Clowns with the 12-24

I really like the composition, lighting and colours.

The corner sharpness is not nice, though. What is your port/dioptre setup? I presume these images were with the 12-24mm?

Alex

Alex I think the same about the corners sharpness....

12-24 with B+W +3 dioptries in the Ike dome port (I dont remember the refer number is the one that work with the 60 and with the 12-24)

Some idea?

Better +4 dioptries?

Thanks!!

The problem is that dome ports produce a curved plane of focus - that is further away from the camera in the corners of the frame. Therefore subjects that are close to the camera in the corners of the frame will be out of focus because they are closer than the point of focus.

A stronger dioptre may help you to focus closer in the corners. Using the correct dome with extention righ for the lens can also solve this problem. But I don't know the ikelite domes well enough to comment in detail.

Camera to subject distance is also crucial here - and the effect is more pronounced when you are close to the subject. It is easy to "hide"/miss the curved field effect when shooting test shots in the pool by using smaller apertures and backing away from the tiles of the pool. That is why is can be difficult to compare different people's tests.

Alex

Alternatively you can compose your pictures without close foreground in the corners - but these images are well composed and I think changing your approach is not a compromise you should make.

Hope this drawing help

Jean -

Your diagrams are great! Is there a not too complicated formula to figure this out based on diameter of dome and lens specs that the manufacture would provide? Seems to me (with the 12-24 anyway) that it has been trial and error with domes and diopters..

M.

Mariano,

FYI,I don't use a diopter with the 12-24mm and Ike domes

Karl

This draws greatly on Jolly's feature

(okay, draw this as you go along ... exercise for the reader)

L ...F......P camera components: film plane, flange, and entrance pupil

____f..e...............D housing/port components: flange, extension, dome

________......r...... pupil-dome distance: interaction

and draw it better than I typed it

It would be easy, if...

1. Port/Housing manufacturers would give

a) The distance between the port flange (call it f) and the camera flange (call it F), for a particular camera/housing combination. Call this distance Ff. This depends ONLY on the housing and its camera.

B) The distance from port flange to dome center (call the dome center D. Measure it from the port flange, call it Df, distance from dome center to flange without an extension. This depends ONLY on the dome.

c) The actual radius of the sphere used for the port ... equal to the max port radius ONLY for hemispherical ports...call this R. This depends ONLY on the dome.

d) The actual distance an extension's flange moves the port (this is usually available). If using an extension, the distance from the outside flange of the extension to the housing flange is ef. This depends only on the extension tube.

Note, no big tables needed ... just 1 dimension for hsg/cmra, 2 for dome, 1 for extension.

and

2. Lens/camera manufacturers would give

a) Distance from entrance pupil (P) to lens flange (F) for all lenses, call this PF.

B) Or, distance from lens flange (F) to film plane (L) for all cameras with distance from les flange to; then we get PF = PL - FL

Note, just one number needed per lens! The pano folks do compile some of these.

Okay, now we know that

*the pupil sits PF in front of the lens flange.

*the dome sits DF = Df + Ff in front of the of the lens flange

* the distance from the pupil to the dome is r = DF - PF

We want r=R; the spherical dome radius should equal the distance to the entrance pupil to minimize problems.

*If r is too big, you're screwed -- the dome is too far out, and you can't bring it closer.

*If r is too small, add an extension of length R-r, the current deficit in distance between the port and the pupil.

That's it.

Now, the pupil location varies for zoom lenses (for the 12-24, a lot), and it is IMPOSSIBLE to have a SINGLE, OPTIMAL dome location/radius with a moving entrance pupil (see link in my comments at the end of Jolly's feature)

Thank's CeeDave, I'm better at drawing than math's, here is another drawing illustrating the relationship between the lens and the dome (sphere).

If you have a sphere (an uncut dome), imagine the center of it where all point are equal in distance, this is the axis or center if you want, now install a lens inside that sphere, remember that a lens of a given focal lenght has an optical axis of it's own, if you center that lens optical axis with that of the sphere (dome), then you have bliss, don't and you end up with blurry edge.

Part of the problem is the zoom, since it's optical center is constantly on the move, so a simple recipe is find the optical axis of your lens at the focal lenght you want to privilege (go to Kaidan web site http://www.kaidan.com/Faqs-Pano.html#5 ) then mark your lens, measure the distance of your dome radius (not the diameter) let's take an 8" dome, it's radius being 4". your dome should sit 4" from the mark on the lens or 3" if your using a 6" dome. the dome physical thickness in theory should be accounted for but I feel it's not enough to worry about. The diopter is to compensate the failure of the close focusing on the part of the lens.

One problem with smaller dome is the short distance the lens need to be able to focus, with an 8" this focusing distance is +/- 16 from film plane, with a 6" it goes down to 12" and a 4" dome is a drastic 8", not too many lens can achieve that close, and when doing split shot you will need to cover that distance, 16", 12" or 8" to infinity depending on your decision.

Hope that helps.

Mariano,  

 

FYI,I don't use a diopter with the 12-24mm and Ike domes

 

Karl

Karl I supose that is the 5503.50 dome port. I try without lens next time....

Good grafics and explanations!

Hi Chris, here is a chart I drew according to your post, check it out and then maybe we can post it as a "print and fill your own info" chart for the member, this could be attached to the text from your post . just pay attention to the part "DOME CENTER" I take it as the axis or center of the dome, but it could be interpreted as the interior lining center position ?

Keep me up to speed.

Jean,

Omigod, that's beautiful! Man, I'd be so much better at my job if I could CAD like you can CAD!!!!

My math/names are a little different than your diagram, as you suggested. I have D being the location of the inside SURFACE of the dome; it is on the lens axis. That is, my point D occurs at the right end of your arrow labeled "Radius (R)".

The only thing that I can think to suggest to make this even cooler is to add lines to compute r, and check vs R. Your drawing makes perfectly clear that r=R is the desired result.

Very unimportant, but my e is the outside edge of the extension, not the extension itself.

If you think this might be used, I'd like to take some time (a day or two) to improve my silly naming of variables and distances a bit and get back to you. I can e-mail or fax you suggestions if you'd like.

Purists might say that R should be the mean of the inner and outer radii of the spheres ... but these domes are pretty thin compared to their radii, so I doubt it matters (esp. since there's often >mm error in knowing the entrance pupil location!).

Thanks hugely ... I am gonna print it right now!

All the best,

Chris

Also note that not all dome ports are perfect hemispheres. The Sea and Sea dome for example is only about 1/3 of a sphere. The center of curvature ® on the Sea and Sea dome does not line up with the mounting flange, it's considerably further back.

The sames is true for the Ikelite dome and a few others.

Cheers

James

James, My understanding is the dome radius is not neccesary the physical portion that we see, but the imaginery part that we don't see are to be accounted for, (the rear part of my drawing is the continuation of the dome in front of the lens), Many moons ago we use to cut 8'' dome for our macro port that appeared to be no more than 4'', (like the cut portion on the drawing appear much smaller than the whole sphere), but nonetheless it is bound to the radius of the whole.

I feel sorry for Mariano, cause we are straying away from is isue, just hope he enjoy's the ride.

Chris, my e-mail is vizart@videotron.ca, I will gladly set up something with you, I like your math's, and BTW it's not CAD it's a software called Corel Draw (Made in Canada Hey!), do wish I coul learn the CAD tough.

Regards

Chris, the best is for you to liquid paper my drawing and re-e-mail it to me with your correction, then we posted it in a separate tread, wish I had as much time for diving as I have for my illustrations :? dreadfull winter....