Originally posted by daacon 
I love a good Pano Josh and this is certainly one was this veritical or horizontal ?Hand held or tripod ? Whatever the method the results speaks volume.
Thanks Dave - I'm assuming the volumes it speaks are good?
... Here's some details about the process I used - I'm really a novice at this whole pano thing, so please don't take this description as any sort of authoritative treatise on panoramas, and if anyone out there sees flaws in my technique (in addition to the ones I point out myself), please speak up
.
This image is made up of 9 vertical (portrait) orientation shots from the Tamron 70-200mm at 70mm. The lens was tripod mounted using the lens mount (which I think is technically the wrong way to go for panos since it will introduce more parallax as the sensor moves WRT the pivot pt.) and the panning was done with the panning base of the tripod head (the Cullmann you may have?). The challenge with panning with the tripod head is that if your tripod isn't level, you'll pan a diagonal slice rather than a horizontal pan... I don't even have a level
, but I've found I can get the pod fairly accurately leveled by just walking around it, adjusting each leg in turn to level the base of the tripod head by eye ("
I think it's level"). Then I swing the camera through a few test pans watching to see if I can detect any vertical movement...
While shooting, I tend to try to overlap each shot with the previous one by about 1/3 of the frame or a bit less (I have no idea if this is the right way to do this - I should probably read a book or a website on shooting panos
, but this seems to work). To do this quickly (fairly important when the clouds are blowing around and the light is changing rapidly etc.) I pick a point that's about a third of the way in on the side of the frame I'm panning toward (let's say the right, since I seem to work from left to right)... Hopefully there's something identifiable in this area (represented by the green area in the diagram below) - I pick this point - a rock for instance, and I pan until it's in the equivalent area on the left of the frame (the orange) - this way the new shot includes the same rock, and everything left of that is in the last shot too, but the right two thirds are new. Then I lock the panning knob and fire the shot (using the remote release or the two-sec. timer when I'm smart). Once you're sure your pan is level, and you've practiced this a few times, the actual shots can be hammered out pretty quick.
My hasty diagrams...
----------or----------
I tend to favor using a fixed exposure setting in manual mode, but I suppose if the pano will have a very large dynamic range, using an auto exposure mode might actually work out - I'll have to try it sometime.
For stitching I'm using
Hugin, which is free (open source), powerful, and confusing, but the latest release is much more usable than the last one was for me... I've found that even if the program automatically finds matching points for you ("control points" hugin calls 'em) it's important to visually confirm that it's correctly identified the right points in each image. In some cases I've hand matched points for up to twenty images - it took a while but turned out nice. This pano had it's points generated automatically using the panomatic plugin, and all the points were right on, so I'll be trying that setup again.
. Thanks for commenting...
.
Similar Threads 
