Jump to content

Recommended Posts

Posted

I will be utilizing 2.0x and 1.4x Teleconverters with a fisheye zoom lens behind a dome.


Currently doing the math for optimum dome positioning and sizing.

 

I am under the assumption that the entrance pupil will move by adding a TC.

And I mean by that, not equivalent to the 31mm physicial space that the teleconverter adds between camera and lens.

 

Can you share your thoughts and maybe even the math for this endeavor?

  • Like 1
Posted (edited)
13 hours ago, Adventurer said:

I will be utilizing 2.0x and 1.4x Teleconverters with a fisheye zoom lens behind a dome.

I'd say it's either-or scenario. You'd  need different port extension for each of them

 

13 hours ago, Adventurer said:


Currently doing the math for optimum dome positioning and sizing.

 

I am under the assumption that the entrance pupil will move by adding a TC.

And I mean by that, not equivalent to the 31mm physicial space that the teleconverter adds between camera and lens.

 

Can you share your thoughts and maybe even the math for this endeavor?

 

I'd say the approach might be less scientific - basically push the lens back as far as not to get vignette at widest setting. The dome size is far more important than some millimeters in positioning. Down below left is 140mm glass dome with TC2x and lens at 8mm (=16mm) and correct position and right is 180mm dome with lens at 15mm (no TC) so far back that it already vignetted. Still the CA (see the blue lines) is so much more pronounced even though the right image scenario is sharper even in corners (but there is another variable like the lens was at 15mm and not 8mm and it didn't have TC on it). 

 

I have no real life samples as to what effect this extreme CA would have on the real image underwater. 

Screenshot 2025-02-03 at 8.04.39 AM.jpg

Screenshot 2025-02-03 at 12.38.02 PM.jpg

Screenshot 2025-02-03 at 12.37.46 PM.jpg

Screenshot 2025-02-03 at 12.41.36 PM.jpg

Edited by RomiK
Posted

Thanks @RomiK for posting the examples. For correct optical performance you have to get the dome size and positioning exactly right on a millimeter level.

 

You are introducing quite a bunch of variables by changing dome size and TC at once. 
 

Knowing that Nauticam and Marelux 180mm domes are not full spheres your test was destined to fail. The 180mm dome with conical rear part was simply not made for that lens. The Canon 8-15 FE without the TC is very likely to perform better behind the 140mm dome, which is (almost) full sphere and suitable for fisheye lenses.

Posted

Coming back to the initial question and topic, the theory and some formulas suggest that by using the TC the entrance pupil (common name nodal point) will move into the direction of the cameras sensor plane.

 

So if the Kenko Teleplus HD 2x DGX for Canon has 35mm length, an additional extension ring of just 15mm might be enough; if you keep using the same dome and it was absolute precisely positioned before you added the teleconverter.

 

In this example we assume the entrance pupil moved 19.78mm towards the sensor. This is just an example to give you a general idea on how it works and what might be going on, with the real 15mm port extension not computed nor tested / verified.

Posted

My understanding is that compromises are involved, and with the Canon 8-15mm the perfect theoretical positioning does not necessarily line up with the best extension, depending on what type of photos you want and what dome you use.

 

For close up wide angle shots, because the maximum effect is possible with a small fisheye dome, taking advantage of the minimum focus distance, and placing the dome closer to the lens, makes a big difference to the effect, creating a larger main subject and a better photo than having a longer extension which theoretically improves image quality but reduces the size of the central subject. For a larger dome, which is not ideal for close-up wide angle anyway, this doesn't apply.

 

 

 

Posted (edited)
11 hours ago, Adventurer said:

Thanks @RomiK for posting the examples. For correct optical performance you have to get the dome size and positioning exactly right on a millimeter level.

 

You are introducing quite a bunch of variables by changing dome size and TC at once. 
 

Knowing that Nauticam and Marelux 180mm domes are not full spheres your test was destined to fail. The 180mm dome with conical rear part was simply not made for that lens. The Canon 8-15 FE without the TC is very likely to perform better behind the 140mm dome, which is (almost) full sphere and suitable for fisheye lenses.

 

These were just a few of tests of different configurations I made... gave you those which I thought would be the most polarized in order to deliver the  point - millimeters really don't matter, there are other things in play and we only need to be in a ballpark figure. Like 30mm vs 35mm extension? No real difference. Like 180mm sharper (minus the CA) than 140mm with Sony TC in ideal position? OMG the 180mm was supposed to be awful wasn't it?  And Kenko was even worse🙈 than Sony. But you would have to spend time shooting charts underwater to understand this. Unfortunately the Internet is full of theories with "demonstrative" pictures from Raja Ampat 😂 which tend to muddy the waters somewhat. So I had to make my own tests. Good luck and I am looking forward to your UW test charts and conclusions 🙂

Edited by RomiK
Posted
3 hours ago, RomiK said:

 

Like 180mm sharper (minus the CA) than 140mm with Sony TC in ideal position? OMG the 180mm was supposed to be awful wasn't it?  And Kenko was even worse🙈 than Sony. 

 

Well, it depends what you define as "awful". 😉 This is usually a very subjective expression of the situation. 
More objective would be:

  • miimeter distance of the entrance pupil
  • milimeter distance of the port (required vs available extension rings)
  • stronger image compression
  • able to shoot the lens more oben (F4 or F8 instead of mandatory F11 or F16)
  • priority on zoom flexibility
  • priority on corner sharpness
  • priority on avoiding chromatic aberrations
  • priority on travel weight lightness

In physics and underwater photography there is usually no free lunch.

Meaning there is one death you have to die (as we say in German).

 

My approach will be to verify the theoretical moved position of the entrance pupil, with a paralax test on a nodal rail.

For that I do not need to go diving or get the system wet.

 

Again, thanks for sharing your results @RomiK !

 

However, I think there are combinations that will perform much better considering IQ in your rig.

I hope you will also be able to benefit from the findings in this thread.

Posted
20 hours ago, Adventurer said:

 

Well, it depends what you define as "awful". 😉 This is usually a very subjective expression of the situation. 
More objective would be:

  • miimeter distance of the entrance pupil
  • milimeter distance of the port (required vs available extension rings)
  • stronger image compression
  • able to shoot the lens more oben (F4 or F8 instead of mandatory F11 or F16)
  • priority on zoom flexibility
  • priority on corner sharpness
  • priority on avoiding chromatic aberrations
  • priority on travel weight lightness

In physics and underwater photography there is usually no free lunch.

Meaning there is one death you have to die (as we say in German).

 

My approach will be to verify the theoretical moved position of the entrance pupil, with a paralax test on a nodal rail.

For that I do not need to go diving or get the system wet.

 

Again, thanks for sharing your results @RomiK !

 

However, I think there are combinations that will perform much better considering IQ in your rig.

I hope you will also be able to benefit from the findings in this thread.

 

All good points save for one I would disagree with. You absolutely have to get the system wet (not yourself necessarily 🙂 - I shoot my pictures laying on edge of my home pool shooting down 😉) in order to test as there are things coming to play like refraction and virtual image. The cutting mat is the best water resistant test bed I have found :-). I did try to play with Petzval surface phenomenon but didn't have that much patience to have meaningful result in case you'd like to really dive in 😁. Cheers 

Posted (edited)
On 2/2/2025 at 10:46 PM, Adventurer said:

I am under the assumption that the entrance pupil will move by adding a TC.

And I mean by that, not equivalent to the 31mm physicial space that the teleconverter adds between camera and lens.


 

So I did my parallax testing on a nodal rail to verify / de-myth the whole thing.

 

The result: the above stuff I read on the internet and which seemed to be confirmed by two reasoning AI models is complete bullshit.

 

The NPP / rotation point / entrance pupil will stay exactly where it was!

 

Looking from the frontglas side at the Canon 8-15mm Fisheye F4 L it is stays at 17.98mm (@15mm zoom position) or at  19.12mm (@8mm zoom position).

 

No change if you add a 1.4x TC or 2.0x TC in between.

 

So assuming your dome position was already perfect before adding the TCs you will have to add:

 

18mm port Extension for the 1.4x Kenko

and

35mm port extension for the 2.0x Kenko

 

But as you may very likely have used a too short port extension to be able to use the full zoom range of the native 8-15mm without vignetting, adding a 50mm port extension for Marelux Housings is the way to go for the 2.0x TC. It is very likely to be the same for Nauticam when using their 140mm dome.

 

Edited by Adventurer
  • Thanks 2
Posted
41 minutes ago, Adventurer said:

 

 

But as you may very likely have used a too short port extension to be able to use the full zoom range of the native 8-15mm without vignetting, adding a 50mm port extension for Marelux Housings is the way to go for the 2.0x TC. It is very likely to be the same for Nauticam when using their 140mm dome.

 

 

Thank you for doing the measurement, Adventurer...👍

 

I do not understand what you mean with your last sentence: when a hemispherical dome as Nauticam 140mm is used and the entrance pupil is placed right at the center of curvature, there should be no vignetting - right?

 

=> When I use my Canon 8-15, Sony A7R5, Nauticam housing and 140mm Nauticam domeport with the 30mm extension recommended in the port chards, I cannot see any vignetting, neither with the 8mm circular fisheye image nor with the 15mm 180°diagonal image...

In accordance no vignetting with Kenko 1.4x/2x or Sony 2x TC, when I correct for the length of the TCs by appropriate extensions...

 

As a sidenote, there is a tread made by Dreifish, where he tested different extensions (in 5mm increments) UW. His outcome was that 35mm extension is optically even better than 30mm for the Canon 8-15, based on perspective UW/OW in split shots: 

I tested both 30mm and 35mm, but since I did not see a difference in IQ, I personally still prefer 30mm as the shorter extension is more convenient (with 35mm sometimes I could see part of the shade in the image, probably due to sensor IS and camera movement during photographing)...

 

 

 

Wolfgang

 

  • Like 1
Posted
1 hour ago, Architeuthis said:

I do not understand what you mean with your last sentence: when a hemispherical dome as Nauticam 140mm is used and the entrance pupil is placed right at the center of curvature, there should be no vignetting - right?

 

Well,... yeah I wish it would be like that, but in reality even the Nauticam and Marelux full sphere 140mm domes are missing a few degrees curvature, where their glas is mounted into the aluminum frame.

The most practical thing on the Canon 8-15mm is to keep an eye on the entrance pupil: It's located approximately directly under the red dot when you have the (original Canon) land sunshade mounted. You can keep an eye on it, if it sticks out sufficiently out of the black port tunnel of your underwater housing.

 

So the native lens without teleconverters would need the 140mm Marelux and Nauticam domes with 47,9 mm Extension to be perfectly positioned, but as these two domes finish their sphere a little too early a 35mm extension is the absolute maximum you can use, before dark corners (vignetting) kicks in with these ports. However the fact that we might all be approx 13mm to short with the extensions explains the less optimal performance @RomiK illustrated in chromatic abberation test-shot when he used the lens natively. 

 

Hence the teleconverters shrink the FOV a little, so we can live with non-full sphere domes such as the 180mm with conical portside. However if the TCs decrease the optical quality or the slightly offset dome is to be judged.

 

 

Posted

Nauticam already has a port recommendation for a Canon 8-15 plus 1.4x in the port chart, they just add 20mm of extension and they have a zoom ring for that combination as well.

 

From what I recall and various evidence of what various manufacturers offer on the internet for this lens it seems that fisheye lenses are not particularly sensitive to positioning the dome port correctly.  A prime example is ikelite 8 inch compact dome which is quite a small segment of a sphere and the Centre of curvature is well behind the dome.  It works and the image samples provided look quite OK at first glance.  I'm not saying that you can't or shouldn't try to do better, but this port is a long way out from ideal placement and still provides quite usable images.  Here is a post with examples.

 

GET the Canon 8-15mm Fisheye Before It’s GONE! // Underwater Review &

 

 

 

 

Posted
On 2/6/2025 at 10:05 PM, Chris Ross said:

I'm not saying that you can't or shouldn't try to do better, but this port is a long way out from ideal placement and still provides quite usable images.  Here is a post with examples.

 

GET the Canon 8-15mm Fisheye Before It’s GONE! // Underwater Review &

 

I would not draw any conclusions from the cited article as there is not a single photo in it which wasn't cropped and heavily processed. 😁

Posted
15 hours ago, Adventurer said:

 

I would not draw any conclusions from the cited article as there is not a single photo in it which wasn't cropped and heavily processed. 😁

I think you are somewhat missing the point, this lens performs acceptably behind a dome where the centre of curvature is around 80mm behind the entrance pupil.  Sure the pics in the link I posted are relatively small, but they are enough to show that they produce  pics a lot of people would be delighted with.  It basically shows that the lens is relatively tolerant of misplacement of the centre of curvature.    If you look at the size of this port, the dome section is only about 20mm high so the lens focuses right down on the port allowing you to get very close to your subject - which makes a significant difference for CFWA, the subject becomes all the more prominent in the frame and small differences in distance makes a big impact with a fisheye lens.  But this of course involves a compromise, with a small deterioration in corner performance.

  • Like 1
  • Thanks for your support

    Logo Logo
    Logo Logo
    Logo Logo
    Logo Logo
    Logo Logo
    Logo Logo
    Logo Logo
    Logo Logo
    Logo Logo

     

     

     

     

×
×
  • Create New...

Important Information

Terms of Use Privacy Policy Guidelines We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.