Daz 3D is part of 
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.Daz 3D is part of
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Here, IMO, is a more pertinent recap. A list of some of the challenges where I invited folks to spend a few minutes to investigate a little deeper:
1. Figure out how to take all the pertinent render passes that are provided by Carrara, and, using the appropriate Blend modes, build a complete “beauty pass” which contains all the components. It’s a great way to learn the basics of compositing, and blending modes.
On your marks….get set….(EDIT: Oh, wait, I did that one for everybody..)
2. So the challenge is for someone to do a little research on this cool effect, grab a photo from the web, and manually generate the appropriate grayscale “depth” image to convert your Carrara render into something that looks like a photo of a miniature scene. (EDIT: Oh, wait, I did that one too)
3. Another easy challenge for those who are interested is to check some trailers for some blockbuster feature films in recent years and see if you can spot the blue/orange color scheme.
4. Take your cellphone or whatever video camera you can get your hands on, make a short video of whatever you want, and use one of the free camera tracking apps to generate tracking info. Then use one of the free compositing apps to bring that all together. I believe Blender has a free match moving feature which does a nice job
5.I’d suggest you try rendering a simple scene with an object index pass in Carrara. Then bring that image into Photoshop (or better yet, a good compositing app) and select an object based on that ID grayscale. Once you have the object selected you can easily make a matte/mask to isolate that image so you work only on that object in your rendered image.
6. Also, note that the post effects are automatically given a Screen blending mode. Anyone know why ??
7.The Realistic Sky simulates the direct sunlight as well as the blue sky light bouncing around the scene and lighting up the objects, but there doesn’t seem to be a pass which isolates that sky light illumination so you can tweak it. Maybe (and I hope) I’m wrong…it would be great if someone could do a little research to see if it’s in one of the other passes.
Again, I don't think anyone has engaged in any of these, or shown any of their work based on their new skills obtained from this discussion. Yeah, a couple of exceptions I suppose, but I'm just trying to get a sense for how much real interest there is in all of this. When people get their hands, for example, on the latest unbiased renderer, they're off and running making any renders they can. I just don't see that level of interest here. Maybe I'm wrong, and people are hammering away with AE or PS or Nuke or Fusion as we speak. I'm just trying to get a sense for where everyone is on this.
And I was especially hoping that maybe Headwax and Wendy would put their intense dislike on hold for a bit and actually participate in this.
I know you guys peek... :) :) :)
Joe, you have made a wonderful thread, the challenges are very interesting.
This thread will be here for a long time to come (unless it get;s lost or daz dies or the net dies or you delete it etc)
Just because people haven't jumped at the chance of doing your challenges don't be dismayed.
Some of us are time poor, or (like myself) have other deadlines to meet, or perhaps they might be too shy to show off their work.
But look at in a temporal sense.
I think we should call this Joes' Layer Compositing Wiki or similar, and uprise you to the membership of the Carrara Hall of Fame.
Arise Sir Joe, you done good!
Edit: sorry just saw the part about putting my dislike on hold. I'll have to think about that one, I just checked my dislike list, it's in the back of my green velor hotpants where I keep it safe and warm, ahh there you are, number 452 on page 36, you might have to wait your turn I am afraid. I have to undislike the other 451 first .
Again, I know it's exceedingly difficult for some to comprehend, but I'm not looking for acclaim or statues erected in my honor. Those things are irrelevant to me, especially in this forum. Maybe others are looking for that and think that's important. Not me.
I'm merely hoping, as I always have here, that people would expand their horizons and come out of their hermetically sealed Carrara dungeons and try new things. Sure, some do, but for others the excuses are constant and apparently insurmountable. And, of course, very emotional.
If people can spend 45 minutes searching thru, and recapping, all past posts over a 2 week period just to try to discredit someone, but not spend 10 minutes trying to experiment with a new technique, it just makes me, again, lose hope that the folks here really are interested in learning and improving. That's my frustration. Emotions and personalities take center stage over skills and learning.
So, Head, does that mean you're not going to participate? :)
Ha ha Joe, I been called a lot of things but that!
Really, I dont have the time, I was supposed to write a big thingy for my agent two months ago and she is still waiting, I am banging it out now, I am supposed to be condensing three books into two pages, writing verbosely, is, as you know, easy.
Writing concisely is hard.
Aoart from that my roof leaks, my frangipani died, I have two shows coming up soon and only half the work done, the local church wants to show a huge Jesus print I made so that they can give a few sermons and relate to it so I have to find 1500 bucks tio get it framed, or a place that will print me out a cheep copy (last one cost 400 bucks) apart from that the surf isnt bad and I am getting fat by sitting here at the vdu
So, sorry, I am time poor.
The flame one looks good.
I have Ron's Flame brushes and it comes with a big pdf and I t hink it would fit in with your teaching adrmirably,
regards
from here
(Yes I am procrastinating in doing my other stuff)
I totally understand.
So maybe you can coax Wendy into participating? Oh, wait, you said you don't have a lot of time. Nevermind. :) :) :)
(just poking some fun at Wendy, since she's been so open with her dislike...)
Maybe I'll do a section on compositing cat images to melt a bit of her dislike. :) :) :)
Wendy and I are having coffee and High Tea tomorrow at a fund raiser for the International Society For Mistreating the Daz Cat (yes, they are all for it) so I will put in a good word for you, or at the very least, have a cupcake for you - only one as I am having trouble fitting into my hotpants already.....
Okay Joe, I said I'd spend time on this last weekend and that didn't happen, but here's challenge 1. I reassembled the render passes from my entry for last month's Carrara challenge in Fusion. The image on the rights from the Fusion screen grab is the composited image while the image on the left is the original full render from Carrara. I screen grabbed the flow area too so you can see how it lined up. I found out too late that I left out at least one render pass that should have held the flow effects on the axe, the lightning streak and the green glow of his eyes, but apart from that it's fairly close. I found need to use the brightness and contrast tool to boost a few of the passes, I brought in a depth blur tool to try to tweak the depth of field in fusion but gained no purchase with that, it really had no effect, note though that my knowledge of how fusion works is limited, I'm fairly new to working with the whole node system so I understand a few things and I'm clueless about others, what I've been doing has been to get the best I can with the tools I understand while slowly picking up with the rest. Lastly, since I was in a compositing software anyway, I threw in a film grain tool and saved out the last image which is perhaps more appropriate to the period of the myth involved.
Next I would want to spend a little time investigating the object index render pass you spoke about before taking a look at the second challenge in your list. Hope I can do that before the weekend runs out, I'll have to pace myself though, so apologies if the turn around time is not so fast.
The passes in case you're wondering
Last ones.
WOW !!!!!! Excellent DADA_universe !!!!!
I'm very impressed.
I'm especially curious how you brought each render pass into Fusion. Did you save each pass out of PS as a .jpg or something like that, and then bring each into Fusion with a separate Loader?
Carrara rendered out the passes as jpegs, I just shift selected all of them from the folder and dragged into Fusion's flow area. That tends to automatically line everything up and connect them with merges. I disconnected everything and then started to compose it as I thought fit.
Oh, okay....just be careful rendering out passes in .jpg format. When you start compositing with already-compressed images, like JPEGS, you're at a disadvantage. Things can get messy real quick. It's better, and cleaner, if you can used an uncompressed format. And if all the passes are layers (or channels) of a single image file, it makes things easier.
Food for thought...
Okay, thanks, I'll keep that in mind, I'm sure I've run into trouble in the past doing just that without realizing what went wrong.
Oh, I've certainly been practicing - wouldn't have been full of praise if I hadn't, because then I wouldn't have seen the results for myself. I'm just not that much given to showing stuff in public.
One interesting thing I discovered is that when I wanted to lighten the sky and used the atmospheric pass for that, it lightened the whole picture, which I didn't want. So, I deleted the portion of the layer below the horizon and that gave me the effect I wanted.
This is what I meant about educating - promoting understanding - as opposed to showing people how to imitate. From the understanding, I was able to work something out for myself without having to be spoon-fed:)
Not only that, but I've been following some related tuts for GIMP. There is nothing I could find on compositing layer passes, but plenty on using channels and making your own "layer passes" from photographs which do not have layers. Now that gets me to wondering why cameras don't offer layer passes! Maybe the very expensive ones do?
A question - when rendering in Carrara PSD, do you set it to include Alpha?
Well, do you WANT an alpha channel? :) :) :)
OK - let me rephrase that...is there any advantage/disadvantage when editing the passes to setting Alpha?
Well, yeah....when you don't want your background to be visible, for example. The Alpha will make it transparent. So say you knew you were gonna composite the rendered image on top of some other background in post, you would render the alpha to make the rendered background transparent. That's just one of many reasons to do it.
And it's also what all the cool people do. :) :) :)
Now, one area that I forgot was available as a render pass in Carrara is the Normals pass. And for serious compositing it can be extremely useful.
What do you use it for? Well, let's say you get a 3D render and you want to tweak the lighting on an object a bit. Yes, I'm talking about tweaking the lighting on a 3D object after it has already been rendered into a 2D image or sequence. And you're tweaking it as if you were back in the 3D app, moving a light around the scene. Go no, right? Well, yeah, you can do that, though there are some limitations. But for reasonable or minor tweaks it can be a great way to go. Keep in mind this can also mean lighting color changes, but unfortunately you aren't going to affect shadows or anything like that. But still it is used a lot and can save a lot of time.
Now, of course if you don't like the lighting on an object you could re-render it, but in a professional situation that can be very time consuming. And if you can do it in post then it will be a lot quicker.
First, what are normals? I think we all have a general idea, but for those who don't it's basically an arrow that points in a direction that is "normal" (another word for perpendicular) to the surface of the object. It provides information about the surface of the 3D object. It tells you, for each point on the surface of the object, where in 3D space that point of the object is facing.
So the reason you can re-light that object in your compositing app is that the normals tell your 2D light if the surface is facing the light, and should be lit, or not facing the light at that position, and therefore won't be lit. A bit hard to explain, but if you still don't understand then just open your modelling app and select the option to show normals and you'll see what they are. Maybe later I'll give a more graphic explanation.
Anyway, what the normals pass does is provide information to the 2D compositing application about how the surface of the 3D object is in the 3D scene. It tells the 2D app the coordinates of where the surface arrow is pointing. And as we said before, those values are encoded as grayscale values in the render passes. And the normals pass provides 3 separate grayscale images. One has the x direction, one the Y direction, and one the Z direction.
So you bring that information into your compositing app, and your compositing app will have some sort of feature to simulate a light in your 2D image. And that light will use the normals info to figure how to re-light the object as if it was in the 3D scene.
Not sure if there's much interest in this, and if there is I'll try to spend some time later on showing an example of how this works. Though it gets a bit complicated depending on the compositing app, so I'm not sure it will be of much use.
Below is the Normal Z pass for the scene I've been using. As you can see, for points on the objects which are facing in the +Z direction (UP), those points are white. And for those facing the -Z, those points are black. And for those in-between points, those are shades of gray.
So if you place a 2D light at top dead center of the 2D image (+Z axis), the light will light all the white areas, none of the black areas (cuz they aren't facing the light) and some of the gray areas.
Of course, the other two normals passes, X and Y, are similar, but on the corresponding axes.
By the way, I recall that in Photoshop you can also simulate the 2D light and it will use the Normals channels. I believe it's the Render/Lighting Effects feature.
Not sure how many are familiar with AE or Nuke or Fusion, but there are also similar features in those. Like I say, it is used a LOT in professional compositing because it's so easy to do for minor tweaks in highlights or whatever you want.
I suppose another challenge would be to experiment with the normals "re-lighting" in the compositing app of your choice and post it here. Should be pretty quick and easy in PS at least.
He he... very interesting !
And one other concept to keep in mind is a bit more complicated, but extremely useful.
And that is the concept of moving the layer or channel grayscale data we've been discussing into the RGB channels of your composite layers. That opens up a lot of features which work on RGB channels to some of your render passes.
For example, if you copy the X, Y, and Z normal channel info into the RGB channels of a render layer, you can now adjust colors and effects based upon direction. Like if you want a blue tint from the +Z direction, you use the Z normal information in the color channel and make adjustments.
It's getting a bit late for me so I'm not being real clear, and it's a bit complicated if you don't use graphics, and I'm too tired to do that right now, but it's an extremely powerful and often used procedure.
Well, in case anyone is getting something out of this I'll keep rambling... :)
One issue that was raised and is very important in the compositing process is the basic question of how you transfer renders to the compositing program. Do I use JPEG or PSD or Cineon or TIFF or TARGA when I render out of Carrara or whatever program?
Well, unfortunately, it depends. Since you probably won't be working on a scan of the 35mm film from a professional movie camera, you won't care much about Cineon/DPX files, even though that's a common format for many professional compositors. Oh, wait, I forgot that Carrara doesn't offer Cineon/DPX.
Anyway, the choice depends a bit on where the image is going. Now you can choose .psd if you're going into Photoshop. And that's good in that it provides the layers and channels that were rendered as passes in Carrara all in the same image file. And when you open it in PS you can easily see what's there and where everything is.
On the other hand, when you get into serious compositing it can actually make things more confusing. Why? Well, when you're new to a compositing app, the fact that the layers and channels are "hidden" inside the single image file, rather than handled as separate files for each pass, can get a bit confusing. Especially with a node-based editor. You have to "know" what the different features are doing and which layers they are working on.
So I would suggest you choose a render file format which gives you separate images for each pass. Yeah, you have to save them separately, and that can get messy keeping track of the files, but IMO it might be easier for you in the long run.
So what format? Well, as I said, JPEG is a lossy, compressed format, and not really designed for serious compositing. But it may work fine for you. Personally, I'd toss a coin and choose from either Targa or TIFF or even PNG.
Now, I'm sure this will start a big argument over which of those is 10% better than the other, and that's fine if it really matters to you. But I suspect for most here it's pretty much irrelevant. So today my coin toss came up with Targa. It may or may not work for you, depending on where you're taking the images, but that's up to you.
Now if we jump back to the discussion of Normal passes, and render the same scene in a .tga file, we get a normals pass that looks like the image below. Why the colors?
Well, because the normals info has been placed in the R,G, and B channels of the image. And that makes it look all RGB-ish.
The good thing is when you load that into your node-based compositor, you can see right away in the thumbnail what type of pass it is. And if you do all of your passes like that, you have a nice stack of images representing the various passes.
And below is an image that is the result of applying a normals pass in a compositing program, and simulating a blue light shining from above on the image I've been using.
The normals lighting that is driving it is on the left, and the result on the right
Joe, thanks so much for all the detailed explanations and examples you have provided.
Inspired by your examples, I tried my hand at trying to "improve" a simple image, in this case a simple landscape. I used a preset (Rolling Hills), and just turned off the cloud layer. The Realistic Sky editor gives some...okay results, but there must be a better way without resorting to expensive rendering settings. Here is what I tried:
The first image is the regular Carrara render straight from the camera, no adjustments. I then turned off the Realistic Sky, used a single Distant light and rendered out some passes, including a custom pass for Diffuse Shadows, which Carrara doesn't have? Anyway, to achieve that I just turned all the colors in the terrain shader white, turned on a white backdrop and rendered that. I have no idea if that is "correct", but I was able to use it as a mask to control the color of the shadows. I then used a depth matte to control mixing the colored shadows in with the render. I also experimented with different blending modes for those shadows, deciding on Overlay.
I used the depth matte several times to drive layers with various color correcting effects, especially to lighten, desaturate and turn more "blue" as the landscape recedes. For my Depth pass, I experimented with Distance Fog, setting the start radius to 0 and then extent radius to the edge of my landscape. It looked similar to the Depth pass from the render room, so if anyone has trouble getting a decent Depth pass you can try the fog trick - you can turn off the lights and it will render fine. Also used the same White landscape shader as my fake Diffuse Shadow pass. Finally, I dropped in a simple gradient for the sky.
Not sure how improved it is...probably needs more of a suggestion of an atmosphere. I need to quit this for now, but any suggestions about what direction to take to improve would be greatly appreciated!
This thread is very informative, thanks again
A few more pictures...I remember reading somewhere, maybe here, about creating a "height" mask for fog...got me thinking and I tried using a Mixer in the color channel with black and white and using Elevation as a Blender. I can't take credit for any of these ideas.
First image is no fog/atmosphere
Second image is the render of the "fog height" mask, with a screen grab of the shader
Third image is that "fog height" mask, set to screen in After Effects...no adjustments
Fourth image is "fog height" mask driven by the depth matte, with a Curve effects on the depth matte to "dial back" some of the fog in the foreground.
Sorry for my scatter brained explanations. The post work is done in After Effects.
Very nice work on the fog effect.
Thanks also to Joe for explaining how to use a Normal pass. Not sure if my CS2 version of PS, knows what they are, but it could be something to look into.
DesertDude, very nice.
The key to most of this is just what you're doing...thinking. Trying to figure what you want, and then trying to figure how to separate that into elements and arrange the tools to work on those elements to get what you want.
And one problem is that there are so many tools and options that it's a bit overwhelming. But once you get the mindset to break your image into elements and work on those, you've won half the battle.
Congratulations for being brave enough to engage. :) :)
You might want to consider downloading one of the free compositing apps out there if you want to do compositing work on stuff like this. PS is fine, but it's far more limited that the other apps.
I'm thinking you might want to try the node compositor in Blender (free) or Fusion (free) or Nuke (soon to be free).