Thread: 24 mm lens with 0.64 speedboster , the math

Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 22
  1. #11  
    Member
    Join Date
    Aug 2015
    Location
    UK
    Posts
    80
    I guess adding a magnifier glass in the mix will up the level of complexity by a notch, will test on a native full-frame lens combo and see if I get a closer match when I don't use a speed booster. Thanks for the link will have a readthrough
    Reply With Quote  
     

  2. #12  
    Senior Member
    Join Date
    Oct 2016
    Posts
    384
    If you modeled a set from a scan, then try to add a camera afterwards it’s not ever going to line up with footage. Typically you take photos from the camera’s perspective, solve for world alignment with a perspective app like fspy, then start building the set on top of your reference image looking through the 3d camera. Or build off of a camera-solved 3d track. Either way you want to build a set looking through the 3d camera viewport with footage as a background image sequence for reference.
    Reply With Quote  
     

  3. #13  
    Member
    Join Date
    Aug 2015
    Location
    UK
    Posts
    80
    Building my set against the scan works fine, It's just when I place my 3d camera when using the speed booster my 3d camera is just slightly in front of my set camera when it matches what is in my plate. I wonder if it has to do with the booster magnifying its not a big difference.
    Reply With Quote  
     

  4. #14  
    Senior Member
    Join Date
    Oct 2016
    Posts
    384
    Why is it important for the 3d camera to be where the scanned camera is? Didn't you say earlier the goal is to composite onto footage?
    Reply With Quote  
     

  5. #15  
    Member
    Join Date
    Aug 2015
    Location
    UK
    Posts
    80
    if everything is modeled to scale its important If all variables matched and I place my 3d camera where the physical camera was it should match reasonable. It kind almost do now but Im curious about the slight mismatch when I use the pocket lens and speedbooster combo. It's my first time using a speedbooster.
    Reply With Quote  
     

  6. #16  
    Senior Member
    Join Date
    Oct 2016
    Posts
    384
    Not necessarily. The paper Aaron linked explains it well beyond the depth necessary to grasp the concept. When you’re building and compositing with 3d spaces, it’s best to use an approximation to get your perspective in the ballpark and then wrangle the rest by eye. There are tons of variables that will make your calculations inaccurate that aren’t available in any modeling or compositing software. Hell even the manufacturer specifications aren’t completely accurate - at least not to a degree for the type of workflow you're attempting.
    Reply With Quote  
     

  7. #17  
    Member
    Join Date
    Aug 2015
    Location
    UK
    Posts
    80
    when we work on VFX for movies we get lidar, photoscan and take extensive surveys of a set to reconstruct it as close as we can, we don't want to build the same set from the perspective of each camera. And we do get very close by capturing the cameras location angles etc when we have everything dialed in. in my case, it might be within the margin would be interesting to do the same test without a speedbooster to see where I end up. I dont have a native m43 lens
    Reply With Quote  
     

  8. #18  
    Senior Member LochnessDigital's Avatar
    Join Date
    Jul 2014
    Location
    Tucson, AZ
    Posts
    647
    Quote Originally Posted by Meshman View Post
    when we work on VFX for movies we get lidar, photoscan and take extensive surveys of a set to reconstruct it as close as we can, we don't want to build the same set from the perspective of each camera. And we do get very close by capturing the cameras location angles etc when we have everything dialed in. in my case, it might be within the margin would be interesting to do the same test without a speedbooster to see where I end up. I dont have a native m43 lens
    You should also try it with multiple lenses of the same focal length and even some zooms set to the same focal length. You'll start to notice variations because of this. So when it comes to modeling, if you're not modeling the entire lens design, then the best you can do is an approximation.

    I learned this all when exploring the larger format "look"/"perspective"/"mojo" by cropping a sensor to 2x and comparing lenses that were half the original focal length.

    While trying to find this often talked about perspective change, what I found is that even the same focal lengths across the same format size showed noticeable perspective changes. On subjects very close to the camera, the differences in each lens became more apparent, but objects further away showed less change. This is when I learned that each lens is drawing its own virtual "position" or center of perspective and simply lining the sensor plane for all lenses was not going to allow me to match the perspectives exactly.





    Aaron Lochert
    Reply With Quote  
     

  9. #19  
    Senior Member
    Join Date
    Oct 2016
    Posts
    384
    This is why Cookes are so good! They customized the way the lens is distorted on purpose, instead of trying to clean all of it.
    Reply With Quote  
     

  10. #20  
    Member
    Join Date
    Aug 2015
    Location
    UK
    Posts
    80
    yeah sure you can even see focal length shift when focusing some lenses so I guess why my lens speed-booster combo is about 10 cm off in world position to get to the correct perspective in the 3d program is not so strange
    Reply With Quote  
     

Similar Threads

  1. Frame / speed math for syncing slow motion
    By whoatemyteeth in forum Cinematography
    Replies: 3
    Last Post: 04-22-2019, 08:35 AM
  2. Help With Metabones BMPCC Math
    By gj91 in forum Lens Discussion
    Replies: 39
    Last Post: 05-17-2015, 01:41 AM
  3. Who's good at math? Crop factors for windowed modes on URSA Mini?
    By FrissionMedia in forum General Discussion
    Replies: 5
    Last Post: 04-20-2015, 08:30 PM
  4. Shutter angle math according to BM: 180 = 360
    By Steve Wake in forum General Discussion
    Replies: 27
    Last Post: 10-25-2013, 02:35 AM
  5. Speedboster + crop sensor lens for pocket?
    By imdjay in forum Lens Discussion
    Replies: 2
    Last Post: 05-12-2013, 09:57 AM
Bookmarks
Bookmarks
Posting Permissions
  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •