On 7/7/2024 at 4:16 AM, DreiFish said:

First -- the 2x TC definitely leads to significant resolution loss (not surprising), but also light transmission loss.

I would like some others to think / verify with me, if the teleconverter needs to be incorporated into the entrance pupil dome calculation. I think yes. So you will get totally different P-I values and flange distances. I think contrary to the canon EF-RF converter, these have to be taken into account.

 

Or does the TC effect and the new flange distance cancel each other out? 

 

I also would point out, that there might a theoretical error for the Canon EF 8-15 when pulling it from optical bench.

 

On 6/30/2024 at 12:59 PM, DreiFish said:

It's not on OpticalBench yet. This is what I've pulled so far from OpticalBench.  

 

The default values there are @ 8mm zoom setting.

If not used with a teleconverter this is the circular fisheye setting and not the diagonal fisheye setting of the lens.

Hence the one that most of the users will rarely shoot, because they do not like a black circled image too often.

 

If you play with the zoom slider on optical bench you get both values.

It will show you 1,63 mm you difference offset for the minimum dome size required.

A minimum dome diameter of just 7,6 cm should therefore be sufficient.

 

But beware!

The theoretical optimal extension on MARELUX for their small dome would be 8,34 cm.

MARELUX is recommending 30mm only. .. I guess beyond this point vignetting kicks in and it's the usual compromise.

However it's worthwile to try a 40mm ring or more to improve IQ if you own it or can borrow it.

 

The RF-EF converter does not change the position of the entrance pupil relative to the sensor for EF lenses, so you can use the same EP distance in your calculations.  The 1.4x or 2x however does move the relative position and you need additional extension equal to the flange-flange length of the tele-converter.

 

I'm assuming you are talking about change of entrance pupil position as the 8-15 lens zooms,  it appears to be a minimal change for this lens.

This topic is one of the best things I've read. Congratulations!

 

Newbie question: can I assume "entry pupil" is where the aperture blades are on my lens just looking inside it?

 

For rectilinear photos (sharks, mantas, etc.) would you go for RF 16mm 2.8 or EF 8-15 f4L + defish?

The entrance pupil is where the aperture blades appear to be.  It can however be a bit tricky translating that to a spot on the lens barrel for calculations, but should be possible to get reasonably close.  There are various ways of determining its position including , looking it up on the Pano tools database,

https://wiki.panotools.org/Entrance_Pupil_Database#Entrance_Pupil_Measurements

testing for the nodal point as described on a number of websites or eye-balling it by looking into the lens.

 

As to what lens to use it is basically determined by reach - how close the animal allows you get to it and this will vary by location - shark feeding shots, they could be very close and some manta feeding stations or they might be quite shy.  For shy creatures something in the range of a 16-35 or 14-28 lens range is normally suggested or a WWL variant with a lens that allows some reach, 28-60 with a WWL gives equivalent to about a 14-30mm lens.

 

The fisheye is generally judged to be not have the reach for shier sharks in particular, even if you were to use a fisheye I wouldn't de-fish it, sharks, mantas etc will look fine shot with a fisheye.

1 hour ago, Chris Ross said:

The entrance pupil is where the aperture blades appear to be.  It can however be a bit tricky translating that to a spot on the lens barrel for calculations, but should be possible to get reasonably close.  There are various ways of determining its position including , looking it up on the Pano tools database,

https://wiki.panotools.org/Entrance_Pupil_Database#Entrance_Pupil_Measurements

testing for the nodal point as described on a number of websites or eye-balling it by looking into the lens.

 

As to what lens to use it is basically determined by reach - how close the animal allows you get to it and this will vary by location - shark feeding shots, they could be very close and some manta feeding stations or they might be quite shy.  For shy creatures something in the range of a 16-35 or 14-28 lens range is normally suggested or a WWL variant with a lens that allows some reach, 28-60 with a WWL gives equivalent to about a 14-30mm lens.

 

The fisheye is generally judged to be not have the reach for shier sharks in particular, even if you were to use a fisheye I wouldn't de-fish it, sharks, mantas etc will look fine shot with a fisheye.

 

Thank you so much! Photophotons entrance pupil database looks very good.

 

However I would like to shoot shipwrecks near home too, that's why I'm trying to avoid "fisheye" look.

42 minutes ago, stillviking said:

However I would like to shoot shipwrecks near home too, that's why I'm trying to avoid "fisheye" look.

 

Each to his own, of course. But unless you've tried FEs on a  wreck and really don't like the results, I wouldn't write them off.

 

Depending how you use it, of course, it's rare that an FE distorts a wreck to such an extent that it becomes obvious because it's rare that wreck images have so many obviously straight lines (weird though that may sound). There are loads of superb wreck pics where an FE was used and you'd be hard pushed to spot that.

 

They have the advantage too of being relatively inexpensive and can be used in small domes: great for transportation and manoeuvre in the water.

 

But, as I say, each to his own.

 

As an addition to the plea for the fisheye: for sharks you can use the 1.4x TC (there is so far only a single person here, who dares to use the 2x TC, most people fear that IQ is not good enough). With 1.4x the shark still has to come close, but IQ will be very good in case it does...

With more narrow angles of view the chances to get a shark into the entire frame are better, but no matter how brilliant the lens/domeport combination optically is, there is so much water inbetween the shark and the lens that IQ suffers in most cases (unless extremely good vis)...

 

Wolfgang

I guess this depends on the shark.

 

Wolf

Did you need to polish the dome after this?

I’m surprised there isn’t more discussion about the Sea and Sea corrector lens. I have one for the EF 17-40L and corner sharpness was noticeably improved. Not WACP wow but better IMO. They only make two sizes, maybe too big for RF lenses? 

There has been some discussion about the S&S correction lens. Especially about availability. As you say, there are, or were, two sizes and a number of members use them especially with 230 domes ad FX cameras. 

I wonder if anyone knows what is special about the S+S correction lens or is it like other low power achromatic diopters?

 

My limited understanding is that single lens diopters (filters) impact image quality whereas achromatic ones are much better. Both just act to reduce the minimum focus distance, helping the lens focus on the virtual image which is close to the sensor and curved.

 

So if the S+S is not available, or is not a suitable size for a particular lens, another weak achromatic filter would likely do the same thing?

 

There are several high quality low-power achromatic filters including the Canon 500d close-up lens, Nikon 5T diopter, and Sigma close-up lens AML 72-01 (which I have and is 72mm threads) which seem to do the same thing but are in different diameters due to sometimes being intended for specific lenses. 

Edited September 6, 20241 yr by John E
typing mistake

They seem to sharpen the corners a bit. It helped with my WACP envy. A tiny bit. 

On 9/6/2024 at 3:54 PM, John E said:

I wonder if anyone knows what is special about the S+S correction lens or is it like other low power achromatic diopters?

 

My limited understanding is that single lens diopters (filters) impact image quality whereas achromatic ones are much better. Both just act to reduce the minimum focus distance, helping the lens focus on the virtual image which is close to the sensor and curved.

 

So if the S+S is not available, or is not a suitable size for a particular lens, another weak achromatic filter would likely do the same thing?

 

There are several high quality low-power achromatic filters including the Canon 500d close-up lens, Nikon 5T diopter, and Sigma close-up lens AML 72-01 (which I have and is 72mm threads) which seem to do the same thing but are in different diameters due to sometimes being intended for specific lenses. 

They are not a diopter, rather a field flattening lens designed to pull the corners into focus, If I recall correctly they may change the field of view a little but are not really allowing closer focus like a diopter would.  Single element field flatteners are common in astronomical applications for example.

On 6/30/2024 at 5:30 AM, DreiFish said:

1. EF 8-15mm zoom needs something like a 35mm extension with the 140mm dome, not the 30mm extension recommended by Nauticam. Here's the first image at 30mm, and the second at 40mm. You can see 30mm puts the entry pupil in front of the center of curvature of the dome, while the 40mm extension puts it slightly behind, and vignettes as a result. 

 

Here are the same images de-fished to highlight the effect. Look at the above-water portion. If the entry pupil is exactly in the center of curvature of the dome, the lines should continue straight above water. If it's too far forward, the above water portion appears smaller. If too far back, it appears bigger.

 

 

 

 

40mm extension is almost correct, so I'd guess 37-38mm is what would be perfect. Of course, the problem is (as shown above) that it vignettes, and removing the dome shade wouldn't fix the issue. It vignettes on the inside of the dome. A 35mm extension thus might be the best compromise -- or 38mm with a wider port, like Marelux's 125mm diameter ports. 

 

Actually this was the MAJOR FIND and great recommendation for Canon Full Frame Shooters.

With Marelux there is 15mm and 20mm which can help you construct 35mm.

Also their 140mm Dome allows you to quick unlock and lock the sunshade underwater.

 

PS: I was very glad Dreifish took my initial test-setup idea for the Samyang 14mm and took so many lenses and port combinations out for a real impressive test drive. And many thanks to @Architeuthis for reminding me of this thread recently - which I totally had forgotten, even though I previously participated in it. 

On 6/30/2024 at 5:37 AM, DreiFish said:

Now, the RF 16mm prime. All shots at F13, ISO 100. 

 

Unfortunately, this lens is not great for underwater. Nauticam recommends no extension at all, and I tried that with the 140mm dome.  But because the lens is so short, the entry pupil is quite a bit further than where the lens sits even with no extension rings, and there's no way to improve it.

 

 

Center resolution is not bad. Corner (or, really, more like edge resolution) is not as good.

 

 

HA !! Good news for the MARELUX Canon Shooter Community:

The 140mm Marelux Dome works excellent with the Canon RF16 F2.8 lens.

Dent sharp at F11 - you can even keep the sunshade on - while it might blow off your socks 😉 

 

As I learned in another thread the reason for this might partially be that we have an approx 8mm (exact 7.9 mm) advantage over the Nauticam Canon housings concerning the flange distance, needed for this lens to get us to straight chessfield-lines (exact positioning of the entrance pupil for doing proper split shots). Here are my quick and dirty test shots, proofs for you.

There are some things (catches?) to be aware of, though:

 

Using such a very small dome gives you a heavily bended virtual image, with any manufacturer. You can see that the alignment of chessboard fields is the same above and below water in the center of the frame, while they start to magnify towards the edges. I suppose dome compression of the virtual image is the root for this. I can imagine that will disappear when using a larger diameter dome or the affordable INON dome with a port-adapter.

Second, as the entrance pupil moves quite a bit forward when this lens is focussing more close you will be able to "make it come out of it's housing cave" by getting really close to your subject. In other words the closer I moved the camera and dome towards the chessboard -> the more the entrance pupil came out -> the better the above and below water alignment (similar size of the squares on the chessboard) got. 

 

The I-P delta is 6.14 mm for the nerds out there doing the math. To give you a reference: the widely appraised I-P delta of the EF 8-15 Fisheye Lens is just 1.63 mm and the for underwater superb Sony SEL-20F18G 20mm F1.8 has a I-P delta of ZERO. Which means that lens does not move it's entrance pupil at all when re-foccussing. It is also a nice verification and proof that every millimeter of positioning counts, when working with full frame camera and small dome ports. Exited to learn, if Super-Moderator @Chris Ross will finally swallow this and agree ? 😉 

 

When you look at the minimum focussing distance chessboard images in large, with LENS CORRECTION ON and OFF processed in LR you will actually learn that this 16mm is a little bit fishy (leaning towards the barrel distortion of a fisheye) and Canon lens corrects the hell out of this optic to get it straight. This will also effectively give you a FOV and look of a 20mm prime lens in topside use. In topside use it's welcome underwater you might want to turn it off and enjoy the vignetting, as a special look of this lens. 

 

When you turn LENS CORRECTION of you will have small black corners, though. 

 

 

 

 

Edited February 7, 20251 yr by Adventurer

48 minutes ago, Adventurer said:

 

 

 

The I-P delta is 6.14 mm for the nerds out there doing the math. To give you a reference: the widely appraised I-P delta of the EF 8-15 Fisheye Lens is just 1.63 mm and the for underwater superb Sony SEL-20F18G 20mm F1.8 has a I-P delta of ZERO. Which means that lens does not move it's entrance pupil at all when re-foccussing. It is also a nice verification and proof that every millimeter of positioning counts, when working with full frame camera and small dome ports. Exited to learn, if Super-Moderator @Chris Ross will finally swallow this and agree ? 😉 

 

 

 

 

 

 

 

A rectilinear lens is always going to be more sensitive to this than a fisheye lens, the way they render lines is fundamentally different to a rectilinear.  It actually not possible to make a rectilinear lens with a 180 deg diagonal field of view, have a look at the formula for rectilinear lenses it falls apart when you try to solve it for 180 deg.  Basically the ends of a straight line need to be at infinity to be rendered rectilinearly.  If you look at the infocus field for a fisheye it's a curve not a line, almost like it was designed to be used behind a dome.

 

I have never doubted that a rectilinear needs to be close to the ideal position.   Fisheyes however are another story.  I don't have the maths to describe it correctly, but experience from lots of people posting about the Canon 8-15 show it's nowhere near as sensitive to dome position.  I don't doubt having the centre of curvature 40 or 50mm away from the entrance pupil will have some impact, but if you are within 5-10mm you will be in pretty decent shape.  Some of it seems be down to the way the image is compressed in the corners, the image scale is a lot smaller and kind of hides aberrations by shrinking them.

 

If you want to do these investigations, then more power to you of course, but if you don't fret about your image corners too much the Nauticam port charts seem to be pretty decent.

 

 

If you have nothing factual to contribute moderating could sometimes just mean reading, Chris.

The port chart of manufacturer X does not help users of brand Y or Z.

 

RomiK showed pretty well that EF 8-15mm is not invincible against raping it behind the wrong domes in wrong distances.

 

Dreifish showed in this thread that the Nauticam port chart is far away from decent and prudent control pays off.

 

Dreifish and I demonstrated multiple times with fisheyes and  rectilinear lenses that every milimeter matters and very often can easily improve your IQ at almost no additional costs.

 

If your housing brand does not offer extension rings in min. 5mm steps you should consider changing brands or buying ports and extension rings from third party port-makers.

I agree with Chris about fisheye lens positioning and port charts. Also, from what I can work out with my system, the Canon 8-15mm with the Sigma MC-11 adaptor and the Kenko 1.4x has a minimum focus distance of about 25mm from the front of the lens. With my 4.5 inch fisheye port and manufacturer recommended extension this puts the minimum focus about level with the glass. With the slightly longer extension that my "chessboard" test suggests, the minimum focus would be inside the port.

 

So the port chart recommended extension is better for CFWA. Since this is an important aspect of this lens I believe manufacturers take this into consideration.

 

Thanks @adventurer for your  response, I should mention that I spent some time comparing the two different extensions with the 8-15mm though it didn't help that the image scale was different in them and spent a lot of time flicking back and forth between the two images till I thought i was starting to hallucinate.  So I downloaded the two images and cropped an equivalent part of each and pasted in a single image to compare them, here it is 30mm is on top:

 

Now I may need to get my eyes checked but the difference doesn't really set the world on fire, if anything there might be a bit more micro contrast on the collar to the right of the 6 in the serial number, but really not much in it to my eyes.  To be fair there may be more to be seen in the original tifs, but based on this I think I could use either one happily.  It would be interesting to see a similar comparison for a rectilinear lens.