skayen

Hello, since my last post, I came to the conclusion that Aeroeza and Mortesv were correct that the tail of my SSD was too long, so I bit the bullet and I set about correcting it. The overall length has been reduced by 7 studs, and the rear engine moved forward by two studs. I took some quick and dirty side shots in LDD to compare against the draft model and it was quite a bit better (or "closer", I think:-) to the draft from Aeroeza. The result, replete with weird colours and sans top-plating, can be seen next to the Korbanth model here: again, I've rescaled the image so the distance from the tip to the widest part of the wings is the same. Unfortunately, I don't like the new model. To my mind, it looks ~ok~ in the top image there, but from other views I think the new tail looks too unwieldy---even inelegant. I think part of the problem is that the angle from the wings to the tip of the tail is too steep on my model, and it's really a problem with the /aspect ratio/ of the tail than the overall length. At least, that's part of the reason, I think. The tail on Pellaeon's outstanding model is closer to what I think it should be. The leading outer edges of the tail are almost parallel to the ship's spine, giving the tail a long, sleek (and menacing!) appearance. I've used the 3x13 wedge plates along the outer edge, but the gradient is too steep and, accordingly, the tail is quite wide by the time it's reached the wings of the ship. Clipping the tail means the aspect ratio (the ratio of the width of the tail to its length) isn't as elongated, which I think breaks the overall appearance. I considered three techniques to fix this. The first was to change the angle from the wing tip to the tail to make it more orthogonal to the ship's spine, which would lengthen the tail slightly without changing the overall length of the ship---but I didn't like the look of this, and, I think the more swept angle in the current version is more accurate. (You can see the effect of a steeper wing-to-tail angle on Aryo Gono's MOC here). I also considered making the plating along the tail thinner by shifting the outer wedges towards the spine by a couple of studs. Unfortunately, this gave the tail a somewhat anemic look because I couldn't change the interior edge that connects the hull to the city without resorting to series of stepped flat plates. I also briefly considered using hinges to swing in the tail at whatever angle I thouht was appropriate, but that would either have obvious gaps in the top plating (like some SSD MOCs) which I could tile over, but not completely (because the joining studs would be at an angle, too...) Anio also separated the plates between the tail and his body. (later: actually, I'm reconsidering moving the outer tail edge towards the spine. I'll come up wiht a version that this the tail by one/two studs and shortens the length by a little and see what happens.) I think, ultimately, it is a battle between many different aspects of building with lego: while some /lengths/ may be accurate, the /angles/ they form might not, and the effect may well appear worse than a balance between errors in both angles and length. In this case, it is many subtle problems with my implementation: the ship is possibly a little too wide, the tail a little too thick, and so on, that I think that, on balance, the tail needs to be longer. But, you're right, Aeroza: there's a lot of doubt, and confusion when I compare many many different instances of SSD resin models like the Korbanth, above, and others---all of which have a clear view of the entire length of the model and appear to have longer tails. In the end, I decided that I preferred the overall look of the longer tail, but I figured it was worthwhile experimenting with the look of the shorter tail just in case it was all kinds of awesome ;-) I've tried to pull better 3D information out of the photographs of the real model. The irony is that I actually work in computer vision (essentially, quicker and dirtier photogrammetry!) and extract 3D from images, but with a lot more constraints about focal length and right angles and so on ;) To this end, I've tried triangulating feature points on the photographs to get a point cloud, but automated methods don't work on this image sequence. I ~could~ resort to manual tracking, and did try that for two views of the top hull, but it drove me to tears---the guy who took those photographs wasn't all that concerned with keeping corresponding points in view across different images ;-D You mentioned somehting about a render from ILM on the StarWars BluRay, and I can see that Mortesv has posted a picture. Can you tell me more about this sequence? Is it, by any chance, an animation of the SSD moving about---perhaps on a turn-table or something? Because if it is, then perhaps you could try ripping the video of that sequence and using a camera tracker to generate a 3D point cloud. I have access to Boujou, too (and others...), but you could also try PhotoSynth and PhotoTourism/Bundler and PMVS to make 3D models from nicely conditioned sequences... Unfortunately, I don't have the BluRay because I am sufficiently annoyed by the persistent meddling with the original movies =/ Also, thanks for your kind gesture to donate bricks, hah :-) I do plan on building the model when I /finally/ finish working on it (and making 3D instructions (*)), and plan on hitting up Bricklink for the pieces. I already own 10221 which I plan to cannabalise to feed my MOC, so I shouldn't need too many more pieces, I hope ;D The only problem is getting lots of desperate shops to economically ship to Australia. (*) which leads me to another question which I'll need to post sometime shortly. Namely: WTH do you do that, given that the automated instructions out of LDD leave a lot to be desired... cheers John

Hi I didn't see this post before, but although late to the party, please let me also extend my congratulations on a magnificent effort! Like everyone else, it seems, I think the lighting really puts your model in a class of its own--apart from the fact it is impressively large and detailed. I would have no idea how to begin work on a 2 year project like that, and I can only imagine the amount of redesigning you'd have had to undergone to see the project to the end. nice work! cheers John

Hi! hah, that is gold ;-) The only thing that could make it better would be if Darth was conducting with his lightsaber :-) cheers John

Hello! wow, your MOC is fantastic. It has really captured the key features of the SSD but in a convenient size :-D I note that you've angled the bottom hull (and the front engines) and have a nice, svelte engine area thing like the "real" model. You've also managed to straighten up the tail, too---some implementations (e.g. mine;-) I think taper the tail too much. I am very impressed. Congratulations =) cheers John

Hello, your MOC is fantastic. Nice work! cheers John

Hi, I came across your thread when linked from a discussion of my own Lego SSD :-) I see that you've also correctly angled the forward engines along the hull. One of the clues that TLG 10221's lower hull is flat is that the engines are parallel to all the other engines (and many other people's MOCs also keep the front engines parallel...!). I also like what you've done with the rear engines, too. One of the problems with my model is that the slope of the bracket from the top of the ship to the engines is too shallow on mine, but you've really nailed yours---and it lines up nicely with the edge of the hull plate. Your greebling is also very nice, particularly the detail invoked by the set of spanners holding the bar along the side. However, I think the last picture in the link illustrating the profile of your model really showcases how much it's improved over the TLG's version: your version has lost the hunch-back look of TLG version, replacing it with the arrow-head-esque profile. I like it ;-) The only problem with working without instructions is that now you can't disassemble and reassemble your model =/ Nice work. Much kudos and respect :-) cheers John

Hi, ok, thankyou for your positive feedback :-D I did actually see Aeroeza's and Mortesv's link some time ago when I was looking around for thoughts/feelings on the shape of the SDD after I assembled my version and thought it was all broken. My original intention was to work out a permutation of the pieces in 10221 to make it closer to what I thought it should be like, and to that end I got hold of the LDD model posted on these forums and started trying to deconstruct it while dragging the pieces aside so I can then be sure that I'm only using the pieces in the original set. But then the angle at the nose bugged me too much so I deleted all that work and started from scratch with the 2x4 wedge pieces and my own reference to images I could find on the 'net. Thanks for pointout out your link again--I've checked it out and now that I've struggled with some concepts (or, at least, identified some features I wanted to preserve in my model), I can see how you've approached them--but I'll write a comment on your thread shortly. BTW, I've added the draft to my profile comparison picture; I hope you don't mind. I am well pleased that the angles of my hull are a perfect match, actually. I guessed the angle because I didn't have a profile picture handy, and I figured 8degrees looked about right, and it seems we both agree ;-) I'd be the first to agree that there are many aspects of my model that I /know/ are not correct, so I definitely agree that my tail is longer than it should be. My tail is three 3x13 wedge plates, and I knew at the time that I really wanted 2.5 plates long. While I could address some of that discrepancy by having two plates side by side and using the bottom layer offset by half a length and coating the top with a different angle, then I'd end up with the obvious layered effect you can see on my model just behind the widest point---ie. you can see that the top layer is noticably shorter than the bottom layer. In the end, I actually liked having a long tail and kept it as 3 elements long. However, building the top plates was actually the first thing I did, and I think I'll revisit that decision in light of the more complete model. However, I am not sure how much I believe the 'scale model reference' illustration as canon. Maybe it was drafted from the model images (?) but if you compare the section after the last engines does extend a surprising amount. To be fair, that image has been taken with a wide-angle lens, and the perspective will definitely distort length ratios, particularly for the end of the tail because it's near to the camera. Consider also this crude image analysis using the length ratios of the last engine structure with respect to the rest of the tail: what this shows is the line along the structure that has been divided into segments of the same length, taken from the 2D image length of the last engine of the model: ie each mark denotes the /same/ 2D euclidean line segment which is taken from the corresponding segment over the engine. Unfortunately, this image is somewhat deceptive, because it doesn't take perspective into consideration: because the tail is slightly closer to the viewer, to maintain the same length in 3D space the equivalent length in 2D should progressively increase slight. I'd happily accept that you'd need to take this image with a grain of salt, but it /does/ suggest that perhaps the real model's tail is slightly longer than the draft image. Anyway, if we pretend that the perspective isn't an issue, then according to the simple image analysis, the length behind the last engine is ~3.75 lengths of the last engine. (The pink vector you see hanging off the tail is the same downwards vector borrowed from the window, which also has a clearly marked "down"). According to the illustration, the length is little more than 3 engines long. Can perspective account for the 'missing' ~25% discrepancy? ... possibly. There's a lot of guess work going in there, because I'm measuring things on different planes---the outer edge of the jet engine is closer to the camera than the mid-point to which I'm comparing it too. But my feeling is that the studio model really is a bit longer than what the draft seems to suggest. I've tried to find better photographs of the studio model which is side-on, but couldn't. The best model I could find is the Korbanth model in the original post, and its tail is arguably quite a bit longer than the draft, too... Regarding the width of the model: I agree that I think the 2x4 wedges make the model a tiny bit too wide. I was initially going to use the 13x3 wedge pieces and hope that by flatting the model, the angle subtended at the nose would also be correct (because obviously you can fiddle with the top-down angle by tweaking the slant angle: trading errors in height for a given wedge piece for errors in the top profile). Some other problems with my model: - the angle of the tail is too steep - the hanger bay was stylised on purpose to be wrong, mainly because no one can see it and it now reminds me of a shark ;) (The angles near the mid section are ~45 degrees, and they should be more like 80)--also, it extends to far to the front and should finish at a right angle, not taper to a point. Again, this was accidentally on purpose because I like the style ;-) - the last engine group is too close together and consequently the angle from the engine group to the ceiling is too shallow. - and probably others But, anyway, I think we can all agree that the TLG's model is too short or too skinny or too anything else you'd like, and while we can happily quibble about just how long the tail really should be, or how wide, I think it ultimately comes down to personal preference :-) I may revisit my tail, but to be honest i quite like the length of it (for various reasons;-), and really---thats ultimately what it comes down to, I think :-) I shall make some more thoughts on your link shortly, but let me say now that I greatly respect your interpretation, too. Thanks again for your feedback! cheers John

Hello, I like the idea. Some thoughts for adding new pieces/scenes include - adding corners so you can make intersections (for the scene where Threepio and R2D2 dash across the corridor admist gunfire) - escape pod exit bay template - possibly some room so Leia can record a holographic message, if it's not same as the escape pod bay You could think about extending the concept to the Death Star, which I'd argue uses very similar themes and include - prison corridors - that room where Han and Chewbacca fight the officers ("is everything allright?" "yes, yes, we're all fine. How are you?") - cells - tonnes of others, including caverns for swinging across, corridors for the fight between Vadar and Obiwan, &c. cheers John

Hello, I've added some more comparison pictures (including to an image from ESB) and close-up of the underside and city to the original post. cheers John

Hi, I think your idea is ingenious and well modelled. Nice work; I like it :-) I realise that this is probably impossible, but it occurred to me that it'd look cool if the clock hands were emanating from the laser cannon, as though the Death Star was firing / scattering golden beams of death ;-) cheers, John

Hi, I, too, think your diorama is outstanding. I like the way you've faded the larger trees in the forest into a clearing for the bunker, including having layers of smaller trees and green tiles: it makes your work look really striking. I don't want to undermine your work---because it really is extremely impressive---but in the spirit of thoughtful comments as 'food for thought' rather than just stunned awe, I was wondering if you've thought about / tried rounding out the tree trunks? Although I appreciate that the different tile heights adds to the bark texture, the trunks still look very square. Perhaps you could try adding another layer of tiles in the middle of each side and with careful (/randomised;-) use of the 2/3 slopes on the edges you might be able to give a more rounded look to your trunks while maintaing the bark-look? cheers, John

Hello, thanks for your kind appreciation :-) Although the exterior is pretty much finished (except that I haven't added any detail to the side trenches yet and I am still considering changes to the city), but it is currently ~3700 pieces and is (virtually) 1.44 metres long. For reference, the length from the nose to the widest point of my model is almost the same as the legnth from the nose to the widest point on 10221---but my version is quite a bit longer. Thanks for your comments regarding the greebling on the city. I wasn't a big fan of the "random pieces" that some versions use. While there is a lot of detail on the model, I think the city displays more structure which I wanted to bring out. Having said that, though, my version does have some element of excessive detail on the outer blocks, but I am musing about scaling that back, too. I'll post more pictures of the underside and the city soon; I just haven't had time to do it, yet. I still need to figure out how to turn the model into a working set of instructions, given that some of the internal structure isn't aligned correctly because LDD rounds rotations to the nearest integer and some of the hinges have fractional rotation =/ I hope to be able to build it at some stage, sure :-) Two people have suggested that it is a little bit too wide. Although it probably is a bit wider than ideal, I think it's actually extremely close. I found an image of a resin model by Scale Solutions which has a pretty decent top-down-view, which I've illustrated here for comparison: (removed: see first post) The top and side images have been scaled so the baseline I mentioned above is the same, but the overlay in the bottom right corner has been scaled so the /length/ is the same. This comparison is not /quite/ accurate, because although I can generate perfectly orthogonal synthetic images, the photograph is at a slight angle which can distort the width somewhat---either by making it larger by tilting to the back of the ship or narrow by tilting to the narrow sides. However, I maintain that it's reasonably straight-on for comparison, and that my model actually seems to be very close to its width (and that the angle formed at the nose is very close). Not exact, sure: but orders of magnitude better than the official lego model. The width of the 'wings' with respect to a Korbanth model is illustrated here, where I've tried to mimmick the camera angle taken by Matt Craig: (removed: see first post) I plan to make a few more comparison pictures with respect to reference photographs of the movie model from modelermagic and the movies themselves, but haven't got around to it yet. Finally, here's a comparison picture with the official model, which I think illustrates that my version is significantly flatter than the amusingly hunchback Lego version. (Also, the top angle is mirrored underneath.) (removed: see first post) thanks again for your feedback. cheers, John