That's just unreal, man. The little details are what make your stuff stand a level above the rest, and it makes me want to try that much harder to improve the design I've done. Wish I had another 7 or 8 hours in the day so I could improve!
Loving the ship, and the texturing, there is something about that image though that makes it look like a model rather than a full sized ship, don't know if its the blurred background, or something else.
Sorry I am being really vague, it looks real, but like a real model photographed on someone's table outside with the background out of focus because you are using the macro lens.
Loving the ship, and the texturing, there is something about that image though that makes it look like a model rather than a full sized ship, don't know if its the blurred background, or something else.
It might be that the "ground" and "sky" are too "perfect" or artificial. It certainly isn't the model. Coolhand. you're in a league all your own! Love it.
You're probably right about the backdrop ruining the illusion a little, though for me the orange lights on the landing gear are throwing me off a bit too, though it has already been mentioned that this area needs further work. Other than that this looks absolutely gorgeous, loving the lighting and texture work. Truly excellent work.
While I wouldn't spend much time to nit-pic the wip-pic, for reference theres at least 3 effects that might make it look a bit 'macro':
1, Theres a little fuzzyness, like a lens effect bleeding over the model from the background just to soften it but its probably too powerful.
2, some areas are too soft & some areas have too little modelled detail, this is what i'm addressing right now as best i can... Also i'm not rendering at a particularly high subdivision level (or sample level) in the wip-pics adding another iteration really does wonders for making the panel lines appear tighter.
c, the final reason is the most interesting and applies to everyone, so listen up;) its to do with texture sampling, basically if you put a texture on something, and render it at 0 angle of incidence then the texture will appear crisp and clean, as you rotate the surface away and increase the angle of incidence to the camera the texture will begin to blur as texture - sampling tries to decide which bit of the texture to display - because you're taking perhaps hundreds of pixels of data and squeezing it into one pixel in the image, what value do you make that pixel? You work that out by sampling, and by increasing sample rates and relying less on textures and more on geometry you can overcome this to an extent.
So for example i'm converting all the hex patterns into actual modelled hex patterns, thousands of little hexagons.. so you can see the sharp modelled hexagons on the bow sides in the last pic, but further back they're more textured than anything so they blur, because the sides are a large angle of incidence. This is why you should always model for example, panel lines rather than using a bump or normal map, because while it will be very convincing flat onto the camera, at any kinda grazing angle you'll just have a blurry mess.
so essentially if you model everything and have the sample rates up high enough you can more or less overcome that, it will always be with you to an extent though... but even now, if you choose the right angle and light it right, you can make a very convincing render with it... the one attached is a couple of years old and even made its way onto the concept ships blog so it must be ok;)
basically i know whats wrong with it (don't even start on the landing gear LOL), i just need lots of support:D But if you feel i've missed something feel free to speak up. Thanks for the comments everyone.
here's my list of fix-ups,
model all hex panels, redesign other areas of the TPS to fit that pattern better.
tighten up panel lines
do more with the burn/weathering effects
redo the materials (to an extent)
ditch the rivets from the panels (a, they're bump and blur out, b rivets? really.)
model the landing gear (its all tubes and placeholders and was never finished)
model the interior
add tons of greeble (in appropriate places like in the wheel wells and, perhaps glowing away behind smoked glass panels)
redo all the thrust housings, redo the 'gills' and tons of other stuff
model the interior again.
perhaps raise the front (by means of a longer front landing gear) to reduce the slope of the ramp by a couple of degrees more & reduce take-off run.
I think thats it, probably enough for me to be getting on with.;)
Ok well while I work on the front some more, another thing I really want to update are the windows. Never worked out a shape I liked for them so I'm totally open to suggestions. The best shape from a practical POV is perfectly circular, or perhaps slightly oval, but maybe that's a little dull, I can't orient them vertically either as although that would look nice, that's perhaps too much like star trek. when i redo them I'll pay more attention to the transitions, so you wouldn't simply have a hole cut out of a panel (ok there's a sort of inset) but you might have a layer of different material surrounding that (perhaps thermal protection) and various layers of glass and seal, probably to keep it relatively simple i might just have one thick piece of armoured glass.
and yes, there are lots of small issues with the mapping in this area when you get close (due to coordinates drifting under subdivision) or they way they sometimes run doesn't always imply a rational structure... most of these issues are on the right of the attached image. So that needs to be fixed, the airlock isn't big enough, the grill looks like crap and there's too much bump - i probably had it turned up for the far-away shots or to show up more under certain lighting conditions.
None of these things are an issue when further away but the entire build has been trying to see how close i can get to the hull and still have it 'hold up' so need further attention... fortunately other areas have a bit more polish and will need less tidying.
Rich, that's no problem, the main reason i'm coming back to this is to take some of its 'modellyness' away, replace fake geometry with real and so on.
The ports look fine to me though they should be re-positioned to the center of the hull panel. One is overlapping into the riveted edge of one panel and that wouldn't happen in RL.
Leo, that overlap is one of the incongruities I'm talking about, something I'd been intending to fix, problems in that image are too numerous to list;) As for the shape currently i feel its just a little unattractive - perhaps even nautical, like the flood holes Rich mentions... mostly there's just something about the curves i dont like, maybe looks a bit old fashioned... they aren't regular enough... like i want a straight line, a regular curve and another straight line, not a curve that varies like those.
I'd been considering something like the italicised trek style window like the shape that Biotech describes. roughly something like that guess, but if you imagine that tilted over, so its almost an almond shape. perhaps to better deal with any gas/plasma flows along the side the of the hull. I just wonder how that would fit with the panels overall. something more like the space shuttle cabin overhead windows might work better in that regard - making fuller use of the space between structural elements as described by the panel lines. I'd maybe lose the windows altogether, but they're an important scale cue, and indeed a source of those ever important little glowing lights.
Before I can tidy up these areas I've addressed a problem with the airlock, which was too small, this one could accommodate several people at once, there's another on the opposite side. The door retracts back and then up, sliding clear - I spent quite a while working out the correct arc for this - as the door is slightly curved and everything is off-axis, the door has to follow a very specific track up into the hull... There will be some extra gubbins to account for the retraction mechanism etc. The Airlock model is one i'd built previously a few years ago, no idea if i'd intended it for this model, i had to modify it to fit. need to redo all the thermal protection to suit the new form and so on. Also a very rough render of some of the internal elements, which are all in a very early state.
Thanks guys, here's another pic. Grafting in a new 4m section, i'd felt like moving the main gear back a tad, but that would have been more difficult, this would shift the centre of mass forwards, making it look more balanced on the main gear and this change makes it look faster, adds internal volume too. Still redetailing and reprofiling the nose too, with a 5 degree droop , you can see some missing panels on the side while i'm putting new panels. Also found a friend that was stashed away in the cargo hold - totally forgot about the hypersonic drones and launcher!
OK made some more test renders to see how its going so here's an update: Biggest change, new windows and interior spaces behind them, probably more like a recreation or mess area than cabins.
Each individual window is grouped with 3 others, these clusters are inset into the hull and a layer of the same armoured glass like substance that makes the windows fills this inset flush with the hull, the space between the inset windows and the outer layer would be open to space, so there's no weight of air pressure on these, other than the dynamic loads of atmospheric flight.
Looks great, however, if the green box is the size of a man, then the windows are too high for anyone to actually see anything out of, assuming the floors in those rooms are on the same level as the floor of the airlock.
Also the airlock doesn't allow for a lot of space around a human, especially if he was wearing a bulky spacesuit with a jet pack.
Tell you what Rich, I'll forgive you because i've not shown a layout of that area yet.;) For the airlock, basically its the same or bigger than similar sized chambers on Nostromo and Discovery, and those seemed to do OK with fully suited actors so I think we'll be fine.
Posts
Sorry I am being really vague, it looks real, but like a real model photographed on someone's table outside with the background out of focus because you are using the macro lens.
It might be that the "ground" and "sky" are too "perfect" or artificial. It certainly isn't the model. Coolhand. you're in a league all your own! Love it.
1, Theres a little fuzzyness, like a lens effect bleeding over the model from the background just to soften it but its probably too powerful.
2, some areas are too soft & some areas have too little modelled detail, this is what i'm addressing right now as best i can... Also i'm not rendering at a particularly high subdivision level (or sample level) in the wip-pics adding another iteration really does wonders for making the panel lines appear tighter.
c, the final reason is the most interesting and applies to everyone, so listen up;) its to do with texture sampling, basically if you put a texture on something, and render it at 0 angle of incidence then the texture will appear crisp and clean, as you rotate the surface away and increase the angle of incidence to the camera the texture will begin to blur as texture - sampling tries to decide which bit of the texture to display - because you're taking perhaps hundreds of pixels of data and squeezing it into one pixel in the image, what value do you make that pixel? You work that out by sampling, and by increasing sample rates and relying less on textures and more on geometry you can overcome this to an extent.
So for example i'm converting all the hex patterns into actual modelled hex patterns, thousands of little hexagons.. so you can see the sharp modelled hexagons on the bow sides in the last pic, but further back they're more textured than anything so they blur, because the sides are a large angle of incidence. This is why you should always model for example, panel lines rather than using a bump or normal map, because while it will be very convincing flat onto the camera, at any kinda grazing angle you'll just have a blurry mess.
so essentially if you model everything and have the sample rates up high enough you can more or less overcome that, it will always be with you to an extent though... but even now, if you choose the right angle and light it right, you can make a very convincing render with it... the one attached is a couple of years old and even made its way onto the concept ships blog so it must be ok;)
basically i know whats wrong with it (don't even start on the landing gear LOL), i just need lots of support:D But if you feel i've missed something feel free to speak up. Thanks for the comments everyone.
here's my list of fix-ups,
model all hex panels, redesign other areas of the TPS to fit that pattern better.
tighten up panel lines
do more with the burn/weathering effects
redo the materials (to an extent)
ditch the rivets from the panels (a, they're bump and blur out, b rivets? really.)
model the landing gear (its all tubes and placeholders and was never finished)
model the interior
add tons of greeble (in appropriate places like in the wheel wells and, perhaps glowing away behind smoked glass panels)
redo all the thrust housings, redo the 'gills' and tons of other stuff
model the interior again.
perhaps raise the front (by means of a longer front landing gear) to reduce the slope of the ramp by a couple of degrees more & reduce take-off run.
I think thats it, probably enough for me to be getting on with.;)
I love it!
Looking forward to the interiors, and the glowy things, love a ship with glowy things.
and yes, there are lots of small issues with the mapping in this area when you get close (due to coordinates drifting under subdivision) or they way they sometimes run doesn't always imply a rational structure... most of these issues are on the right of the attached image. So that needs to be fixed, the airlock isn't big enough, the grill looks like crap and there's too much bump - i probably had it turned up for the far-away shots or to show up more under certain lighting conditions.
None of these things are an issue when further away but the entire build has been trying to see how close i can get to the hull and still have it 'hold up' so need further attention... fortunately other areas have a bit more polish and will need less tidying.
Rich, that's no problem, the main reason i'm coming back to this is to take some of its 'modellyness' away, replace fake geometry with real and so on.
How about it you took a rectangle, then leaned the vertical edges, 45 degrees, then round off the corners?
I'd been considering something like the italicised trek style window like the shape that Biotech describes. roughly something like that guess, but if you imagine that tilted over, so its almost an almond shape. perhaps to better deal with any gas/plasma flows along the side the of the hull. I just wonder how that would fit with the panels overall. something more like the space shuttle cabin overhead windows might work better in that regard - making fuller use of the space between structural elements as described by the panel lines. I'd maybe lose the windows altogether, but they're an important scale cue, and indeed a source of those ever important little glowing lights.
Each individual window is grouped with 3 others, these clusters are inset into the hull and a layer of the same armoured glass like substance that makes the windows fills this inset flush with the hull, the space between the inset windows and the outer layer would be open to space, so there's no weight of air pressure on these, other than the dynamic loads of atmospheric flight.
Also the airlock doesn't allow for a lot of space around a human, especially if he was wearing a bulky spacesuit with a jet pack.
THanks!
I am loving those texuters.