Enterprise-refit 3D Model

Discussion in 'Fan Art' started by Lestatdelc, Jul 23, 2018.

  1. Lestatdelc

    Lestatdelc Lieutenant Junior Grade Red Shirt

    Joined:
    Jul 11, 2018
    Location:
    Portland, Oregon
    In doing some Googling around for some image searches for the Officer's Lounge fo the TMP era USS Enterprise-refit, I happened upon some images in Dan Goveir's great 1701-A 3D model thread. I have been working on a similar project as his for a while now. My project has evolved from, and based off of original production drawings, filming minature photo reference photos, etc. of the Enterprise.

    I actually own some of the set blueprints referenced early on in Dan's thread (such as the engineering set) and I have some blog posts about it on my website:

    [​IMG]

    I am currently doing a deep dive on the bridge and deck 2-3 superstructure. The "official" blueprints produced by David Kimble back in 1980, are close but not super accurate. As some minor differences vs. those drawings and what was actually built in the filming miniature by Magicam are not accounted for. Such as a ever so slightly thicker saucer which has bene confirmed by high-resolution photo analysis of the original decal set produce by Astra Image (RAA) for the filming miniature, and in conversations with Jim Dow (who was lead model-maker at Magicam on the Enterprise) and Richard Taylor (the art direct with RAA who was the lead designer & who was Andrew Probert's art director). Some of the discrepancies are due to how the miniature was fabricated, such as not actually accounting for the thickness of the vacuform plastic sheeting used over the original plaster spun buck. The buck being made to the original Richard Taylor (original bridge) drawings (two of which I own as well).

    I also have several of the original officer's lounge miniature set blueprints as well as original "new" bridge design sketches, all drawn by Leslie Ekker:

    [​IMG]

    [​IMG]

    Anyway, I did some more reworking to get a smoother and better aft 2-3 deck slope over the weekend to more tightly match the model, and specifically the aforementioned interior "set" miniature.

    Here's a video of a quick fly around of the new 2-3 components of the model. This is very preliminary baseline model. I will be using it to then work backwards from it, to build out the longeron, stringers, spars and ribs, for the underlying structure of a presumed 23rd century take on semi-monocoque construction of the Enterprise-refit.

     
    Galileo7, Pauln6, blssdwlf and 4 others like this.
  2. Lestatdelc

    Lestatdelc Lieutenant Junior Grade Red Shirt

    Joined:
    Jul 11, 2018
    Location:
    Portland, Oregon
    I had cross-posted the video above over at The Fleet Workshop website, one of the regulars there asked me to clarify when I mentioned the miniature, if I was talking about the lounge model, and not the full ship miniature, and if so, do the windows on the ship model match the detailing on the lounge model?

    That lead to a lengthy answer about the couple of months work to get to this point in the superstructure model which might be good to add here.

    The curvature baseline of the windows (top view) is a match visually to some of the reference of the exterior “ship” miniature, to the drawing of the interior “set” miniature. That curved window baseline is just a fraction below the slanted kick-board “wall” that defines the sunken seating area surround and the raised planter areas. That slanted baseboard goes up at a 60º angle to a short 90º vertical section to join with the “window” wall that slants up as a back wall and curves upward and becomes the ceiling.

    So those basically align and match, and I got those to match down to the millimeter (finally).

    However the large dashed outer curvature line, which is labeled as indicating the “outside seam line for ref. only. do not build” is not accurate to the exterior aft superstructure. The windows and entire lounge area is further aft in relation to the superstructure saucer seam line. That dashed reference seam line in the interior “set” miniature drawing is simply too far back from the windows to match the relative position in the exterior model. So that dashed lines curve is off as to shape/curve to the actual exterior seam.

    Since if was only meant as a conceptual reference point, and wasn’t part of the interior miniature that was built, I had to forgo trying to somehow make that reconcile (which it simply cannot as drawn). And let the exterior superstructure and saucer seam line land where it needed to be to match the distance up the slope for the baseline of the windows. That required taking into account the depth fo the window “alcove” that goes around the superstructure and is inset into it.

    For the height of the windows, I went with the measured height of the aft view drawing of the interior, which when projected back and “cut out” of the back sloping wall.

    You can see how it was built in this screenshot. The magenta colored rectangle matches the measured height (converted into the same scale) taken from the Ekker interior “set” aft elevation drawing, which is then projected onto the sloping curved wall:

    [​IMG]

    It pretty much aligns with the exterior miniature photo references and placement. And after considerable effort in getting the compound curves to allow that rectangle (projected back) to get it to match the interior wall seam (which you can see in the "bottom view"—upper left—in the split view screenshot).

    So that window baseline position and window height are what I ended up having to use as the cross-reference point between interior and exterior.

    Working back and forth to get the complex compound curves in all axis around the sloping back wall to match the miniature profiles, and top view when it meets the saucer section (which is itself a shallow compound curve was a bitch.

    Because that swept out shape has to interact with the saucer—itself as compound curve—in such a way that it match the top view exterior shape of the superstructure/saucer seam line, but also match profile in curvature, while sweeping around and transitioning in such a way so that window baseline matches the interior drawing line, and exterior reference photos.

    It has taken me literally a couple of months worth of trial and error modeling it in 3D to get it to jive with all the photo references. The variables and the challenge being working with the inset depth, hull thickness at the windows section to get it to all match (at the windows) in curvature to both interior and exterior.

    Here is a simplified step-by-step of the challenge.

    Starting out from production design drawings getting the known profiles of the fore, aft, and port/starboard vertical edges, and creating them (shown in magenta) is pretty straight-forward:

    [​IMG]

    Likewise the uppermost deflector grid ring:

    [​IMG]

    …the saucer profile:

    [​IMG]

    …and the seam outline (in a flat plan top view):

    [​IMG]

    Taking the saucer profile—revolving it for form the saucer surface itself—then taking the seam outline and projecting it down onto the saucer profile gets you the seam in three dimensional space (again fairly straight-forward):



    [​IMG]

    Now things get tricky.

    Determining the key transition vertices points between the aft vertical slope path to forward angled slope line (which is actually a 202.5º revolved shape for the front part of the superstructure)…

    [​IMG]

    …was the deceptively hard thing to work out.

    Because defining those transitional sloped vertices (the two highlighted in magenta below) not only need to keep the transition smooth between the start and end slopes (while matching what is seen in the photos of the exterior miniature)…

    [​IMG]

    …but critically slope inward and curve lateral in a way that from a surface…

    [​IMG]

    …that can then be inwardly offset by the amount of alcove inset and the window thickness…

    [​IMG]

    …in such a way that interior curvature of that intersection meets at the interior window baseline.

    All while doing so in a way that slopes upward to form the interior window plane properly to match the interior shape and dimensions of the lounge miniature which is again derived form projecting the window height rectangle (magenta) onto the interior sloped curved surface to form the window baseline and top edge (yellow shape)…

    [​IMG]

    This can then be used to project up onto a plane intersecting with the column angles as defined in the Ekker top-view drawing to form the raw window shapes outlines (yellow)…

    [​IMG]

    …which in turn are extruded to create shapes that intersect with the aft inner aft surface …

    [​IMG]

    …to be cut out of the surface…

    [​IMG]

    …which can then be “shelled” to form the thickness of the raw window framing, the corners of which are then filleted to round the corners (the two windows on the right with the filleted corners highlighted…

    [​IMG]

    And so on.

    It took over two months of tinkering to get it to match correctly in all those areas. It would be easy to simply ignore the interior wall and window baseline of the interior, and get something that looked pretty close to just the exterior. Many of the 3D models I have seen others make also don’t have the very subtle vertical bowing of the front and aft profiles, and simply make it a straight angle for them.

    Where things don’t match perfectly to the interior drawings are the the top view in the Ekker drawing. The windows are not quite a “long” from bottom to top (when viewed from above) as what is in the interior “set” miniature drawing.

    Here however the Ekker drawing is not consistent with itself in that the the aft elevation window height doesn’t reconcile with itself to the top view window heights, given the slope of the cross-section side elevation in the same set of three orthographic in the Ekker drawing.

    But the 3D model does match vertically in height to the interior aft elevation drawing, and match to the shape, size, and position to the exterior miniature reference images.

    So that’s what I went with since it was the best solve.

    I did however use the horizontal radius of the ellipsis in the top view drawing to determine the circular radius of the rounding (filleting) of the window corners, which with “cut through” the sloping aft wall do pretty much match the shape when viewed directly from above and from the back/front (i.e. exterior view/interior view looking out).

    I used the center points of the top view Ekker drawing (which have centerline dimensions outward) to define the angles of the columns and the convergent origin point. Which was crucial to determine the two key transition slope vertices for the outer shell in forming the overall proper outer superstructure base surface.

    At first blush you would think it would not be too hard to model it. But when you start getting down to the millimeter (in-universe) and try and get it to all to line up and make sense, while still matching the exterior miniature forms is hard enough. But to get it to reconcile with interior miniature, and do so in a way that is faithful to what can be gleaned for the production drawings (which have some internal inconsistencies in them) and from the photo analysis of the exterior was really complex.

    Anyway, enough of my rambling for now.
     
    Last edited: Aug 7, 2018
    DanGovier, David cgc and Spaceship Jo like this.
  3. Terranimperial

    Terranimperial Lieutenant Junior Grade Red Shirt

    Joined:
    Jul 8, 2018
    This is some good work! I wonder how many polygons just this section has.
     
  4. Lestatdelc

    Lestatdelc Lieutenant Junior Grade Red Shirt

    Joined:
    Jul 11, 2018
    Location:
    Portland, Oregon

    Thanks.

    I can give you a precise answer.

    Zero!

    :hugegrin:


    I am using Moment of Inspiration (MoI) 3D Modeling software, which is NURBS-based software and not polygon-based 3D modeling. This means it is all mathematical curves or shapes (tough it handles straight line segments as well of course). But having it not be polygon-based means it is wholly scale and resolution independent, with no loss of the mathematical purity of the vector-based Bézier curves, unlike the majority of other 3D software applications which usually use polygons because it is faster and "lighter". This is because in polygon-based modeling everything, even curves, spheres, etc. are made out of polygons instead of being pure mathematical shapes. With NURBS you can literally zoom in forever and there is no degradation of forms or shapes. Circles. eclipses, etc. are true curves, not approximations of curves.

    Not only that, but MoI allows me to still work natively in Adobe Illustrator (which I use professionally all day long) and allows me to ping-pong back and forth between those applications.

    The other aspect that’s great about MoI is that it handles exporting out a variety of file formats and 3D geometries well, and handles the tessellation (converting the pure geometry into faceted polygon approximations) fairly well. It can be exported to whatever level or scale is needed.

    Once it is in any of the other polygon-based “mesh” formats, I can then clean up and simplify any of the complex faceting that the conversion from mathematical curves into polygon geometries might introduce into the shape—if needed.

    So for example I can export the NURBS geometry I work up, to .obj files—which polygon mesh-based programs work in—and open into SketchUp (which it can also export into directly) for whatever I want or need, be it 3D printing, rendering or animation.
     
    Last edited: Aug 7, 2018
    B.J. and Sibyl like this.
  5. Sibyl

    Sibyl Caffeine Pill Popper Premium Member

    Joined:
    Apr 13, 2015
    Location:
    Sibyl's Scrambled Mind
    Moi certainly looks like something to consider. Not too expensive for a NURBS modeler, for sure. Rhino is more than twice that.

    I'd love to be able to model in NURBS (I currently use what is lovingly referred to as NURMS (Turbosmooth in 3ds Max, otherwise knows as Subdivision Surfaces). Unfortunately, NURBS in Max is--well--less than desirable to work with. There's a plugin that will convert Sub-D meshes to NURBS, but it's as much as Rhino.

    I do love both methods, but there are a lot of occasions that NURBS would be preferable to me, especially with 3d printing curved surfaces.
     
  6. Lestatdelc

    Lestatdelc Lieutenant Junior Grade Red Shirt

    Joined:
    Jul 11, 2018
    Location:
    Portland, Oregon
    The past few days I began working on the generic corridor sections beginning with the "curved" radial sections as produced in Lora Johnson's Mr. Scott's Guide to the Enterprise as well as the numerous photo references and blu-ray screencaps. I did this to begin to try and reconcile the deck levels in the primary hull, which obviously impact how the deck and window levels within the bridge superstructure are arranged.

    Needless to say, this opened up a can of worms and the discovery that almost all the internal blueprints out there made a huge erroneous assumption about the basic number of decks within saucer. Most fan produced internal blueprints usually have eleven decks in total within the primary hull. Basically echoing the overall assumptions made with the old Franz Joseph's TOS Enterprise Booklet of General Plans (the grandfather for Trek "blueprints") produced in 1975:

    [​IMG]

    Almost all the plans have the bridge as deck one, two decks within the superstructure, two decks between the superstructure and the two main decks that compose the main "disc" off the saucer, then four more decks as the saucer reduces down to the lower sensor dome area. Though some fan-produced plans have only three deck and not four in the lower portion below the main "disc".

    However going back over some Andrew Probert production sketches one in particular stood out. It was one where he was proposing a location for the rec. room deck within the saucer under the bridge superstructure—and the sketch was done before the redesign of the bridge superstructure and the lower sensor dome were redesigned when Richard Taylor and the folks at Abel & Associates were still responsible for the VFX designs—which shows that the assumptions most people have about the number of decks, at least in the refit-Enterprise of TMP is wrong.

    In that sketch is shows that there was there originally only one deck intended to occupy the superstructure below the bridge deck:

    [​IMG]

    This is also confirmed in the classic TMP cut-away poster drawn by David Kimble (who also inked the official "blueprints" which were produced in conjunction with the film's release. While many state that the poster is not "canon" (and indeed there are some minor errors in it) it does reflect the overall design thinking of the art department/design team at the time of the film's release:

    [​IMG]

    One of the minor errors in it are things like it being clear from the film itself that the corridor where Decker encounters Ilia as she makes her way to her quarters from the turbolift is deck five (as shown on screen) but is drawn as being deck four in the Kimble poster.

    [​IMG]

    That said, eliminating the need to try and force a second deck into the superstructure solves a myriad fit issues with deck spacing, turbolift shaft heights and clearances as well as window/portal arrangements in the saucer. Going with six decks in total (counting from bridge to bottom deck of the main disc before the narrowing taper) instead of seven allows an overall three-meter internal deck framing that conformably accommodates the corridor framing as built on the Paramount soundstage. The "straight" corridor sections of course eventually being repurposed for the TNG corridor sets.

    [​IMG]

    The main problem it solves in the upper superstructure is it nicely resolves the problem of the fact that on the exterior of the filming miniature the portals around the inset alcove are pretty much equidistant vertically within the superstructure height. Trying to cram two decks into that space not only makes turboshaft stations highly problematic, but no matter how you fudge it, you end up with portals that are ether blocked by the deck itself, to sitting at the floor or at the top of the ceiling (depending on which way you fudge it.

    This solves that problem entirely, and allows the bridge to sit lower in the superstructure and will help make the twin turboshafts work better since they no longer need to come up as high into the bridge level structure.
     
    Last edited: Aug 29, 2018
    Galileo7 and Tallguy like this.
  7. Lestatdelc

    Lestatdelc Lieutenant Junior Grade Red Shirt

    Joined:
    Jul 11, 2018
    Location:
    Portland, Oregon
    A video showing a little more work on the saucer edge to start working on the deck level framing at the saucer rim:

     
  8. Mytran

    Mytran Rear Admiral Rear Admiral

    Joined:
    Aug 5, 2009
    Location:
    North Wales
    Nice solution to those otherwise ridiculously high windows on the lower rim! :techman:
    A little precursor of the Ten-Forward lounge, perhaps?
     
  9. DanGovier

    DanGovier Lieutenant Commander Red Shirt

    Joined:
    Jul 31, 2011
    Location:
    Somerset, UK
    Having a raised area is a nice idea. I naturally tend to build everything on one level, and it's a bad habit because in real life things are rarely completely flat :P
    My solution to the windows is to have a lounge as tall as the Rec Deck, but with a mezzanine area for tables etc. The mezzanine doesn't need to conform to normal deck height, so it can be positioned for optimal window placement.

    It's great to see your build coming to life, I'm looking forward to seeing more :)