring flooding, log jams and such. Without the resources for labor back in the day. The area was still farming and industry was an infant. This image is actually the birth of the railway as men clear the canal to lay track. Yes, our rivers were not deep enough when not at flood stage.  Many many arteries feeding and draining.

Thanks, that helps. Interesting about running the railroad along the canal bed. Not the only place that happened. One of the reasons there are so few of the original canals left.

You all are probably getting a little tired of seeing this fellow, but he's my test model and I keep finding things to fix. This may be the last version. I stumbled across a 19th-century sewing manual that includes patterns for many everyday clothes, so his outfit now should be more in line with what working people wore in 1824: oversize pullover linen shirt, "poor boy" suspenders, straw hat. Not sure the trousers are 100 percent correct but these are good enough for this pose.

The clothes look really good. Great job.

Quote from: sboerner on April 24, 2020, 01:57:45 PMYou all are probably getting a little tired of seeing this fellow...

Nope!!!! ;)


Clothes indeed look great!! Maybe some more work on the shoes?

Will some have beards? I know they were pretty darn common then.

Never get tired of this, Steve, so keep on posting. That guy looks great, you get better at every stage. The folds in the blouse are superb. Did you do this with Marvelous Designer?

Thanks, Ulco. Yep, the clothing is done with Marvelous Designer. The brilliant thing is that it's pattern-based, so if you can find historical sources for patterns you are set. No excuses now . . .

And the folds? Are they inside MD?

The folds are all inside MD. The clothing is designed using a t-pose of the model. Once it's ready you can place a posed version of the same model (the vertices must match exactly) and then morph between the two. The simulation handles the clothing during the morph. You can also import animations but I haven't tried that yet.

Clothing textures and a light wrinkle map were added in Painter.

Thanks Steve. Never realized it also did such folds!

I had quite a bit of fun with this before the big C diagnosis, Nov of 2018.
https://planetside.co.uk/forums/index.php/topic,25713.0.html

I went through the trial version and put off buying due to graphics card issues. I would love to be rich enough to just use this program exclusively on one unit. I am a seamstress IRL, making women's and children's clothing, and simple men's designs. I made dresses, formals, suit jackets and trousers for women back in the day IRL. Then doll clothes for the infamous Cabbage Patchers....

I think, like you said there is a repository of old clothing patterns from way back afaik. Tell me if you get in a jam, and I can graph paper you something. It seems as if you have a good handle on this already.

My sis has made 50 quilts since retiring 2 years ago.....free motion stuff.

Thanks, Luvs. I don't know how I missed your thread when I was looking into MD earlier this year. My understanding is that the program is a spinoff and that the company specializes in CAD programs for clothing designers. So it's designed to be used by people (unlike me) who know what they are doing. I can see why you'd like it. The interface is very intuitive (perhaps the best I've seen in a 3D application) and I'm learning the terminology and figuring things out.

Thanks for the offer and I'm sure I'll have questions! Some of the patterns I've run across are completely inscrutable.

After disabling displacements and shading to speed up navigation in the 3D view, placing the mannequin placeholders went pretty quickly. I wasn't sure how this idea would work, but it looks like it will be a real time-saver. Now I know how many models are needed and exactly where they will go. (Allowing for adjustments once they are replaced with real models, of course.)

As of now there are 26 unique models. Seems like a lot but I think it's doable. Counting duplicates and populations there will be more than 100 figures in the finished scene.

I sometimes use landmarks for placeholders, but these are nicer. Only thing you have to think of is the (subtle?) displacement when placing the models. That would give them a different Y than in this render.
I had some 'trouble' lately when changing the geometry of the ground after having edited some pops. Some instances sank into ground then. One the one hand it would be great if an edited instance would automatically sit on ground again like the rest of the pop, but on other occasions it wouldn't. Checkbox (auto-sit) perhaps...

Wow, what a project!

Oh, Ulco is so right! Good luck with that! Hope no pop or obj vanishes somewhere...

QuoteI had some 'trouble' lately when changing the geometry of the ground after having edited some pops. Some instances sank into ground then. One the one hand it would be great if an edited instance would automatically sit on ground again like the rest of the pop, but on other occasions it wouldn't. Checkbox (auto-sit) perhaps...

So true. And because there is no planet surface in this scene, there are no compute terrain nodes to connect to. So population placement has been . . . interesting. In some cases attaching them to objects works, in others I've had to set them to float. Either way I'm adjusting the Y position manually, sometimes by instance. Don't worry! I won't let anyone sink into the surface.

Creating the human figures is going much more quickly than expected. Almost all of the poses are done, and about a third of the models are shaded and ready to place. It's become an assembly line. Just waiting for some reference material to arrive by mail to finish up some of the clothing.

Quote from: sboerner on May 03, 2020, 04:14:20 PMDon't worry! I won't let anyone sink into the surface

Thats good to know!! ;D


Cool worker with that hammer!

Awesome! I have to be honest and tell you I got motivated with this build and a couple of others currently. If you look close you will spot a hammer in my new ATV rover tool box. ha! 

Keep up this terrific stuff! :)

Wonderful guy! I love the shading on this, a bit of dirt, nice folds (yeah!) and bump.

Wonderful models sboerner. Why, even the "preview" shot above looks interesting on its own!

A few more. Almost done with the 20+ poses that will be in the scene. Very little of the detail here will be visible, because most of the workers will be in the distance behind clouds of dust. But this is good practice. Working with these models every day is giving me some good ideas as to how the meshes and rigging might be improved. I'm planning to do a new generation later this spring or summer and those will be better.

I like them even now as they look.

Really great models!
The facial expressions of the guys in the first image made me laugh though. They both look extremely drunk I'd say... ;D ;D ;D

QuoteThe facial expressions of the guys in the first image made me laugh though. They both look extremely drunk I'd say... ;D ;D ;D 

;D ;D

Well, they did in fact distribute whiskey rations several times a day, up to 24-32 ounces total for each worker.  :o I wouldn't want to be the guy holding the drill.

But yep, I'd really like to animate the eyes to avoid that thousand-yard-stare.

Fantastic characters! Do you do any post on the clothing after MD, like smoothing normals or sudividing? I start with particle distance 20, then switch to 10 or even 5, and remesh to squares. But even then some angles/polys show up quite hard in TG, and I take them through poseray at least to recalculate normals, or ZB to polish (you have to take care not to tear the seams though, so I mask these first).

Quote from: Hannes on May 06, 2020, 03:21:16 PMThe facial expressions of the guys in the first image made me laugh though. They both look extremely drunk I'd say... (https://www.planetside.co.uk/forums/Smileys/fugue/grin.png) (https://www.planetside.co.uk/forums/Smileys/fugue/grin.png) (https://www.planetside.co.uk/forums/Smileys/fugue/grin.png)

I thought the same ;D 

But how could they withstand the hard work without their whiskey rations ;D

Quote from: sboerner on May 06, 2020, 04:06:44 PMBut yep, I'd really like to animate the eyes to avoid that thousand-yard-stare.

In a few cases it's easy. I used a "technique" animating my Robert Patrick Terminators eyes to avoid the thousand-yard-stare by "animating" the texture maps:

(This is only a quick example I made in some minutes. And it only works if each eye is a separate object)

Eye_Directions.jpg

Thanks, guys.

Quote from: undefinedDo you do any post on the clothing after MD, like smoothing normals or sudividing? I start with particle distance 20, then switch to 10 or even 5, and remesh to squares. But even then some angles/polys show up quite hard in TG, and I take them through poseray at least to recalculate normals, or ZB to polish (you have to take care not to tear the seams though, so I mask these first).

No post work. In MD I work with a particle distance of 20 but export at 10, keeping the triangles. I experimented with converting to quads but they broke up the folds. Everything gets converted to triangles by the renderer anyway. For these models I'm trying to keep the weight down. But for a hero object I'd still try to avoid having to do post work, and might try a particle distance of 5 and export with fabric thickness enabled.

QuoteI used a "technique" animating my Robert Patrick Terminators eyes to avoid the thousand-yard-stare by "animating" the texture maps

Nice idea!

Thanks, Steve. I might leave it then as well. It's supposed to be better for post in ZB as that doesn't like triangles, that's all.

I understand completely — my first inclination was to output quads and then refine in Mudbox, in my case. But MD can produce high-density meshes so it seemed best to let it go. Are you using ZBrush to smooth the result, or to reshape it? MD has a sculpt tool but I haven't tried it yet.

I use it to smooth and move out any inconsistencies, but it's not the best procedure as seams are easily torn. I'll have a look out for the sculpt tool.

You might be able to merge the shells and weld the vertices along the seams. I'll take a look at that.

Quote from: sboerner on May 08, 2020, 12:04:19 PMYou might be able to merge the shells and weld the vertices along the seams. I'll take a look at that.

Or you could use ZRemesher and projection.

Yes, but then I'd have to redo UVmapping, and MD provides good straight maps by itself. Or can you project UV maps onto a new mesh? Sorry, btw, for it getting a bit offtopic.

The thing that sold me on MD is that it produces such clean, distortion-free UV maps. The UVs are not stretched around folds, the way they are when you use sculpting tools to model fabrics, so any applied textures appear the way real cloth would. I don't think I'd mess with the UVs. 

I'd recommend trying to do as much in MD as possible, adjusting the fabric parameters and simulation settings (gravity, etc.) to fix things if they don't seem right. You can pull the fabrics around to some extent while the simulator is running, but let it do the work.

Yeah, this is seriously off-topic but no point in starting a new thread.  :)

Quote from: Dune on May 09, 2020, 01:56:15 AMYes, but then I'd have to redo UVmapping, and MD provides good straight maps by itself. Or can you project UV maps onto a new mesh? Sorry, btw, for it getting a bit offtopic.

There are ways to do that.
But you can of course just weld the vertices in ZB, either on import or when already imported.
Forgot to mention that yesterday, sorry.
And, of course, sorry Steve.

No problem at all. We're all learning here.

Thanks for your input guys!

Couple more. The eyes are rigged now, and it makes a difference at this distance. It won't really matter in the current scene, but it may in later ones. The texture maps are only 1K so there isn't a whole lot of detail.

Just a few more to go.

Quote from: sboerner on May 10, 2020, 12:37:10 PM...and it makes a difference at this distance

Oh yes! Great!!

Great! especially the two on the left got a kind of personality... likable characters, surely having funny stories to tell after work having a beer... :)

Fantastic guys! Indeed the left ones are having some fun, I love those expressions. Like they're about to throw that stone into a pool with sunbathers around.

And I really like those clothes; Marvelous!!

Only one crit/remark; shouldn't they be closer together to carry a stone that big/heavy? Well, maybe I shouldn't have said that, as I know it's a pain to alter a gesture/pose even in a small way.

I really like the work that is being done on hands here with the fingers and arms precisely touching the objects.

They don't "sink" into the boulders and tools, neither do they "use the force" making wheelbarrows move a centimeter or two ahead. Something that, unfortunately, is rampant in many productions. A really good work.

Superb figures but yes, the forces acting on the figures aren't quite right.

Thanks, guys. I appreciate the comments about the poses. They'll be placed near the box of the crane as if they were getting ready to toss in the rock, so maybe they should be holding it higher than normal. But it doesn't look very heavy. If it looks wrong in the scene I'll rework it.

Thanks to Cyphyr and Dune (https://planetside.co.uk/forums/index.php/topic,27878.new.html#new) I was finally able to finish a high-resolution benchmark rendering. The scene is nearly finished. Two more figures need to be placed in the foreground (the contractor and chief engineer), and a layer of dust will be added to the rock piles.

I'll probably remove a few figures from the center group to reduce congestion. They're actually spaced several meters apart, but the perspective is compressing that.

Looks good.

Really great work!
One of these renders, you can watch again and again finding new details!

You managed to bring it all together really nicely. I love it. Indeed, so much detail to watch. Really like the tarpaulin too (I did one as well lately). Must have been heavy... and dangerous work, looking at the pile of rock upper right in the crop. The stones can easily roll down again. You could make an addition; (a) sturdy wooden barrier(s), like they do in avalanche regions, but that might not be historically correct.

Great work!

Looks really cool! Though I'd say the puddle are a little too dark. Not really muddy looking, but like a dark sludge or ink is settling in them. The characters all look really well done, and life-like in the scene, and the upper level of the trench looks really nice.

Absolutely amazing work, Steve!!! Agree with Ulco and Jordan, but it might be only nitpicking.

Here is the contractor. This figure is based on a real person, but one for which (conveniently) there are no surviving portraits. So I have some latitude. His name was Darius Comstock, a Quaker, pioneer, entrepreneur, inventor, abolitionist. He led a very interesting life.

Almost done. Just need one more figure to represent the chief engineer.

Quite the waistcoat! Such an interesting work. Good job! :)

Thanks! I've spent a lot of time lately going through old 19th-century tailor's guides and pattern books. Really confusing at first! But starting to make sense.

What a great figure again! Working with patterns is pretty hard indeed. Even trying to get the parts together in an orderly way is sometimes harder than (I) thought. Especially with more complex parts, it's pretty difficult to stitch the right ones together, and get the measurements in order. And when moving a point to give more leeway (say under the arms), you have to update all line lengths.
So I give you credit for your suit! And the rest.

Superb work, great attention to detail and excellent research too!

Quote from: mhaze on May 30, 2020, 03:58:24 AMSuperb work, great attention to detail and excellent research too!

I couldn't describe it better!
Great! :)

Thanks, everyone. The engineer is a couple rungs up the ladder, so he'll have a top hat and a nicer shirt.  :)

QuoteEspecially with more complex parts, it's pretty difficult to stitch the right ones together, and get the measurements in order. And when moving a point to give more leeway (say under the arms), you have to update all line lengths.

Yes, and often it's not easy to figure out which seam to tweak first. My respect for clothing designers, dressmakers and tailors has increased considerably since I've starting playing around with this. I keep a calculator open and usually work with measurements turned on. It's starting to get easier.

Congratulations, I just watched the whole TOPIC. It's a colossal amount of work and very well done. :D

Fantastic character and great tailoring!!!

Great look and Character!

The chief engineer, who will be checking up on the contractor's work.

This is the last human character. I want to add another horse or two and that will be all of the models.

Great! They really look old-fashioned, and it's not only the clothes.

Sweet ones again!

You two could print your figures as real 3D objects too. Would be very nice to look at i am sure :)

Thanks! That's an interesting idea, for sure. On the other hand, 2D is more than enough of a challenge . . .

This is probably the final version . . . time to call this one done.

Looking forward to the next one (actually I've started already). Compared to this one it should be a piece of cake!

This is absolutely amazing, Steve! So many details!!
However I have the impression, that the lighting looks a bit flat. Are there additional lightsources?

Very nice work.

Thanks, guys!

There is a second sun, very weak, 0.2 brightness (if I remember right), to fill in the shadows slightly. And in post I brought the highlights down to emphasize the cloudy sky. So the flat lighting is intentional, though I may have overdone it slightly. Duly noted!

Just my thoughts, I hope you don't mind: I have the impression, that increasing the environment light's strength looks better and more natural than adding additional lightsources, when you're rendering with the legacy renderer.
However, a killer image!!

Not at all! This is rendered with the path tracer, however. The environment light would not have the same effect, correct?

The scene is a lighting challenge because of the deep, deep shadow cast by the left wall of the cut. Freely changing the orientation of the sun is not an option because I want it to be historically accurate. The channel is on a N-S axis, facing south, so you will either get a heavy shadow on either side or the sun shining directly into the camera. I went for a late-morning sun in early summer to minimize the shadow, but it still needed some help or the poor fellows under the wall just disappeared.

Add to that the heavy overcast and several layers of dust clouds . . . this was not a pleasant place to work, so my goal was to make the atmosphere look thick and oppressive . . . almost like a storm is about to break.

Stunning work! I admire the amount of research and asset production you do before being able to render a final image. I must agree with Hannes, though understanding your vision. Maybe not lighting up certain areas but darkening others (local secondary cloud out of view) may have a good effect, turning it into a more gloomy working area.
The GI doesn't have effect when path tracing, afaik. so you could use the camera's exposure to lighten up certain areas. Or burn a hole in the sky for local lighting up.

I also look forward to seeing what you're up to next...

Quote from: Dune on June 20, 2020, 02:01:40 AMThe GI doesn't have effect when path tracing, afaik

GI does, but increasing the environment light's strength doesn't, when path tracing.

So, I'd love to see this scene rendered with the legacy renderer, and with a stronger environment light instead of additional lightsources. Even though this is kind of cheating as well, it might be worth a try.

Huh? This is GI, isn't it? That is what I meant, increasing this won't be taken into account by the path tracer. What GI do you mean?

I meant that the path tracer uses some sort of (brute force-) GI, and the only thing, you can't do is change the environment light's value. Maybe I'm totally wrong, but that's the way I understood it.

We're on the same level, as that's what I understand too. I just used a local (out of sight) slightly illuminated cloud hung above my scene to lighten up the shadows a tad in PT. Could have used a plane too, but cloud it became.

Yes, that should work as well.
I did something similar in my King Croc image.

As far as I know the environment light has no effect with the path tracer -- GI is provided by secondary rays. So instead of a fill light (either direct or indirect) maybe the best way to provide more shadow detail would be to increase the max ray depth from 5 to something higher. I haven't messed with this yet, and it will be interesting to see what difference that makes in render time. I'll disable the second sun and will run some tests to see what happens.

You're probably responding to some of the changes I made in post, too. I'm a long-time print guy and my instinct is to open up shadows (where you always get some dot gain on press) and reduce highlights (to make sure there is a dot). I'll try to use a lighter touch.

I do appreciate the feedback. When all is said and done I'm going to trust my own eye. But you two have produced many great images so I know I'd better listen!

Thanks for your kind words, Steve! :)  Looking forwards to your next steps.

Albedo of clouds plays a big role in your overcast shadows (whiter clouds "capture" more light and less shadows to muddle actual object shaodws). Also, upping your soft clip effect may help as well. Things to consider.

Interesting. One of the main problems with this scene is maintaining the density of the cloud cover. If the landscape is properly exposed the clouds are almost completely blown out. Are you suggesting that I should *decrease* the albedo of the cloud cover, to make the clouds darker to begin with?

The renderings are 32-bit exr, so changing soft clip shouldn't have any effect.

Quote from: sboerner on June 20, 2020, 04:22:39 PMInteresting. One of the main problems with this scene is maintaining the density of the cloud cover. If the landscape is properly exposed the clouds are almost completely blown out. Are you suggesting that I should *decrease* the albedo of the cloud cover, to make the clouds darker to begin with?

The renderings are 32-bit exr, so changing soft clip shouldn't have any effect.

The opposite. As than the clouds transmit more light to the surface more like a normal overcast day. This allows populations; people, buildings, and vegetation shadows to be more visible. Like it is looking into a forest or what not. Denser, darker clouds means more overall shadow on the ground, to hide other shadows, kinda muddling the realism.

I'm out on a walk now under overcast and rain, take this shot of the forest and a clearing.

Thanks, Jordan. That's given me some ideas.

Here is a new rendering with the fill light removed. I increased the density and coverage of the top cirrus layer to soften the sunlight, then increased the camera exposure value (now at +2.5 stops) to compensate. I also increased max rays to 8, which seems to improve the shadow detail at little cost in rendering time.

All of the cast shadows are much more diffuse, including the one cast by the left wall. The rendering has picked up some noise, no surprise there, so that will need to be dealt with. (AA is now at 9 and max paths at 25.)

I used a luminance mask to isolate the sky and bring it back into gamut, but otherwise there's very little post work here.

Fantastic! This is coming along fast. The lighting looks leagues better. The forest has it's deep shadows, the grass on the ground has it's blades more defined due to internal shadowing, and the wall shadow looks richer.

I agree. Much better to my taste!

Yes looks better indeed.

Smoke, dust, steam, clouds, reflection in puddles;... Tophats, waistcoats, overalls, caps and boots ...Wheelbarrows, hammers, chisels, sledges, carts, horses...Terragen has progressed to a new level with this project. A lifetime of struggle building this canal accomplished in a few short months...AWESOME render and project!

Thanks, everyone. I have to agree that the lighting is much improved. I guess I needed a push – lesson learned. Should now just be a matter now of fine-tuning the renderer.

(And thanks for the kind words, Luvs. It's been a good project, for sure.)

Just awake, but astonished by the difference. Great work, Steve!

Great work, indeed.
And nice improved lighting now.

Thanks – and I appreciate the constructive criticism. I agree that it's a big improvement.

Running tests now to optimize the sampling. Because of the shadows and cloud layers this one is going to need higher settings than I normally use. I need the machine for other tasks, too, so it will probably will take a few days.

I guess your reflections in the canal will take quite some time, in relation to the clouds. That's what I noticed yesterday in a PT render. Usually for (not too) wet mud and finely displaced ground I use non-RT reflections, which are fast and certainly good enough. No difference to RT-reflections, especially under vegetation and such.
But I believe that PT really calculates all as, well, paths (comparable to RT), so that takes far more time. Certainly if clouds are involved.
So maybe it's much faster to do a GI cache of the whole, render the top half with clouds and the lower half without, or a simple v2 overcast sky, instead of several deep v3 cloud layers. You might miss some cloud reflections in open pools, but on ground it won't make much difference. Theoretically that is. You may have problems with the sunlight though, you'd have to test it.
On the other hand... you can also take a short vacation :P

Quote from: Dune on June 23, 2020, 01:30:14 AMI guess your reflections in the canal will take quite some time, in relation to the clouds. That's what I noticed yesterday in a PT render. Usually for (not too) wet mud and finely displaced ground I use non-RT reflections, which are fast and certainly good enough. No difference to RT-reflections, especially under vegetation and such.
But I believe that PT really calculates all as, well, paths (comparable to RT), so that takes far more time. Certainly if clouds are involved.
So maybe it's much faster to do a GI cache of the whole, render the top half with clouds and the lower half without, or a simple v2 overcast sky, instead of several deep v3 cloud layers. You might miss some cloud reflections in open pools, but on ground it won't make much difference. Theoretically that is. You may have problems with the sunlight though, you'd have to test it.
On the other hand... you can also take a short vacation :P

Pretty sure PT ignores whether ray trace is checked or not, and it is technically ray traced, well path traced.

Yes, that's what I think, and that's why it's probably slow, especially with lots of small displacements. Would be cool if some sort of non-RT/PT reflections could be made in PT mode, that give a reasonable 'reflection' and be as fast as in deferred rendering.

Quote from: Dune on June 23, 2020, 05:16:14 AMYes, that's what I think, and that's why it's probably slow, especially with lots of small displacements. Would be cool if some sort of non-RT/PT reflections could be made in PT mode, that give a reasonable 'reflection' and be as fast as in deferred rendering.

These "reflection" paths shouldn't be much slower than the diffuse paths, and typically they will be faster if the roughness is above about 0.2. If you were to replace those with the reflection approximation from the Standard renderer it would wash out the detail in the shadows and make the image look a lot more like the Standard renderer, but you'd still be paying in render times for PT on the diffuse lighting. I doubt that would be a good quality/time ratio, and you might as well use the Standard renderer instead.

Ah, that's good to know, Matt. Thanks. All worries taken away  :)

The total render time was about 17 hours the last time I did a full frame at 3840x2400, which I'd consider working size for this image because of all the detail. I don't think that's bad at all. Since the image is tiled I can do sections overnight and keep the machine free during the day. So a couple of overnight sessions and it's done.

Right now I'm rendering small patches of the image at that resolution to find the optimization sweet spot. So custom sampling settings, etc., are animated over a series of small crops. Those run overnight too. It's a gradual process but it lets me move on to other projects during the day.

Revisited this scene to add the improved character models at right. One thing led to another and I ended up rebuilding the cut and replacing the rock walls. (Had some help with that. Thank you, Ulco.) Much happier with the result. A few other large- and small-scale changes. Tidied up the node network. Finding legacy -- non-PBR -- default shaders in the object networks was a bit of a surprise and a reminder of how much TG has changed in the last two years.

Great work.

Thanks, Kadri. Should have a new one to show in a couple weeks.

This is looking really, really good! I can't wait to see where this ends up, great work.

Thanks! And my bad, my previous post was not clear. This scene is finished for now. The "new one" is an entirely different scene, the Seneca River crossing that is posted in a different thread.