Your noodling is a lot more practical. Even if your waterfall pano was
printed on a moderately wide carriage printer (24 inches) it would make
a 2x5 foot print at 170 ppi with no interpolation. Viewed from 18-20"
even someone with 20-20 vision should not be able to see any pixellation
at 170 ppi.
I'll have to keep in mind for the future that the horizontal pixel size
for a 5D is 4368 pixels which gives 182 ppi uninterpolated on a 24"
print. And for the 10MP E-things it's not much smaller. With 3648
pixels available you'd get 152 ppi on a 24" print or 203 ppi on an 18"
print.
It should also work well for interiors. A 35mm lens in vertical format
covers about the same vertical height as a 24mm lens in horizontal. And
a 28mm lens in vertical format gives about the same as an 18mm in
horizontal format. Stitching a few of these together will give a very
high quality image with plenty of pixels for perspective correction if
needed.
Yeah, why do we want to do multi-row panos? The heavy duty, single row
Jasper Engineering device is looking better if you're going to shoot
panos of close objects. Maybe not needed outdoors but definitely
indoors. But, if we're going to stick to one row it's also very easy to
build your own custom pano head from wood and especially easy if you
built if for a single camera/lens combo.
<http://photography.about.com/gi/dynamic/offsite.htm?zi=1/XJ&sdn=photography&cdn=hobbies&tm=20&f=10&tt=2&bt=1&bts=0&zu=http%3A//www.creativepro.com/story/feature/12727-2.html>
Chuck Norcutt
Moose wrote:
> Chuck Norcutt wrote:
>> ..... But, using it in horizontal mount, you can simply
>> rotate the camera around the mounting screw in the camera's tripod
>> socket. But as the lens tips up and down there is a small error
>> introduced in the position of the no-parallax-point.
>>
>> Assume a 100mm lens and 21 degree horizontal angle of view on a 35mm
>> camera. (but vertical in this case since the camera is mounted
>> vertically) I'll also assume (just ball park) that the
>> no-parallax-point is about equal to the focal length away from the film
>> plane. To take a 3 row pano we have to tilt the camera both up and down
>> 14 degrees to get a 1/3 field of view overlap between bottom, center and
>> top frames. Final vertical coverage is 49 degrees. Anyhow, tilting the
>> camera up or down by 14 degrees will mis-position the no-parallax-point
>> by about 3mm. Even for relatively close object (like 25 feet) the error
>> is on the order of 1 pixel or less. As usual, YMMV and my calculations
>> may be all wet... but I think it will work.
>>
> OK, I know this is a theoretical construct, but still, I have
> alternatives to explore.
>
> Why go to all that trouble? Put on a 35 mm lens in vertical orientation.
> That gives 54 deg. vertical coverage, and you can take fewer horizontal
> shots, too, if you want; no stitching vertically and less stitching
> horizontally.
>
> So I can see only two reasons to do it the hard way. "Because it's
> there", of something similar, I can't argue with.
>
> The only other reason I can see is that it's for an enormous print that
> will be viewed from relatively close. If it's that sort of project, I'll
> bet the budget would cover a proper two axis panorama rig.
>
> As it happens, my nine shot waterfall panorama was shot vertically at 35
> mm. No calculations, I just set the zoom to give me the vertical
> coverage I needed.
>
> The merged panorama is 10,375 x 4,094 pixels. Assuming it would be
> viewed from at least a few feet, and using quality upscaling, it should
> look great at 72 ppi. That works out to 12 x 4.7 feet.
>
> Just noodling further.
>
> Moose
>
> ==============================================
> List usage info: http://www.zuikoholic.com
> List nannies: olympusadmin@xxxxxxxxxx
> ==============================================
>
>
>
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|