How to shoot photos with CombineZ(version)

COverton Jan 31, 2011

  1. COverton

    COverton TrainBoard Supporter

    1,939
    179
    36
    It is exceedingly important to minimize the field shift of the camera between successive focus depth shots when asking CombineZM or ZP to stack them to render a very sharp focus from macro-close to the far backdrop several feet away.

    What I mean is, you have to teach yourself to look for identifiable items at the corners of the screen or view-finder, anything just visible, and then line up that same view after you have set your focus depth and are ready to trip the shutter. You don't have to be even 98% accurate between shots, but as your error goes between each image, CZP/M must do more integration and ends up discarding more 'trash' all around the image it can manage to sharpen...and it is like a weird picture frame that must be cropped.

    That's for a tripod. Look closely at your selected view, note the placement of a couple of items, remember the view, tilt the camera and take your depth, half-depress the shutter release, re-aim, lock, and trip the shutter. A timer should be used so that you don't impart a shake when you depress the button. Repeat for as many shots as you think you'll need to get a sharp view all the way through the viewed diorama.

    Can't use a tripod practically?

    Now the tricky part...no tripod use is possible because you have to set the camera down close to the layout's surface to get the human-level view you want. And that view is well inside the fascia anyway, and hard to reach.

    What I have done is to fashion a platform onto which I can slide the camera tight against immovable, not easily movable, smooth abutments, one behind and one on the far side of the camera. In my case, I use the smooth aluminum jig for turnouts that come from Fast Tracks. Hey, multi-use is the name of the game in this hobby!!

    I prop up the jig, smooth side up, with bits of cork roadbed, wedges of wooden door jam shims...whatever will hold it steady (no rocking!). I also have the points filing jigs for a straight #8 turnout and a #6 double-slip. These are heavy machined steel blocks. I place their smooth sides facing the camera spot, and back them with real railroad spikes, also quite heavy.

    The idea for no-tripod use is to take up the camera, find something else close to your desired focal distance, half-depress, then fully depress and the timer will start. I always use 10 seconds. That gives you time to restore the camera to its accurately aligned aiming point slowly and deliberately, just as you would walking away from a lit fuse. Because I use halogens in two separated and parallel tracks, I cast shadows if I don't also back up and squat. This all takes time, folks, and I usually have to position the camera far back and high enough that I have to climb a three-step stool.
    More time. So use 10 seconds.

    I hope this is easy to understand. The platform must not budge, but neither can the aligning back and side-stops you use, whatever they may be. The best way to accomplish the re-positioning is by sliding the camera against them using a side-slip so that it nests against both precisely the same way each time you place it.

    Wait for the beeps and shutter to go, climb over and retrieve, pick your next target depth, set the auto focus, trip the shutter, and on to exposure # whatever.

    That's it.

    It is indeed tedious, but surely the results are evident. You do get quicker, as in everything you must master. It is to the point now where I can stage a shot, set up the platform, take a test image, adjust platform for optimal field, change camera exposure time, etc, and have a full shot and process by CZP inside of 30 minutes.

    Adding all the steam and stuff using Sagelight takes me another hour at least.
     
  2. COverton

    COverton TrainBoard Supporter

    1,939
    179
    36
    Shoot...I cant edit the title of the post.
     
  3. COverton

    COverton TrainBoard Supporter

    1,939
    179
    36
    A picture is worth 1000 words, so here is the latest WPF photo's platform. The jig is much too long to balance safely, so I used the smooth plastic case for some needle files. The rest should be obvious based on my description in the first post.

    [​IMG]
     
  4. Mark Watson

    Mark Watson TrainBoard Member

    6,000
    1,317
    85
    I've tried focus stacking before but never got the technique of incrementing the auto-focus down.

    After reading your technique, I decided to try again with better success. :D

    Here's what a single photo would look like.
    [​IMG]

    And the stacked focus result (9 individual photos combined)
    [​IMG]
     
    Last edited by a moderator: Feb 7, 2011
  5. COverton

    COverton TrainBoard Supporter

    1,939
    179
    36
    Very good, Mark. The depth is excellent. Was my explanation about the closer subjects, say 1/3 of the way between the lens and the nose of the diesel, clear enough for you? I didn't cover that above, but in the WPF thread.
     
  6. Mark Watson

    Mark Watson TrainBoard Member

    6,000
    1,317
    85

    Thanks Crandell. Yes, that explanation was crystal. For this photo, however, the first focus point was the nose of the point loco. As you can see, there's nothing interesting in between. :p

    Tomorrow it's supposed to be sunny. I think I'll take my showcase module up to the roof and try to re-create this photo. :)
     
  7. COverton

    COverton TrainBoard Supporter

    1,939
    179
    36
    That would be a great candidate, Mark. Personally, I would seriously consider posing a humanfigure with his/her back to the camera and appearing to be intent upon this earth-shaking monster roaring past. If nothing else, it would be a great opportunity to hone your skills at getting the closest shots in focus sufficiently discretely to get the foreground nice and sharp. In fact, you could experiment with the rear fender and bumper of an appropriate car in the lower left corner of that photo.
     
  8. Bruce-in-MA

    Bruce-in-MA TrainBoard Member

    995
    1
    23
    Nice tutorial.:thumbs_up: I also use Combine ZP and find that tripods are generally useless on most points on my layout. My camera 'zig' is a bit more primitive - a block of wood for the base and a little duct tape to keep the camera fixed on it.

    I'm still trying to figure out ways to get my camera into certain areas on my layout. I wish I had a large robotic arm to mount my camera on. :tb-wink:

    Actually, I wish digital cameras would just integrate this directly into the camera. Then you could just set it down and press one button and let it work through a set of shots by itself - and then combine them in camera. Someday...
     
  9. Wolfgang Dudler

    Wolfgang Dudler Passed away August 25, 2012 TrainBoard Supporter In Memoriam

    3,794
    352
    49
    I can't change zoom in Macro, I have to use wide angle only.
    Therefore if I want to go very close I have to take ALL pictures wide angle. But you have anyway to cut the stacked picture.

    Wolfgang
     
  10. Mark Watson

    Mark Watson TrainBoard Member

    6,000
    1,317
    85
    I believe this option already exists using a webcam... I dont have any sources though.
    (Just something I think I read once). :)
     
  11. Bruce-in-MA

    Bruce-in-MA TrainBoard Member

    995
    1
    23
    I have the same restriction. To get around it, I take all the shots in wide angle (and macro), run them through Combine, and then take the output into a photo editing package and crop it. So long as the focus of your shot is not too far in the background, the cropped shot should turn out okay.

    I hope that would signal a trend, and it will soon be found as a standard feature in Macro modes.
     
  12. COverton

    COverton TrainBoard Supporter

    1,939
    179
    36
    Okay, I understand you now, Wolfgang.

    With my Powershot A710is, I can set a bit of zoom, say 15%, and it will still allow the macro to be used. The only difference between that and wide field is that the point of closest possible auto-focus moves out quite a bit further when I have pre-loaded a bit of zoom. If I run the lens out to about X3, or halfway through the optical zoom range for the A710is, I can't use macro mode either.

    And as you and Ray must do, and Mark, every rendered sharp photo must be cropped judiciously around the entire perimeter to get rid of the 'weird effects' frame that surrounds the "keeper" portion of that image.*

    *For those who are wondering what this is about, the programme integrates all the frames, discards a bunch of unusable stuff, fills gaps by more integration, discards some more, and then reveals on your screen a very sharp image right to the depth of your last focus. However, at the periphery of this sharp image is a 'reflected' portion that must be cropped. If you had rails extending down close to the corner of the image, at about 1/2" toward the edge they will appear mirrored, and look like a chevron forcing the rails to kink sharply and rise at an incident angle. To get rid of this unwanted artifact of the generative figuring, you simply open the crop utility of your favourite image processor, and take the crop boundaries right up to the kink on each side of the image. What is left is your 'keeper' part. So, in effect, you always lose the outer 5-8% of your frame typically. It's no biggie...most of us routinely crop images to bring objects to a larger/closer size anyway, and that discarding step is merely in concert with something we'd all probably do anyway.
     
  13. COverton

    COverton TrainBoard Supporter

    1,939
    179
    36
    I guess I should just show a photo of this mirrored periphery so viewers can see the artifact I mentioned above. In this image, notice the rusty rails. Once you know how to look for this croppable artifact, you can go around the perimeter of the image and see the effect on things that are not linear as the rails are. Note also that the thickness of the stuff that should be cropped on one side of the image is not necessarily found in symmetry on the other sides. Sometimes one or more sides are thinner.

    In this image, it can also be seen that I cloned out a lot of undesirable elements, such as window sill, tree shadows on the backdrop, etc.

    [​IMG]
     
  14. Mark Watson

    Mark Watson TrainBoard Member

    6,000
    1,317
    85
    Well Murphy was helping me out with today's attempt to re-create my first AC-12 photo. First I hauled up my showcase module and accessories and set everything up. Then I came back down to get the locomotive, cars and camera equipment. Well by the time I returned, the shadow from the building next door reached my set! :tb-hissyfit:

    Luckily there was a patch of daylight still down towards the front of the building, so I moved everything out of the shadow and into the sun.. slash.. wind tunnel. :tb-hissyfit:

    Luckily it wasn't too hard of a breeze and I was able to shield the light/loose stuff fairly well.

    So without further ado, here is the result. :)

    [​IMG]
     
  15. COverton

    COverton TrainBoard Supporter

    1,939
    179
    36
    That's pretty darned cool, Mark. It looks nice and sharp all the way through.

    What are your feelings about it?
     
  16. Tracy McKibben

    Tracy McKibben TrainBoard Member

    1,181
    98
    24
    I messed around with this a bit a few months ago. This photo resulted from some work with CombineZ and The Gimp:

    [​IMG]
     
  17. Mark Watson

    Mark Watson TrainBoard Member

    6,000
    1,317
    85
    Thanks Crandell. I made it from a 14 image stack, but I still got quite a bit of ghosting, particularly around grab irons and where thin objects stand out a distance from their background. Once the stack is combined though, it makes cloning much easier. :p


    Tracy, I remember that one. If that tree wasn't so dominating in size, I think you could have passed that one off as a prototype! :D
     
  18. Hytec

    Hytec TrainBoard Member

    13,965
    6,903
    183
    My wife received a Canon Powershot SD980 for Christmas that I believe has this panoramic feature built into its firmware, though it may be a feature of the software that came with the camera. She took 5-6 overlapping images of our living room and selected them for a panorama. The final product is a seamless panoramic image that covers approximately 270 degrees. The firmware (or software) selected common areas of overlap between each successive image and merged them all into a single image. There is a slight distortion at the merged interfaces due to the fisheye effect of the wide-angle lens, but that is hardly noticeable in the overall effect.

    I believe I understand the general theme of this thread, so I hope my comment is not a hijack. If so, please forgive me. However, I believe this feature would be applicable with macro-images of layouts also.
     

Share This Page