Page 1 of 12 12311 ... LastLast
Results 1 to 20 of 238

Thread: C-47

  1. #1
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963

    C-47

    Starting a model of a C-47.

    Took over 400 reference photos of the actual aircraft, slowly building it piece by piece. Should be quite accurate when it's done. I am looking for any modelling crits from people that are in the know. Eventually it will have D-Day stripes and the whole bit.

    I used agisoft photoscan to import the rough scan and all camera positions in blender. I then retopo the entire scan a panel at a time. Some areas have more detail than others. I use the camera positions and reference photos I took of the actual aircraft to fix errors and model details the scan missed. The aircraft is the C-47 out front of Evergreen aviation museum in mcminnvile oregon usa.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	flying with clouds.jpg 
Views:	102 
Size:	373.5 KB 
ID:	441222  

    Click image for larger version. 

Name:	Render53.jpg 
Views:	55 
Size:	148.3 KB 
ID:	441223  

    Click image for larger version. 

Name:	Render10 opengls with overlay.jpg 
Views:	129 
Size:	534.2 KB 
ID:	441224  

    Last edited by NRK; 15-Jun-16 at 21:39.
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



  2. #2
    Member
    Join Date
    Oct 2007
    Location
    Poznan, Poland
    Posts
    711
    Good start, these nacelles look quite good!

    However, it seems that you have troubles to obtain the perfect circular cross-section shape of the NACA cowlings.
    Am I right, guessing that you split these cowlings into separate panels? And they are smoothed by the subsurf modifier? I would suggest to form them as a single piece (at least: do not cut them along the circle radius). In this way it is easy to obtain a nearly perfect circular cross-section of a subsurfed surface. Then you will create convincing impression of these panels using textures (bump maps, reflectivity map, etc).

    It is still possible to cut such subsurf-smoothed surfaces along radius, but you have prepare appropriate layout of the mesh before. (To do it, you have to create the root cross-section first, containing a double vertex in the point of the split, and still having a perfect circular shape. Then you can extrude such an edge loop into the shape of NACA cowling. Finally, split this mesh along these double vertices - cutting subsured mesh along topologically "sharp" edges does not alter its shape). Well, these and other methods you can find in my book (its address is at my signature, see below).
    Last edited by Witold Jaworski; 17-Apr-14 at 03:54.



  3. #3
    show us some of the pics as you go along
    any specific model - so many DC3 have been made for military ?

    in latest built there is a new smooth function in N panel
    try it seems more powerful then the other one!

    happy bl
    Last edited by RickyBlender; 17-Apr-14 at 07:41.
    Nice 3D models = 25 % K.I.S.S + 25 % common sense + 25 % Inspiration + 25 % Knowledge
    Win 10 64 bits / Thanks and a Happy 2.79 with Cycles



  4. #4
    Member theoldghost's Avatar
    Join Date
    Jun 2008
    Location
    U.S.A. The Southeast section of Virginia
    Posts
    1,659
    What a historical plane. As a 1950s Airborne Soldier I did not jump the C-47. They had moved on to the C-119, C-123 and C-130. But, many of our older sergeants had jumped it in combat. And, of course the part it played in WWII is legendary. That Skytrain going over England headed for France must have been chilling for those looking up on D-Day.

    Please consider an animation when you're done.

    Click image for larger version. 

Name:	C-47-on-DDay.jpg 
Views:	423 
Size:	190.6 KB 
ID:	303296Name:  C-47.jpg
Views: 6195
Size:  35.9 KB



  5. #5
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963

    Thanks for the comments.....

    Thanks for the comments, I'll try to prove my accuracy with some side by sides soon. I actually don't have any subsurf applied at all as of yet. I have a workflow I've been refining for some time. I'll elaborate more as I progress.

    So I've been working on the Pratt & Whiney R-1830 Twin Wasp a little along with the variable pitch propellers. I'll post side by side with the reference photos when I'm a little further along.
    Click image for larger version. 

Name:	Render3.png 
Views:	250 
Size:	406.3 KB 
ID:	315252


    Note: I know the propellers are backward on one side, I'm just using the mirror modifier for now, at some point I'll turn it off and tweak the differences.
    Last edited by NRK; 15-Jun-14 at 11:08.
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



  6. #6
    are you modelling the whole engine ?
    lot's of work on these - I think 14 pistons

    happy cycles
    Last edited by RickyBlender; 20-Apr-14 at 08:02.
    Nice 3D models = 25 % K.I.S.S + 25 % common sense + 25 % Inspiration + 25 % Knowledge
    Win 10 64 bits / Thanks and a Happy 2.79 with Cycles



  7. #7
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963
    Originally Posted by Witold Jaworski View Post
    Good start, these nacelles look quite good!

    However, it seems that you have troubles to obtain the perfect circular cross-section shape of the NACA cowlings.
    Am I right, guessing that you split these cowlings into separate panels? And they are smoothed by the subsurf modifier? I would suggest to form them as a single piece (at least: do not cut them along the circle radius). In this way it is easy to obtain a nearly perfect circular cross-section of a subsurfed surface. Then you will create convincing impression of these panels using textures (bump maps, reflectivity map, etc).

    It is still possible to cut such subsurf-smoothed surfaces along radius, but you have prepare appropriate layout of the mesh before. (To do it, you have to create the root cross-section first, containing a double vertex in the point of the split, and still having a perfect circular shape. Then you can extrude such an edge loop into the shape of NACA cowling. Finally, split this mesh along these double vertices - cutting subsured mesh along topologically "sharp" edges does not alter its shape). Well, these and other methods you can find in my book (its address is at my signature, see below).
    Thank you so much for your work, my coworker bought your book and got a lot out of it. I'm experimenting with my own methods which works better in some ways and worse in others. I didn't realize the cowlings were perfectly round, I'll tweak them in thanks
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



  8. #8
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963
    Originally Posted by RickyBlender View Post
    are you modelling the whole engine ?
    lot's of work on these - I think 14 pistons

    happy cycles
    Yes 14 pistons, haven't decided how detailed I'm going to get yet, nice thing about it is do it once then x14
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



  9. #9
    well more like 2 X 7 = 14 I think
    I did a 9 pistons radial engine not easy or low poly to get nice 3D model !
    but nice challenge in any case
    did not take the time to make this big one
    i began doing some of the B-17 B-24 and I might do this engine



    happy cycles
    Nice 3D models = 25 % K.I.S.S + 25 % common sense + 25 % Inspiration + 25 % Knowledge
    Win 10 64 bits / Thanks and a Happy 2.79 with Cycles



  10. #10
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963

    3D Scan workflow.

    Using the pictures I took of the actual aircraft I converted them into a 3D model using Agisoft photoscan. I then know I have all of the proportions and scale correct. I also have closeup photos of every area of the plane. Here I'm modelling the underside/nose using the workflow

    Attachment 303980

    You can also see from the screen shot I have the position of all 400 cameras and their corresponding photos so I can render from the EXACT perspective of the photo and texture paint as well.
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	WIP 3D Scan with Model.png 
Views:	298 
Size:	1,000.9 KB 
ID:	316181  

    Last edited by NRK; 19-Jun-14 at 23:25.
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



  11. #11
    how come the scan is so noisy !

    did you happen to take a scan of the piston ?
    very difficult to get good drawings for these !




    thanks
    Nice 3D models = 25 % K.I.S.S + 25 % common sense + 25 % Inspiration + 25 % Knowledge
    Win 10 64 bits / Thanks and a Happy 2.79 with Cycles



  12. #12
    Member
    Join Date
    Oct 2007
    Location
    Poznan, Poland
    Posts
    711
    Originally Posted by NRK View Post
    Using the pictures I took of the actual aircraft I converted them into a 3D model using Agisoft photoscan. I then know I have all of the proportions and scale correct. I also have closeup photos of every area of the plane. Here I'm modelling the underside/nose using the workflow

    <picture>

    You can also see from the screen shot I have the position of all 400 cameras and their corresponding photos so I can render from the EXACT perspective of the photo and texture paint as well.
    A very interesting method! For sure it helps greatly to catch the exact proportions of the modeled object. (For example, in the classic methods it is difficult to determine the exact proportions of the landing gear to the fuselage or wing).
    On your picture it seems that this method produces relatively rough reference surface. Relying on this basic data, do not forget about the technological details of this airframe. In the times when it was designed engineers often shaped fuselage or nacelle cross-sections using combination of arcs. I won't be surprised, if the middle part of this fuselage and the tail has a symmetric cross-section, based on two ellipses.



  13. #13
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963
    Originally Posted by Witold Jaworski View Post
    A very interesting method! For sure it helps greatly to catch the exact proportions of the modeled object. (For example, in the classic methods it is difficult to determine the exact proportions of the landing gear to the fuselage or wing).
    On your picture it seems that this method produces relatively rough reference surface. Relying on this basic data, do not forget about the technological details of this airframe. In the times when it was designed engineers often shaped fuselage or nacelle cross-sections using combination of arcs. I won't be surprised, if the middle part of this fuselage and the tail has a symmetric cross-section, based on two ellipses.
    I'm glad you like the idea, I really respect your effort and attention to detail and your input is GREATLY appreciated

    Originally Posted by RickyBlender View Post
    how come the scan is so noisy !

    did you happen to take a scan of the piston ?
    very difficult to get good drawings for these !
    All 3D Scans are noisy, this is noisy because the computer is extrapolating the data from the photographs which are blurry etc, it's actually not as noisy as it looks, the areas of the scan where I had detailed photos are quite acurate. The scan gave me some details of the pistons but not much. I'll have to model the pistons the old fashioned way.... google images. The good news is the scans caught enough of the engine that I've got the scale and placement of the pistons, just not the details.
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



  14. #14
    I search web and was not able to find a lot of pics on the piston's shape!
    I don't have enough details to do a nice 3D model

    hope you have enough info to do it and will wait till you show me more details
    then I might be able to make my own model !

    happy cycles
    Nice 3D models = 25 % K.I.S.S + 25 % common sense + 25 % Inspiration + 25 % Knowledge
    Win 10 64 bits / Thanks and a Happy 2.79 with Cycles



  15. #15
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963

    Quick Update Render

    Lots of work to do... but moving along nicely. I have a question... how should I do rivets. I'm considering using hair with object and 'painting' them on. Another option is to literally paint them on with bump maps when I'm done. I could use dupliverts as they would be nice to instance... any thoughts?

    Attachment 305013
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	Render4.png 
Views:	123 
Size:	352.1 KB 
ID:	316179  

    Last edited by NRK; 19-Jun-14 at 23:24.
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



  16. #16
    Member
    Join Date
    Oct 2007
    Location
    Poznan, Poland
    Posts
    711
    Personally, I create rivets using bump maps, otherwise it would be too much work. (I have never seen finished model with modeled rivets, although I saw one or two unsuccessfull attempts to do such thing).
    I have impression that the cross section of this fuselage near the wing should be more regular. (Use curves of greater radius, and change it gradually). You have fixed the engine cowlings, but the shadow distribution on their surface reveals that they still do not form a perfect circle. Such problems with regular shapes happens, when you have too many vertices in the mesh (it is difficult to place all of them properly). It looks like a weary skin of a very old plane .



  17. #17
    are you using cycles ?
    try in latest build the new snooth function

    for riverts duplivert is good but need to do it for 3 axis !
    unless there is a trick here !

    this is good if your planning to do close shots
    otherwise just do a simple bump map !

    happy cl
    Nice 3D models = 25 % K.I.S.S + 25 % common sense + 25 % Inspiration + 25 % Knowledge
    Win 10 64 bits / Thanks and a Happy 2.79 with Cycles



  18. #18
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963

    Little bits of progress...

    Unfortunately I'm having trouble getting the right side of the aircraft to compute with my little computer (the file is too huge to add more chunks to the 3d scan model) Also the real problem is I don't have the center of the aircraft in real world space from the 3d scan and thereby my mirror isn't accurate. I spent a long time thinking about it and I can't seem to figure it out. I started writing a python script that would allow you to move the model and then go to edit mode and move the mesh back to it's original position to allow me to iterate the object center to the correct position/rotation. Anyway my python script wasn't working too well, so I got frustrated and roughed int he fuselage.

    As you can see I just roughed in the fuselage, lots of detail to add. Started the landing gear, kind of all over the place, I tend to model by brush strokes and then refine. I know the leading edge of the tail is messed up. I think I'm going to try painting on the rivets with hair and then baking them to bump if I get into trouble....

    Attachment 306951
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	Render5.jpg 
Views:	135 
Size:	136.4 KB 
ID:	316178  

    Last edited by NRK; 19-Jun-14 at 23:23.
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



  19. #19
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963

    Another quick test render.

    Threw in some more rough shapes... zeroing in on details...
    Click image for larger version. 

Name:	Render6.jpg 
Views:	126 
Size:	77.2 KB 
ID:	316177
    Last edited by NRK; 19-Jun-14 at 23:22.
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



  20. #20
    Member NRK's Avatar
    Join Date
    Aug 2009
    Location
    Washington, USA
    Posts
    963

    Little bits of progress...2

    Working on the landing gear... lots of detail in there, trying to figure out the mechanism.

    Attachment 309096
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	Render9.jpg 
Views:	123 
Size:	82.7 KB 
ID:	316176  

    Last edited by NRK; 19-Jun-14 at 23:21.
    We spend years in school to prepare for this life... how much time have we spent preparing for the next?

    My long term Film Project:Glasses the Movie



Page 1 of 12 12311 ... LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •