Large domes and edge sharpness: sources

1 hour ago, Interceptor121 said:

The issue here is that the camera working distance is 28cm
A lenses like this one commands a very large radius to focus on the glass most likely in the region of 15cm so you are looking ideally at 12” dome
Dome size is driven by working distance lens construction and positioning of entrance pupil
The most important characteristic of a lens is to be able to focus close otherwise when you wrap it in a dome there is very little range for focus and as consequence depth of field
Depth of field however is a consequence not the driver
With rectilinear lenses optimised for flat frame field of curvature becomes predominant and this is why edges blur
With a fisheye lens focussing at 15cm and subject to barrel distortion field of curvature appears to be countered by the distortion that goes in the opposite direction
When I shoot my canon 8-15mm I get better edges underwater than I get to n land and this is because the dome increases depth of field and the lens distortion counters the field of curvature at least this is my theory
For clarity I would not recommend anyone a small dome for a lens that can focus at 28cm in fact I wouldn’t recommend getting that lens at all
Unfortunately even on mirrorless most rectilinear lenses focus between 25 and 28 cm requiring ideally 12” domes or a radius of >15cm


Sent from my iPhone using Tapatalk

I'm not sure I follow your arguments here, my 12-40 Oly zoom focuses at 20cm and performs really well in a 170mm dome.

On fisheyes, I took these shots first a Panasonic 7-14 at f4 and 8mm, second a Panasonic 8mm fisheye focusing really close so I have something out of focus at f4.5.  You can see the in focus area is a straight band at a constant perpendicular distance to the sensor plane. The out of focus field is pretty uniform closer than this band across the frame.

The out of focus area in the fisheye is close to circular around the in focus part.  So the depth of field extends radially out from the entrance pupil.   The lighting is not perfect but it shows what is happening.

I would suggest that part of the reason fisheyes work so well is that in focus  area is a hollow sphere centred on the entrance pupil, roughly the same shape as the virtual image.   It is almost like they were designed with a dome port virtual image in mind.

It's not perfect of course and the in focus area is probably not a symmetric circle and possibly neither is the virtual image so you need to stop down to get good corners but not as much as you do with rectilinear lenses.  And no doubt there are other aberrations from imperfect placement, chromatic aberration etc that add into this.

I'm not sure I follow your arguments here, my 12-40 Oly zoom focuses at 20cm and performs really well in a 170mm dome.
On fisheyes, I took these shots first a Panasonic 7-14 at f4 and 8mm, second a Panasonic 8mm fisheye focusing really close so I have something out of focus at f4.5.  You can see the in focus area is a straight band at a constant perpendicular distance to the sensor plane. The out of focus field is pretty uniform closer than this band across the frame.

The out of focus area in the fisheye is close to circular around the in focus part.  So the depth of field extends radially out from the entrance pupil.   The lighting is not perfect but it shows what is happening.
I would suggest that part of the reason fisheyes work so well is that in focus  area is a hollow sphere centred on the entrance pupil, roughly the same shape as the virtual image.   It is almost like they were designed with a dome port virtual image in mind.
It's not perfect of course and the in focus area is probably not a symmetric circle and possibly neither is the virtual image so you need to stop down to get good corners but not as much as you do with rectilinear lenses.  And no doubt there are other aberrations from imperfect placement, chromatic aberration etc that add into this.
 

What if we use a flat port underwater with a rectilinear lens?

What if we use a flat port underwater with a rectilinear lens?

7 hours ago, stillviking said:

What if we use a flat port underwater with a rectilinear lens?

You get something like this:

Instead of this:

We had somewhere a discussion, whether the Nauticam 230 domeport is a domeport for fisheye lenses or for rectilinear lenses (I do not remember where it was, but it was not completely in-topic, anyhow). It is also not completely in-topic here, but I think this is the right location to post a statement from Nauticam about this issue:

I asked: "I have a question regarding "230mm Optical-Glass Fisheye Dome Port II" - SKU # 18812:  The radius of curvature is 120mm (?), but the diameter is less than 2x120=240mm. Is this now a complete hemisphere as the name implies (Optical-Glass Fisheye Dome..."), is this the deviation from perfect hemisphere neglectible for fisheye use or should one put only rectilinear lenses behind this port?"

Just now Ryan Canon from Nauticam US responds: "You are correct - the dome is not a perfect hemisphere, but we have found in our testing that this slightly smaller segment doesn't create any noticeable image quality issues."

=> So the assumption, that, for practical purpose, the 230 dome is a fisheye domeport, intended for fisheye lenses, is correct.

=> Nevertheless, many prefer the 230 dome (r=12 cm) over the 250, truly rectilinear, dome (r=16cm), also for rectilinear lenses, probably because it is smaller, lighter and cheaper. Also Nauticam recommends the 230 port over the 250 for many rectlinear lenses in their portchards (in another request, some time ago, Ryan wrote me that they recommend the 230 over the 250 for lenses, where the 250 domeport does not bring noteworthy improvement over the 230 (because of size, weight and price)).

=> A smaller, less hemispherical, domeport with r=12cm would give identical results as the 230, but is, at present, not available...

Wolfgang

32 minutes ago, Architeuthis said:

We had somewhere a discussion, whether the Nauticam 230 domeport is a domeport for fisheye lenses or for rectilinear lenses (I do not remember where it was, but it was not completely in-topic, anyhow). It is also not completely in-topic here, but I think this is the right location to post a statement from Nauticam about this issue:

 

I asked: "I have a question regarding "230mm Optical-Glass Fisheye Dome Port II" - SKU # 18812:  The radius of curvature is 120mm (?), but the diameter is less than 2x120=240mm. Is this now a complete hemisphere as the name implies (Optical-Glass Fisheye Dome..."), is this the deviation from perfect hemisphere neglectible for fisheye use or should one put only rectilinear lenses behind this port?"

Just now Ryan Canon from Nauticam US responds: "You are correct - the dome is not a perfect hemisphere, but we have found in our testing that this slightly smaller segment doesn't create any noticeable image quality issues."

 

=> So the assumption, that, for practical purpose, the 230 dome is a fisheye domeport, intended for fisheye lenses, is correct.

=> Nevertheless, many prefer the 230 dome (r=12 cm) over the 250, truly rectilinear, dome (r=16cm), also for rectilinear lenses, probably because it is smaller, lighter and cheaper. Also Nauticam recommends the 230 port over the 250 for many rectlinear lenses in their portchards (in another request, some time ago, Ryan wrote me that they recommend the 230 over the 250 for lenses, where the 250 domeport does not bring noteworthy improvement over the 230 (because of size, weight and price)).

=> A smaller, less hemispherical, domeport with r=12cm would give identical results as the 230, but is, at present, not available...

 

Wolfgang

The 250 dome is even less of a  complete hemisphere

The radius of curvature per specifications is 16 cm so the dome is actually only able to support 102.75 degrees in fact it is called wide angle port

This port should be used for rectilinear lenses that do not focus very close that typically require this curvature

It would not make any sense to use this for a fisheye as the angle of view is very small so this is a good solution for not very wide rectilinear lenses up to 18mm

This port considering most lenses are at least 16mm is really a strange one

So to recap

There are only two full domes in Nauticam catalogue

4.33" acrylic

140mm glass

There are then 3 glass wide angle ports

180mm with 109 degree fov and 11cm radius 

230mm with 147 degrees and 12 cm radius

250mm with 103 degrees and 16cm radius

Of the above only one is kind of suitable for fisheye consider horizontal field of view is 172 degrees so the image is subject to more barrel distortion than a correctly positioned dome

We then have the following acrylic ports

8.5 dome port no specifications provided unclear what the radius and angle of view are

As of today if you want to have a fisheye lens with the same field of view that is offered on land the only options are the first two 4.33" and 140mm no others support the full field of view

The second options seems to be the 230mm to use  fisheye with additional barrel distortion and reduced field of view

The 8.5" acrylic port remains to be investigated it is not clear if it is a full hemisphere or not and what the angle of view is

Why the 230mm would have a benefit over the 140mm dome considering the positioning is suboptimal remains to be seen. I cannot understand why additional barrel distortion improves corners it should deteriorate them

8 hours ago, Interceptor121 said:

Why the 230mm would have a benefit over the 140mm dome considering the positioning is suboptimal remains to be seen. I cannot understand why additional barrel distortion improves corners it should deteriorate them

Possibly because other factors are at play.  In any case Nauticam mark the 140mm dome as most optimised, though the criteria for that is not known.

Fisheyes certainly do not suffer from corner issues to the same extent as rectilinear lenses as the shape of their focal field is different to a rectilinear lens as I noted in a previous post.  A rectilinear lens you can see a parallel plane to the sensor is in focus taking a shot of using focus peaking in the viewfinder.  With a fisheye the focal field is roughly circular so a near perfect match to a curved virtual image.

Yet I also observe that Ikelite recommends their compact dome port 75344 which is a very shallow dome segment as the CFWA solution for the Canon & Nikon 8-15 lenses.  The lens needs be positioned way up close to the port glass well forward of the centre of curvature to avoid vignetting .  I'm not saying for a second that the image including the corners is as "good" as a fully optimised setup, but people use the Ikelite setup and are happy with what they are able to shoot.

There are inconsistencies in Nauticam port chart depending on who tested the set up

On the nikon f chart the best dome for the sigma fisheye is 230 in all others format and lenses is 140mm

There is no reason for that other than mustard said so?

For me getting a fisheye to reduce the field of view and then moving backwards compared to the glass because the dome takes some space is not logical

Fisheye lenses perform terribly in the corners they don’t do well at all topside the focal plane is curved opposite to a dome due to barrel distortion

Maybe this effect is countered by the dome and therefore things improve unclear

If the dome helps improving the fisheye lens then smaller dome would be better but there are considerations of depth of field

I think the confusion here depends on the fact that corners are affected by field of curvature which is similar but different from depth of field

Sent from my iPhone using Tapatalk

Just a personal data point in this thread - I shot my Nikon 16-35 behind a 230mm dome (and 90mm extension) for years, and I was always disappointed with the corners.   Even more so considering both the dome expense and the hassle of traveling with it.  Sure wasn't a payoff.

Then I added the Sea and Sea Internal Correction Lens, which I think was designed for this dome/lens combo.  Or generally 16-35 lenses.

It made a big difference!   Before the ICL I cropped the edges and corners from virtually every shot, no matter how stopped down I was.  With it, I'm shooting at F5.6 - F8 a lot, though for shots where edge-to-edge is both important and where the subject ranges from fairly close to distant I may stop down as far as F14.

I'm now shooting Nikon 14-30 behind the same 230mm dome (50mm extension), and I added the 82mm ICL to it.   Seems to work very well.  (I'd hope so, since the price of the ICL jumped from $400 to $750 in the past couple of years.)

I have one OK example to show.  I suspect it nears the best performance I'm going to get with this combo.  It was shot at F14 and shows a lot of lettuce coral from edge to edge.

https://www.cjcphoto.net/roatan2022/images/page3.htmlk

There are definite limits that I'm still trying to nail down.   Basically, when part of the edge subject is pretty close (I'm trying to learn just how close) and then extends off in the distance, the closer edge is just not going to be sharp.  This is when you actually focus somewhere in the center at a greater distance.  As a hypothetical example, I may have an edge that is 4-5 feet from me and a subject like the one in the link, you get decent edges.  But if that close part of the background is another foot closer, you just can't make it look good while still focused on that more-distant subject.  DOF doesn't cover things, and the results are bad.  But simply moving the camera back a bit can make a huge difference.

2 hours ago, phxazcraig said:

I have one OK example to show.  I suspect it nears the best performance I'm going to get with this combo.  It was shot at F14 and shows a lot of lettuce coral from edge to edge.

Thanks, Craig. That's really useful. Look pretty good, I reckon. That ICL is a chunk of change though especially in addition to the 230 dome.

4 hours ago, phxazcraig said:

I'm now shooting Nikon 14-30 behind the same 230mm dome (50mm extension), and I added the 82mm ICL to it.   Seems to work very well.  (I'd hope so, since the price of the ICL jumped from $400 to $750 in the past couple of years.)

I also have the Z 14-30 + SS ICL + 230mm dome and get good results. I have to adjust the distortion slider a bit but I think that’s due to the lens profile in LR. I paid $600 at Reef Photo. Here is a shot at 14mm:

https://larryhallas.com/portfolio/reef-and-sunball/

This post was about fisheye
For rectilinear lenses the predominant issue is field of curvature
This is influenced by angle of view and where you focus
If you shoot something flat os different than shooting something where the focus point is closer than the other part of the frame
The sea and sea field flattener claims an improvement of 2 stops to obtain equivalent edges but is meant for particular combinations lens port


Sent from my iPhone using Tapatalk