[mverta]

This is a reference shot... slightly wider lens, but not much. And most importantly, it doesn't look huge, either. Looks like a model.



This is just one of those times where an old adage proves true:

On film, with a physical model, you can get 80% of the way there and the eye will buy it. With a digital model, you can be 95% of the way there and it will still look fake.

This thing looks huge on the screen, but...


_Mike

[dutch_designer]

Instead of birds you could add some tie-fighters or some other smaller ships.
Unfortunately there's no haze to emphasize depth in space.. I agree with you though that a lack of geometry detail you're going to end up with an approximation, but stuff like size should be convincing even on small images where such detail wouldn't show up. Freefloating stuff like this without any clear real-world references and athmospherics is perhaps even impossible to convince properly without using some tricks like you mentioned before.

[DELETED]

DELETED

[JDHill]

It would be nice, to use "sun" without "sky" for that kind of renders...

...not meaning to hijack Mike's thread (so I'll try to keep with the starfield theme )...but the Sun can still shine when using MXI/HDRI background:



...and if Intensity = 0:





~JD

[iandavis]

Mike,

have you thought of a flock of flying stormtroopers?

in the movie what sells these ships as big, is that they are rarely seen without other elements which scale is known... like chasing the falcon... or with tie fighters swarming around. Without these, it does look like it's at the largest... the size of 1701... which is a little smaller then it's supposed to be I'm thinking.

[jdp]

very cool mike!

some thoughts about dimensions:

I also think that a reference is needed, because the situation it is too much uncommon.
perception is always relative. you don't really know how you can judge those things without experiencing the actual thing. also the distance from the object is important.

dimension is relative to space and time and it depends on the experience of it. to experience the space you need movement and/or references. here we don't have nothing to compare the object in space (references) nor in time (movement).

you are probably right about the details but in this case the problem is that details are maybe too far for being noticeable, so nothing would really change. Maybe a lot more tiny lights would help, but I am not sure since the surface is too white to notice them....

to summarize: you have an homogeneous background, you don't have reference objects, you are moving in a pure 3dimensional ambient without ground references, you don't know the distance from the object becasue you don't have visual distorsion (which is correct, but it let you only know that you are really far away), you have a static image, you are using only the eyes for judging the space.... you can't even tell how big you are!

man it would be though in these conditions...

looking at this skylab photos, I say that I have the same feeling as in your render: how big this thing is?





maybe images are, in the end, always "fake"....

btw, I wonder how much energy this monster needs to operate!

[michaelplogue]

Very cool Mike!

Maybe you could lower the f-stop a bit to give it a bit more DOF. I realize that the camera must be miles away, and you'd probably not get much of a DOF effect in real-life. However, It may give a bit more of an 'imensity' effect...

[Voidmonster]

I suspect that one of the problems you'll run into with this is simply that many of the cues we use to determine scale when looking at distant objects just don't apply in space. All the niceties of air are very important to our perceptions of scale. Check out this photo gallery of Antartica: http://www.luminous-landscape.com/locat ... tica.shtml Specifically, the bit about light.

[GM5]

Since you really wouldn't have DOF and there is no atmospheric haze... it calls for adding some TIE fighters - both standard and one Advanced x1 Mike Verta version

[thxraph]

solid mike, solid

i agree with Greg Montague for the tie fighters

thxraph

[Richard]

Nice work mike!

For me what is losing your scale is the proximity of the light source. If it were as you acknowledge further away I feel the fall off would not be so dramatic along the long lenght of the ship.

From your overhead shots I imagined the texture would add scale effectively in the original angle but not, bummer! I'd suggest a detailed texture but no bump may be a better solution!

I guess also it's hard as we haven't seen anything this big photographed in space and honestly when I've seen photo's of our own surrounding planets taken by probes and stuff the planets and moons always look surprisingly small.

How small does our own moon look!

[jfrancis]

Part of the problem is the behavior of little local lights on the dark side.

Look how evenly lit the open bay is under the ship. I wonder wIhat kind of lights would light that so evenly? A really big light source in comparison to the bay.

You might go for more small local spills and more darkness between. It's expensive to light things that well, even in Star Wars universe. (I'm talking about areas the sun doesn't reach)

http://www.falconcrest.com/images/gpaulni.jpg

http://web.syr.edu/~smjohn01/Dome_with_fireworks.jpg

http://www.chinfo.navy.mil/navpalib/shi ... 3night.jpg

---

Also, you could try darkening the back end to suggest distance to see if that helps. You say there's no dark haze in space? There's no sound, either, but that never stopped anyone from adding it.

[jfrancis]

On film, with a physical model, you can get 80% of the way there and the eye will buy it. With a digital model, you can be 95% of the way there and it will still look fake.

All other things being equal, I think audiences will buy a bad helicopter before they will buy a bad spaceship, because they know the space ship is fake.

[mverta]

Turned off all the lights to see how much total impact they're having on the problems. Also improved the keylight.



On one hand, I think improvment in the keylight is helping a touch, but I think the absence of lights is actually taking two steps back. It's possible that really good surface lights could make a difference, but I don't think they're going to take this image to the level of believable scale I'm after.

It's true that this image is worst-case scenario; that to me makes it a perfect candidate to learn stuff from. If I can manage to make this not suck at some point, then I'll really have something...

_Mike

[LarsSon]

Looking better - somehow. But the lack of texturing and reference objects
is making it not real for me. Even stars could do some trick.

Because the ship is in total emptyness it's really hard to measure the size.

If you don't mind, one example what reference gives to image.
I know that you know all the things about compositing, anyway.
This was just one night play month ago.




-LarsSon