Oh, what a give away!
marble
Posts: 7,500
Just wondering what, for you, is the absolute give-away that an image is 3D, as in DAZ Studio compsed and rendered? For all the sophistication of lighting, PBR, etc., there are certain things which are instant tell-tales.
The first is razor sharp straight lines. The edges of walls, for example are usually so unrealistic.
Second is clothing. There is no way that real skirt pleats and folds follow a curve from hip to thigh when sitting. Yet 3D skirts, unless they have been draped as dynamic cloth, always do. Pants never bend well either.
I'll also add hair although some is better than others. When I can see the ribbon structure, I'm instantly put off and so many of the hair styles on sale here are like that.
The 3D humans are becoming very realistic. It is a pity that their environments and add-ons are not so advanced.
Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
Eyes. 95% of the time, if I can see the eyes, I know it's a CG render, even if everything else is "perfect"
Also hands/arms not quite lying against a clothed body, since the smoothing mods want to wrap that clothing around the arm/hand.
The whole image being very crisp and clean. I know HD/High shutter speeds are a real world thing, but it tends to make a CG image look like a photograph of a diorama (at least to my eyes.)
Not directly DS related, but I was recently at a friend's house and they had Jurassic Park on their HD screen - the Raptors being CGI was far more obvious than it was on my comparatively smaller, non-HD screen.
Have you seen the Star Wars spin-off, Rogue One? They used a CGI Peter Cushing and, with all their skills and technology, it looked like a CGI Peter Cushing.
Yeah. I was calling him "Cut Scene Cushing" because it looked like animation from a video game cut scene. A very good one, to be sure, but still quite obvious.
It is a philosophical question as much as it is a technical one from my perspective. Since the technology is capable of pseudophotorealism, there tends to be a preoccupation with that nowadays that I personally find somewhat bewildering. Nobody would look at a traditional oil painting and expect it to look like anything other than what it is. Nobody would go into a gallery and examine such a painting with a magnifying glass to see if it stood up to some sort of photoreal scrutiny. Apart from the fact that you would undoubtedly be immediately shown the door, the paintings are not meant to be seen in such a way - they are meant to be viewed from a distance in their entirety, not examined one little piece at a time. Unfortunately, in a way, computer technology allows us to do just that with digital images. If you spot a nice render, do you zoom in? Do you stop at 100%, the level at which presumably the image was intended to be viewed, or do you go further in order to apply some test of photoreality?
As I said, the technology has advanced to the point already where simulating the real world in rendered images is largely possible, but that does not necessarily mean that that should be a universal goal or some kind of holy grail. Anyone wishing to achieve what you have described certainly has the tools available to do that if that is what they want, whether by modifying an existing product, or making their own or editing in postwork. Out-of-the-box products with the kind of "real world" detail you describe may have their uses and there may even be a suitable market for them. But there is a cost. Such details take more time and the price charged will reflect that. Not everyone may be willing to pay the price. Such products also tend to be heavy on system resources in use, and not everyone would be happy having their systems brought to their virtual knees by them. Only a relatively small number of potential customers would have machines or be prepared to purchase machines that are capable of dealing with whatever is thrown at them. There are already a few products sold here that some have said are problematic in this regard.
I guess what I am saying is essentially that the once-predicted widespread use of flying cars never materialized not because of a lack of technology, but because such vehicles would be prohibitively expensive and impractical for most uses. Most of us are content to keep four wheels on the ground, and that applies figuratively to our use of computer graphics as well as literally to our use of automobiles.
If enough people demanded the kind of detail that was envisioned, someone would make it. Today. Otherwise, probably not, and anyone wishing that will be left to do it for themselves. Which is an option.
I'm not advocating one thing or another. For myself, if I want photorealism, I'll grab my camera and take a picture. Guaranteed photorealism every time. If I want to create an image using virtual content in 3D, I don't expect anyone to mistake it for anything else. I would prefer to have it viewed purely from the perspective of whether anyone liked it or not based upon what it is and not some arbitrary criteria amounting to "I can tell that it is not a photograph". Um, duh, yeah.
OK, but some small changes could be made. "Roughing-up" wall edges, for example. Or improving the way conforming clothing drapes. I'm assuming these things are not out of reach for the hobbyist but perhaps you are right and they might add significant cost.
As for subjectivity, while I don't mind someone knowing I've used a CGI programme, I want it to look as natural as possible. I've just bought the VWD dynamic draping plugin for that very reason - because I'm tired of sending dresses back to Blender to morph the drape.
Actually, I rarely share my images anywhere (they are poor in comparison with many I've seen). The hobby is for my own pleasure but I find I get quite demanding of myself and the software.
For me the two quickest give-aways are the clothing and the hair. Second is the sharpness of the image. Almost invariably the images that make it up to the DAZ Gallery favorites are blurred and bloomed in post-process.
Looking at real photographs of real people, especially by professional protographers those photos have flattering blur in them. That's what I was telling my friend when he kept buying digital cameras with higher resolutions - that his was was not going to let him take pictures of her anymore if he kept it up. Those high res shots are unflattering to even young people and bring out details that people don't notice even if you're standing face and face unless your are consciously looking for them and focusing your eyes on those features. Eyes don't work like hi-res digital cameras.
Totally agree with you there. I love photorealism, but I love really good 3D art even better. It can look like a painting or a photo, as long as it looks good, which of course, is a matter of opinion. I got into 3D art to do book covers after not being able to find what I wanted in stock photos. Many graphic designers are taking photos and making them look more painted for book covers. They're actually taking out some of the photorealism.
Approaching it from the 3D side of things gives far more options. I don't have to find a model or picture that looks like my character, I can create one, often for less than it would cost for buy the photos or hire models. Yes, there's still a massive investment of time, energy and some money involved, but in the end, I have something I can use over and over again.
I love 3D for what it is and what it can do. And every day I see people using it in new ways that amaze me.
Hair, weird AO, and odd depth of field... Usually... Mostly hair.
I certainly don't have any problem whatsoever with anyone pursuing realism in their renders, and I wasn't necessarily advocating anything. Variety is the spice of life. Using the painting analogy, there are or were artists (Ken Danby comes to mind) that used realism very successfully in their work, but that doesn't mean that it is something that should be universally the goal of every artist. Different strokes, as they used to say.
In addition to what I mentioned about zooming in, what nonesuch said about what a camera actually captures is true. When we are looking at photographs or even renders, we have the luxury of examining them in every possible detail. They represent a moment frozen in time in a sense, and our eyes and minds don't percieve the world around us in that way. It isn't a perfect analogy, but I liken the ways our eyes percieve things to a motion picture camera, rather than a still camera capturing a particular subject at a particular point in time from a particular point of view. Our eyes are recording continuously, adjusting focus, adjusting to light, moving around, etc.. What our eyes are capable of seeing, versus what would be recorded if our brains processed what we saw as a series of still images, would not be the same. Take HDRI as an example. HDRI photographs are based upon what the human eye is capable of seeing, by compositing photographs taken with different camera settings. Fair enough. But are such composited images really showing what we see when looking at that subject at a precise instant in time? Or do they show what we would be capable of seeing if we stood there long enough to make the movie? If you are standing in a darkened room looking out through an open doorway on a bright, sunlit day, what do you see? If you are focused on what lies through the doorway, your eyes adjust to the bright light and focus on the objects visible in the distance. If you suddenly shift your gaze to the darkened corner of the room to one side, you will momentarily be unable to make anything out clearly until your eyes adjust to the reduced light levels and change focus to any objects or details that are much closer. If you took a picture of the room with the open doorway in the centre, what should it show? An HDRI image would repeat the shot over and over with different settings, some showing details at a distance, others closer and compensating for the lighting gradient. What our eyes are capable of seeing. But, since a photograph, or a render, is supposedly capturing a moment in time, is HDRI realistic? Does it really show what we see at a given moment in time? The problem is, as I have said, we can examine a photo or a render as long as we want like we would the actual scene in real life and expect it to show what we would be capable of seeing through the doorway and in the darkened corners. So what is real? Blury, out-of-focus objects in the distance and clear in-focus objects closer to us, or vice-versa? A clear view of what lies beyond the doorway outside in full sunlight and murky darkness within the recesses of the room, or vice-versa? As I said, it tends to become something of a philosophical question. And in the end, it really comes down to what you like and what you want, rather than a question of reality perse.
Whilst razer sharp wall edges are a give away on many images, to be fair, a lot of modern models no longer suffer this. Indeed that is one of the things I look for in the promos of an enviroment prop when I am considering buying it. Well modelled buildings will have smoother, uneven edges, and anywhere you see two large polys meet up at 90% is an indication of poor modelling, only really acceptable for things that appear in the far background.
Since you mention sets and so on, presumably we're not talking about the HD portraits that I've seen that are pretty indistinguishable from a staged photograph, so I'll toss in 'poses' and 'lack of diversity' in models, environments, body types, skin blemishes... I've never seen any Daz render that looked even remotely like a realistic casual snapshot but plenty that look like staged studio photos.
Realism is a style of art as valid as any other. - Collages - Abstract(s) etc...
The goal is either emotive or a historical documentation. Feel this or this happened, here's the moment...
Somewhere the lines blur and you have the mix.
The problem lies in any medium where you become aware of the process. The strings behind the scenes.
Stage being different than film.
A written story where you become aware of tropes and cliches and writer short cuts. Expositions where the characters speak in un-natural ways or behave in illogical means that advances the plot/story.
Many of the points of illustration- generated from imagination - is to bring things that do not, cannot or not without great difficulty -- exist into a medium we can absorb.
If this was real what would it look like?
Since it's a picture, I'm taking in, it should probably look like a picture.
Since that's the goal of so many artists -- well....
Also, we don't count the hundreds of times everyday we are fooled.
And I'd say "fooling me" is probably the bar. Not Photo-real, but fooling the eye of the average person.
And the bar changes on how well we understand technology.
The first "photoshopped" picture (before photoshop existed) fooled many people who saw it because they didn't know THAT technology existed so the only choice for their brain was that it was real.
The first "fake News" story was believed for the same reason.
You know Peter Cushing passed, you also know what CGI is.Suppose you didn't know who that was? Would it be SOOO obvious? Maybe not, He looked even worse at the end of Return of the Sith, where was tiny and far away on screen. Rogue One is a marked improvement.
-------------------------
I agree with DeWater Rat --- the eyes usually. And it still seems to be focus. like somehow the eyes are never looking at anything or the right thing. Can'tr explain it better, but usually that.
The winners do the sunglasses thing or fix it in post.
Member jeffam112368_9a28fbd572 is giving it a good go:
http://www.daz3d.com/gallery/#users/133158/
Facial expressions. Yes, real people's mouths do widen and appear to ride up closer to their noses when they smile. Yes, the space between real eyebrows crinkle when real people frown. But more than a little bit of these expressions look really odd on Daz figures. I personally spend a lot of time tweaking mouth width and height when my characters smile, and I try to use brow inner/outer height to convey displeasure rather than cranking up the anger/disgust to high levels. Even so, I don't feel like I'm all that successful.
Yes, the Genesis 3 faces are not near plastic enough...
I agree that facial expressions are a giveaway, as are poses, which often have a sort of 'stiff' look to them, as if people are holding the poses rather than relaxing into them. The 'chest out and stomach sucked in' thing most of the females seem incapable of abandoning doesn't help with that one.
Also, fingernails, which are always clean and neatly trimmed, even on the medieval peasants who have been spending their days working in the fields or scrubbing floors. I've never seen a DS character with a tattoo that looked convincing either. They always seem to me as if they'd been drawn on and inked in with a biro.
I'm quite happy with posing. Maybe that's because I only use bought poses as a starting point - I generally spend quite a long time getting them just how I want them. For some reason, I can't use the posing tools and always use the parameter sliders - I don't know whether that makes it easier or harder, though.
I agree about smiles though. The mouth seems to contort unnaturally.
This is one of the closest to photo real I have seen that involves a human figure:
http://jghbxybr.deviantart.com/art/impatience-updated-650442091
Warning: This image has full nudity so you need an adult verified DA account to view.
Regardless of the adult nature of the image, the lighting, the woman's pose and skin, and especially the straw she is laying on are very well done, and at first I was convinced it was a photo. However if you look hard at the face, and a few other aspects you can see it is a render. The image was done in DS using IRay, so no high end application was used.
Hair.
That render is so good. I wish I could get light and shadow like that without having noise in the shadows. Also, the geograft and prop hair look very realistic. The straw must have been a headache. And I wish I knew whether the artist morphed the clothing manually or whether you can buy those.
If you check this product: http://www.daz3d.com/everyday-for-genesis-3-female-s there is a tshirt posed in a very similar way, although if the Russian artist used that for this image, I do not know. He would have used a different texture if he did.
Thanks for pointing to that - wishlisted! I think I must have missed it because I usually skip over Warrior Queen stuff and AS tends to be into that.
Honestly, true, complete photorealism is beyond the reach of even the best computers, and keep in mind that even the perfect "photorealism" of a high resolution photgraph of a real object isn't completely accurate to what our eyes see. Most of us see in stereo and our brain is constantly merging two images together, biasing back and forth as needed, so the odds of the interoccular and lens bias of any 3D photo system perfectly matching your own is much less likely than your chances of being struck by lightning. In the end, EVERY imaging system is simply a compromise between the limits of the storage medium and the pickiness of the viewers.
With that in mind, consider that ANYTHING digital is fundamentally incapable of perfectly reproducing reality, because reality is analog and composed of an infinitely varying assortment of slopes, grades, particles, influences and degrees of illumination. Just as CDs and DVDs don't reproduce sound perfectly accurately, but come within a compromise range that the majority of people find acceptable (with mp3s being a noticable step below that that a lot of folks have issues with), and as a 4Ka digital screening of a movie passes muster for most eyes as long as you accept that it's a flattened reproduction of a 3D world, CG modeling is an assembly of compromises, but with an even larger set of variables to deal with. And the more familiar the subject matter becomes, the more likely our eyes are to find things that don't feel right to our brains. Right now CG is at the point where furry animals and scaly reptiles can be generated that will be deemed acceptable by most, but when we get to humans, we're still walking with one foot in the uncanny valley.
In the end,I think, the best we can do it is fake it and hope that other people find the end result appealing, just as at one time the limit of the technology was an artist weilding a brush of oil bound pigments, and before that pigments blown with a straw on animal fat on the wall of a cave. What looks acceptable to us now will be the things we groan about in ten years and that our grandkids laugh hysterically at in 50.
And in the meantime, if an image causes people to react in the way that we want them to, that's a far greater measure of whether the image is sucessful than whether it look like a flat photo of the same item.
I didn't really mean to spark a debate about the merits of photrealism. I just wanted to discuss what glaring simulations of real world objects spoil the renders for people. For me, it happens to be cloth draping, razor-sharp edges and hair. I can excuse eyes, smiles, far-from-real texures, etc., because I'm not trying to make a photograph. I'm trying to suggest reality, not reproduce it. Cloth that defies gravity suggests tin, not cloth. Hair that looks composed of strips of painted plastic misses the mark of suggesting whipsy strands of hair. Buildings with walls that could slice bacon does not suggest brick or plaster.
That's all I'm saying. There's nothing photoreal about a Renoir or a Miyazaki animation but you look and are happy to believe. You don't look and think, No! That's just wrong.
Personally speaking, I always notice the pose/posture/expressions first. I have seen some very flawless, very well done images BUT, the pose just gives it away.
I suddenly thought of this as I thought about what gives away a digital scene: https://vimeo.com/202516691
Excellent :)
Clothes, Hair, Eyes. Usually in that order.
Thanks for that: really impressive.
Going back to Marble's original comment and the example of video that I liked to, the problem is that we have a great ability of hyper-vigilence to what constitutes the "uncanny valley". When we look at a digital image, we almost always see it in a digital image context so our hyper-vigilence kicks in and we instantly start noticing what is wrong. On the other hand, we watch a movie like Avengers: Civil War and see a scene that is shot entirely on a green screen set that looks like an airport tarmac where we have already picked the obvious cgi of Iron-man or Ant-man but ignore that they weren't even filmed outside. One assumption that I make is in car commercials is that it is CGI unless you can see someone open the car door and step out. Likewise, I never assume that two people in an add were there at the same time.