I should start by thanking interested_observer for this thread I don't need no Pano Head... and Paul Bruins for his original concept/webpage.

I have been frustrated trying to take panos which included moving water and a similar comment in the Panorama-orama - Post your Panos thread reminded me of the thread by interested_observer.

Now being on a budget which didn't allow for 6 K-1s and owning a K-x and two 18-55s I thought I'd go for 3 K-x/18-55s and set about acquiring the extras.

Meanwhile I did some CAD concept drawings:



Followed by some mock-ups and a prototype:





After receiving a second K-x I found there was a problem with my drawings (I still don't know what) and the images didn't overlap.

After changing the angles I tried again:



Which showed some promise:



At least the waves lined up

The third K-x arrived:



And I went out to do some testing.

And found the IR remote often failing to trigger all 3 cameras especially in bright sunlight. That makes it a real pain trying to work out which images go together.

Stitching the images has also been more difficult, Microsoft ICE doesn't like them and Photoshop requires more work.

But this is where I'm up to (more later):

[alfa75ts]

I managed to figure out another 3 shots that go together.

The waves match up!

The beach looks like a banana! I guess the sky is bent too.

[35mmfilmfan]

Very interesting approach - nice to see some DIY too. When I was 14 I built my own enlarger, which cost around 12/6d in real money, using my only lens, Meritar 50mm reversed! Managed some decent 10x8s on it too.

If I ever get sufficient funds to acquire more cameras, I may steal (sorry, 'adapt') this idea for my own use. In the meantime, please keep showing the results.

[alfa75ts]

Test shot showing joins. Parallax is still an issue up close, there are duplicate bullrushes? either side of the joins.

Attached Images

 

[alfa75ts]

A bit of an update.

First, I have modified one of my remotes so it has 2 leds which increases the spread and will hopefully trigger all 3 cameras more reliably. If not, I will have to make a strap which places a led in front of each sensor.

Secondly I have thinking about the error in the overlap that puzzled me.

I turns out to be one of scale. My initial drawings (see 1st post) were too close to the cameras where the distance between them had a greater percentage effect on the overlap.

I also realised that the angle between the cameras has to be less than the Angle of View of the lenses.

My initial thoughts were to use an angle of 50°, it looked good when I drew it up but when I tested it there was no overlap so I tried 40°. Turned out when I drew it up a bit bigger at 50° the overlap disappears at about 1 metre.

The overlap at 40° is 14.9% at 100 metres.

Images:

1/ Remote
2/ At 100 metres
3/ In close