Site Sponsors
Results 1 to 24 of 24

Thread: Getting the camera parallel to the building facade?

  1. #1
    Senior Member
    Join Date
    Jan 2010
    Posts
    431
    Post Thanks / Like

    Getting the camera parallel to the building facade?

    I have to do some frontal images of a row of buildings, and am struggling with getting the camera parallel to the facade. It seems the slightest degree off leads to a perspective slope away to one side or another. Is there some simple trick to getting a camera in parallel alignment with a flat fronted architectural subject you face?

    I'm stitching 3 vertical images together, and one side or another often reveals I didn't get it quite right! I know I can do horizontal keystone in C1 if needed, but I'd like to get it right in the camera before I resort to that (plus it makes stitching alignment difficult when you keystone 3 images by differing amounts)

    Thanks in advance

  2. #2
    Member
    Join Date
    Jun 2013
    Location
    US
    Posts
    41
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    I suppose a lot of this depends on the particular camera that you're using.

    With a view camera, spirit levels and grids on the ground glass are helpful. My 5D3 also has an electronic level which is useful for this.

    Since you're stitching, even a slight variation between images would be noticeable. Lightroom has a perspective adjustment which might help.

  3. #3
    Senior Member
    Join Date
    Jan 2010
    Posts
    431
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    Alpa Max.

    Its not the vertical/horizontal thing that's the issue, its the parallelism (right word?). You can get the camera level and vertical easily, but there is nothing to tell you that the rotation angle you set it is precisely parallel to the long frontage you face. If you are out by just a few degrees, the building slopes away.

    Yes there is keystone correction in the software of PS, Lightroom and C1 - they all have it, but its better to get it right from the get go.

  4. #4
    Super Duper
    Senior Member
    Shashin's Avatar
    Join Date
    May 2011
    Location
    Florida, USA
    Posts
    4,501
    Post Thanks / Like
    Images
    141

    Re: Getting the camera parallel to the building facade?

    Are you sure these buildings are straight. Old architecture is not always level and plumb.

    Can you get to the facade? Find the center of the building at ground level and place a mirror on that or parallel to that. Position the camera so that you can see its reflection in the center of the mirror. You just need a small mirror. That would the simplest method.

  5. #5
    Member
    Join Date
    Jun 2013
    Location
    US
    Posts
    41
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    Quote Originally Posted by narikin View Post
    Alpa Max.

    Yes there is keystone correction in the software of PS, Lightroom and C1 - they all have it, but its better to get it right from the get go.
    Agreed. I'm just not aware of a method that is precise in the field and won't require further tweaking as you do the stitch.

    Doesn't the same grid that you use to determine being parallel in the vertical plane also apply to the 'rotational' aspect?
    Likes 1 Member(s) liked this post

  6. #6
    Workshop Member
    Join Date
    Nov 2007
    Location
    Brooklyn
    Posts
    4,043
    Post Thanks / Like
    Images
    1253

    Re: Getting the camera parallel to the building facade?

    center yourself and measure to L and R edges

  7. #7
    Senior Member stephengilbert's Avatar
    Join Date
    Nov 2007
    Location
    Santa Monica, CA
    Posts
    2,275
    Post Thanks / Like
    Images
    2

    Re: Getting the camera parallel to the building facade?

    "center yourself and measure to L and R edges"

    Another use for the laser distance meter. The enablers never sleep.
    Likes 3 Member(s) liked this post

  8. #8
    Member
    Join Date
    Jun 2013
    Location
    US
    Posts
    41
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    Edit to my previous post.
    I assumed (and probably incorrectly) that you had ground glass focusing. Looking at your camera, it appears not to be the case so my comment about the grid doesn't apply.

    How about shooting tethered, using live view->loupe->grid overlay?

  9. #9
    Senior Member
    Join Date
    Jan 2010
    Posts
    431
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    The digital back (Phase IQ series) has an optional grid that can be switched on. I'll try that, but it's still guesswork deciding what is directly & exactly opposite you in a building/street. Worth a try though!

  10. #10
    Workshop Member
    Join Date
    May 2008
    Location
    NYC
    Posts
    3,275
    Post Thanks / Like
    Images
    7

    Re: Getting the camera parallel to the building facade?

    +1 measuring left/right features with distometer.

    It will place you parallel to the facade with essentially perfect accuracy - ASSUMING the features of the building are square and symmetrical.

    Or if you have an Arca with an eModule Cloud you can use that instead.
    Doug Peterson , Digital Transitions | Email
    Dealer for: Phase One, Mamiya Leaf, Arca-Swiss, Cambo, Eizo, Profoto
    Office: 877.367.8537. Cell: 740.707.2183

  11. #11
    Senior Member stephengilbert's Avatar
    Join Date
    Nov 2007
    Location
    Santa Monica, CA
    Posts
    2,275
    Post Thanks / Like
    Images
    2

    Re: Getting the camera parallel to the building facade?

    I hate to disagree with Doug or John, but wouldn't measuring the two edges of a building only guarantee that the camera is in the center? Couldn't it still be askew?
    Likes 1 Member(s) liked this post

  12. #12
    Member
    Join Date
    Jun 2013
    Location
    US
    Posts
    41
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    I took the question to mean that you were having difficulty in determining that you were absolutely perpendicular to the row houses so as not to introduce any keystoning in the horizontal plane, i.e. left to right.

    Certainly if you could accurately measure the sensor plane to a uniform surface of the building would be a solution. I'm not aware of any convenient tools to measure that accurately over the distance that you are most likely from the building though. Not sure what a distometer is, as I've only heard that term in an optometrists office? I'd love to know more about what it is.

    My assumption is that using the grids (I think they are even movable as well as scalable on that back?) would get you close or at least within the margin of error of what you can see. Would still most likely require some transformation in PS or other software at scale to make it really accurate for the stitch.

  13. #13
    Senior Member stephengilbert's Avatar
    Join Date
    Nov 2007
    Location
    Santa Monica, CA
    Posts
    2,275
    Post Thanks / Like
    Images
    2

    Re: Getting the camera parallel to the building facade?

    Leica's laser rangefinders are called "Distos." (http://ptd.leica-geosystems.com/en/index.htm)
    Last edited by stephengilbert; 15th June 2013 at 21:18.

  14. #14
    Senior Member danlindberg's Avatar
    Join Date
    Aug 2010
    Location
    Spain & Sweden
    Posts
    1,196
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    Measuring left and right with the disto is ofcourse theoretically correct. But in practical doing I have found it quite cumbersome and difficult. Instead I am using a 'low tech' solution that works splendid for me.
    Since I got the Alpa Sports Finder (basically just a frame) I am using the upper and lower horizontal bar to line up the roofline and baseline. Squinting my eyes and stepping back a little from the cam/finder shows even the slightest skewed position. Especially when you can jogg between two reference lines.
    I do not suppose that you have the sports finder so what you could atleast just try is to use the whole top of your Alpa Max as a reference line (levelled of course), lower yourself so that you are looking sort of right through this line/top of cam and turn the whole cam left and right until you are seeing the same amount of your building (and its reference line) far left and far right.
    Alpa FPS MAX TC | Alpagon 32Hr | Helvetar 75 | Schneider 120N | Leaf Aptus II 5 Leaf Credo 60 | www.danlindberg.com
    Likes 1 Member(s) liked this post

  15. #15
    Senior Member
    Join Date
    Jan 2010
    Posts
    431
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    Quote Originally Posted by FoodShooter View Post
    I took the question to mean that you were having difficulty in determining that you were absolutely perpendicular to the row houses so as not to introduce any keystoning in the horizontal plane, i.e. left to right.
    Correct.

    I think the easiest way might be to use rising front to set one of the buildings natural horizontal lines, like an upper row of windows, against one of the DB grid lines, and see if it stays true left to right. Any angular misalignment would reveal itself in this line veering off, as it keystones away or towards you, assuming the camera is leveled correctly. Sliding the back left to right (on a stitching camera) would increase any horizontal keystoning's appearance, and make it easier to spot.

    Thanks for all the suggestions here, people. Food for thought!

  16. #16
    Senior Member
    Join Date
    Jan 2010
    Posts
    431
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    Quote Originally Posted by danlindberg View Post
    Measuring left and right with the disto is ofcourse theoretically correct. But in practical doing I have found it quite cumbersome and difficult. Instead I am using a 'low tech' solution that works splendid for me.
    Since I got the Alpa Sports Finder (basically just a frame) I am using the upper and lower horizontal bar to line up the roofline and baseline. Squinting my eyes and stepping back a little from the cam/finder shows even the slightest skewed position. Especially when you can jogg between two reference lines.
    I do not suppose that you have the sports finder so what you could atleast just try is to use the whole top of your Alpa Max as a reference line (levelled of course), lower yourself so that you are looking sort of right through this line/top of cam and turn the whole cam left and right until you are seeing the same amount of your building (and its reference line) far left and far right.
    Good suggestion Dan, and kind of similar to what I propose doing.

    BTW: is there any piece of the Alpa system you don't own?! It seems they showed you the catalog, and you said "Yes"?

  17. #17
    Senior Member danlindberg's Avatar
    Join Date
    Aug 2010
    Location
    Spain & Sweden
    Posts
    1,196
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    Quote Originally Posted by narikin View Post
    BTW: is there any piece of the Alpa system you don't own?! It seems they showed you the catalog, and you said "Yes"?
    I don't have a film back.......


    Alpa FPS MAX TC | Alpagon 32Hr | Helvetar 75 | Schneider 120N | Leaf Aptus II 5 Leaf Credo 60 | www.danlindberg.com

  18. #18
    Workshop Member
    Join Date
    Nov 2007
    Location
    Brooklyn
    Posts
    4,043
    Post Thanks / Like
    Images
    1253

    Re: Getting the camera parallel to the building facade?

    stephen; if the two distances were the same, the camera would be in the right (and unique) spot. you would then have to pan the camera to center the L/R sides of the image in the view. of course this method would require you be centered as well as parallel.

    using the mirror method will get you parallel with no other constraint

  19. #19
    Senior Member
    Join Date
    Nov 2008
    Location
    Chicago
    Posts
    550
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    mirror method (assuming its parallel to building plane) is quite clever and simple.

  20. #20
    Senior Member
    Join Date
    Dec 2010
    Location
    Warks, UK
    Posts
    550
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    Quote Originally Posted by dougpeterson View Post
    +1 measuring left/right features with distometer.

    It will place you parallel to the facade with essentially perfect accuracy - ASSUMING the features of the building are square and symmetrical.

    Or if you have an Arca with an eModule Cloud you can use that instead.
    This would work if you can guarantee to compose symmetrically relevant to the features or edges.

    ...better to resort to a bit of elementary mathematics...

    use a Pythagoras right angled triangle (e.g. 3,4,5)

    pick a point in the centre of a plane, opposite the desired camera position

    pick another point on the edge

    measure the three sides of the triangle (including the distance along the face of the plane) and adjust until you are perpendicular to the centre point...

    then, if the middle of the image is on the centre point, you know you are perpendicular

    ...this, of course, assumes that your lens is perpendicular to your film or sensor plane, so, if you cannot accurately zero your movements it will not work

    The points might have to be above doors or windows, but, if they were at the same level it would not matter

    ...or use the digital compass on your iPhone... as metalwork in the building might effect the accuracy of a magnetic compass, digital is better... you do not need reference straight edges on the camera - just

    take a bearing along the front of the building,

    take a perpendicular bearing from the camera position

    note which part of the building is opposite the camera and point the camera at it.

    if you cannot set the camera movements to zero, take a bearing across the back of the camera, parallel to the sensor


    ( I have not posted for a while as I have been ill, but, last week, they got my heart to beat in Normal Sinus Rhythm, so I am now officially normal)

  21. #21
    Workshop Member
    Join Date
    May 2008
    Location
    NYC
    Posts
    3,275
    Post Thanks / Like
    Images
    7

    Re: Getting the camera parallel to the building facade?

    Quote Originally Posted by dick View Post
    ( I have not posted for a while as I have been ill, but, last week, they got my heart to beat in Normal Sinus Rhythm, so I am now officially normal)
    Glad to hear!
    Doug Peterson , Digital Transitions | Email
    Dealer for: Phase One, Mamiya Leaf, Arca-Swiss, Cambo, Eizo, Profoto
    Office: 877.367.8537. Cell: 740.707.2183

  22. #22
    Senior Member ondebanks's Avatar
    Join Date
    Feb 2009
    Posts
    518
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    Quote Originally Posted by stephengilbert View Post
    I hate to disagree with Doug or John, but wouldn't measuring the two edges of a building only guarantee that the camera is in the center? Couldn't it still be askew?
    That was my exact thought also. Centering is a separate problem to parallelism or "squaring on".

    The mirror method seems best. When we're aligning a telescope system and need to square on the focuser or an instrument attached to it, we use a derivative of the mirror method, such as a collimating eyepiece or laser collimator.

    Building on this idea...if you wanted to be really spot-on, you could rig an adapter to hold a laser collimator along your camera's optical axis, and swivel the camera until the returning beam (reflected from the mirror on the building) landed straight back on the camera.

    Ray

    (Damn, I should patent that!)
    Likes 1 Member(s) liked this post

  23. #23
    Workshop Member
    Join Date
    Nov 2007
    Location
    Brooklyn
    Posts
    4,043
    Post Thanks / Like
    Images
    1253

    Re: Getting the camera parallel to the building facade?

    or look at all the mirror solutions referred to...or use the enlarger aligner laser dot, or...

  24. #24
    Senior Member
    Join Date
    May 2011
    Location
    Hong Kong
    Posts
    691
    Post Thanks / Like

    Re: Getting the camera parallel to the building facade?

    If you're struggling to see it, you should check the accuracy of your levels. Assuming your levels are accurate then you should see when it is correct as you look through a finder with properly marked lines or use a groundglass grid or liveview grid (or align with the edges). If you have a panning head, as you swing your camera around you should see the parallelism error and be able to locate the mid-point, even you are off-centre to the building.

    It's really no harder than that (unless you have Shashin's identified problem of a building that sags or isn't straight in the first place). Practice on something square at home, or a modern building.

    And don't feel bad about getting it slightly wrong occasionally, even Ansel Adams did so.
    Likes 1 Member(s) liked this post

Tags for this Thread

Posting Permissions

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