Skip to content

Dome ports for the Z8: some measurements and an interactive planner

Featured Replies

4 hours ago, Chris Ross said:

Another consideration is that depth of field is 1/3 in front of focus point and 2/3 behind for the most part, though in macro it distributes closer to 50/50 if I recall correctly.

That's correct.

According to my depth-of-field calculator, the depth-of-field ratio only begins to shift very slowly towards 1/3 - 2/3 starting at a focal length of 70 mm.

Keldan
  • Author
14 hours ago, dentrock said:

Fantastic write-up! Thanks!!!

But still curious as to whether the Howshot 140 in fact is slightly more of a full half-hemisphere than its Nauticam 140 equivalent, as appears to be the case from your comparison photo in this thread above. This would make Howshot a better choice as a 140, to eliminate vignetting and improve alignment in marginal situations.

Thoughts?

Both domes are almost identical. Their radii agree within the measurement uncertainty (Howshot 68.5 mm, Nauticam 69 mm). The Howshot I measured directly; its visible glass height is 56 mm. For the Nauticam 140 I have no direct measurement: I used Interceptor121’s published values and checked them on a photo Dave posted here. In the photo the frame hides where the glass starts, so the height of 63 mm is measured from the base of the mount, and a few millimeters have to be subtracted. That is why the Nauticam currently looks slightly more hemispherical (170° vs. 160° maximum angle of view from the dome center), but this is an upper limit. The data are now in Table 4 of the article, and the planner uses the same values. If someone can send me the exact visible glass height, or a sharp side photo with a ruler showing where the glass leaves the frame, I will update both the table and the planner.

5 minutes ago, Seewolf said:

Both domes are almost identical. Their radii agree within the measurement uncertainty (Howshot 68.5 mm, Nauticam 69 mm). The Howshot I measured directly; its visible glass height is 56 mm. For the Nauticam 140 I have no direct measurement: I used Interceptor121’s published values and checked them on a photo Dave posted here. In the photo the frame hides where the glass starts, so the height of 63 mm is measured from the base of the mount, and a few millimeters have to be subtracted. That is why the Nauticam currently looks slightly more hemispherical (170° vs. 160° maximum angle of view from the dome center), but this is an upper limit. The data are now in Table 4 of the article, and the planner uses the same values. If someone can send me the exact visible glass height, or a sharp side photo with a ruler showing where the glass leaves the frame, I will update both the table and the planner.

The n120 NA-140mm visible glass height from the base is 55mm.

  • Author
8 hours ago, Chris Ross said:

It is a lot of work to put this together, not least putting into an article so we can follow along, so thanks for sharing this.

A couple of suggestions based on my reading through the article, it seems easy to follow but I did find as I was reading certain points I was thinking what about.....? but it was then covered a paragraph or so below. For example the statement: " A dome avoids this: a ray that crosses the glass at a right angle is not bent, so the lens keeps its full angle of view." Then it is explained this only happens when the entrance pupil is centred properly a couple of paragraphs below.

Also where you say:"Difficult are foreground close to the dome in the corners and concave scenes such as caves; there only stopping down helps." Then talk about where to focus which also helps. Another consideration is that depth of field is 1/3 in front of focus point and 2/3 behind for the most part, though in macro it distributes closer to 50/50 if I recall correctly. So stopping down and focusing on something a bit closer is a good combination.

On the tool I'm not sure how it handles vignetting, it seems like some of the extension combinations would likely vignette with the 8-15 fisheye at least on the Nauticam housings, but it's not clear to me where this is reported in the tool.

Thanks for reading it so carefully and for the useful comments. I have changed the article accordingly.

Structure: “A dome avoids this” now says right away that this only holds when the dome center sits at the entrance pupil, with a pointer to the figure that shows what happens otherwise. The point about close foreground and concave scenes (caves, overhangs) needing more stopping down is now in the section on corner sharpness.

The 1/3–2/3 rule: behind a dome it doesn’t hold. The lens doesn’t focus on the subject but on its virtual image, and the dome squeezes everything from the glass to infinity into a zone only about three dome radii in front of it. Depth of field is roughly symmetric around the focus point in terms of the virtual image, but in real distances underwater almost all of it lies behind the focus point. An example from the planner’s model: Z 17–28 at 17 mm in the Sea&Sea 170AR at f/11 (standard circle of confusion). With the focus at about 0.7 m, everything from 0.3 m to infinity is within the depth of field: 0.4 m in front of the focus point and everything behind it. So for scenes that recede into the background, focusing a little behind the main subject gives the best overall result.

Vignetting in the planner: it is there, but the label was not clear. The result “Dome rim from pupil” (now renamed “Vignetting: dome rim seen from pupil”) gives the angle at which the dome rim, or the inner tube of a port, is seen from the entrance pupil. Next to it, a chip compares this with the corner of the FX or DX image: Covers, Borderline or Cuts. In the drawing, corner rays that are cut are shown dashed. Shades are not modeled, and coverage for the fisheye is approximate, so a borderline result should be checked with a test shot.

Fisheye ring length: I agree that fisheyes are less sensitive. By my calculation, a dome that is 10 mm off center gives the 8–15 with TC only 1–2 px of color fringe in the corners, compared with about 4 px for the 17–28 in the same dome. My pool test with the Howshot 140 and a 21.5 mm ring confirmed this, with less than 1 px. So for the fisheye, vignetting rather than exact centering should decide the ring length.

  • Author
3 hours ago, Dave_Hicks said:

The n120 NA-140mm visible glass height from the base is 55mm.

For the Howshot 140, I measured 56 mm, so the radii and visible glass heights are the same. I will update the planner. The Howshot dome is of high quality: "HOYA BK7 optical glass coated with a digital-specific ARD multi-layer coating 99.8% light transmittance". It could come from the same source as the Nauticam, although this is speculation.

OK, thanks. Visible glass height being the same for both domes should mean there is no difference in the two when it comes to vignetting.

Reason I am interested is that Nauticam port chart recommendation for their (old) N120 140 dome with new Sony 8-14 fisheye is 35.5mm extension (i.e. just their standard N100-N120 adapter). This seems a little short to me, given that I estimated the lens EP at 68mm from lens mount in the centre, at 30cm focus.

I can confirm that adding 10mm (total 45mm ext) to this introduces minor vignetting from the shade, at 14mm.

Yet to dive test this to see if adding that 10mm is worth it to improve corners, while requiring me to remove the shade.

Kraken Sports

Account

Navigation

Search

Search

Configure browser push notifications

Chrome (Android)
  1. Tap the lock icon next to the address bar.
  2. Tap Permissions → Notifications.
  3. Adjust your preference.
Chrome (Desktop)
  1. Click the padlock icon in the address bar.
  2. Select Site settings.
  3. Find Notifications and adjust your preference.