DreiFish

I'm curious about the way the Sony FE WA zoom lenses perform also. Especially from the perspective of getting a lens that works well with a 180mm dome -- Interceptor 121's tests suggested that both the 16-35F2.8 GM and Tamron 17-28 III do fine in that configuration. Probably helped by the 16/17mm wide end. On RF mount, I'm struggling to find a WA lens that performs great in a 140mm or 180mm dome. My hope was that the RF 15-30 might be that lens because of its shorter MFD.. but.. if anything, I guess my tests have shown me that they definitely perform better with a bigger dome. I guess the grass is always greener on the other side. I'm contemplating a side-grade to a Sony A1 system in part because of access to the 16-35F2.8 GM lens (also for access to the Nikonos 13 conversion, an AF-capable Laowa 10mm prime -- which works well enough with the 140mm dome -- and 1/400s flash sync speed). So stay tuned.. maybe I'll do these tests with the FE WA zooms in the future. I hope the results in that case won't be that the RF WA lenses were better after all 😄 (Joking and GAS aside, the differences that can be discerned in these controlled tests may not really matter much for real world images. It's always good to keep things in perspective -- probably any of these lenses with any of these dome combinations could produce great images with the right subject, composition, etc.. Doubt anyone seeing my photos on Instagram on their phone would ever notice the difference in corner sharpness if I got the other stuff right...) It's an interesting question. Three possible explanations (that I can think of): Minor differences in focus -- especially since I had to manually focus the lens. Possible, but I doubt that would change the microcontrast, which is the biggest difference I see in the center crop Differences in how close the lens entry pupil ends up compared to the center of curvature of the dome. Although the Nauticam 140mm port is fisheye port, it's not quite exactly half a hemisphere, whereas the Matty Smith dome is. I don't understand the physics well enough, but seem to recall that misalignment of the entry pupil would mostly result in image degradation in the corners, not in the center of the image. So it's probably not this. Probably the most likely reason is actually the amount of water I'm shooting through. I tried to keep the framing the same, which means roughly the same distance from the subject to the camera sensor with both domes. But since the test set-up is in a small tank, and the subject filling the frame was a 15" chessboard, that means the camera is quite close to the subject. With the 427mm dome (213.5mm radius), the front of the dome is almost touching the subject, which means you're shooting through very little water. Only a few centimeters of water as you can see in the shots of the test setup below. With the 140mm dome (69mm radius), it's an additional 144.5 mm of water you have to shoot through, with all it's associated loss of contrast and resolution. This might very well explain the difference. Which also means you wouldn't see such a clear difference in center resolution/microcontrast in the real world where you'd be using a 230mm (120mm radius) dome vs a 140mm dome (69mm radius) and where the subject is 1-2 meters away or more. Goes back to my earlier comment that while differences can be seen in extreme test conditions like this, for real photos of real subjects, other factors may completely obliviate the difference. No, but I would assume you end up with the exact same distance from the sensor to the front of the housing with the N120 housings as you would from the sensor to the front of the 35.5mm n100/n120 adapter on the Nauticam Sony housings. Phil's tests with the Marelux housing seem to confirm that, as the extensions needed on Marelux Sony housings for the Laowa 10mm line up with the 0-5mm extension I think is optimal on the Nauticam N120 housings when you account for the Marelux Sony housing being about 20mm deeper than the Nauticam Sony Housing or 15mm shallower than the Nauticam RF housings.

Ok, I've retested the Laowa 10mm with the Nauticam 140mm dome (dome shade removed), no extension on N120. This time I made sure the aperture was F11. The results are a bit improved, which suggests my earlier test must've been at a wider aperture by mistake. Here it is on the right compared with the Matty Smith Custom 427mm dome on the left. Observations: The 140mm dome needs a slightly longer extension to put the entry pupil in the center of curvature than the 427mm dome. This is because the latter is more hemispherical than the former. Still, no extension on N120 comes pretty close. A 5mm-10mm extension might be optimal, depending on the size of the dome. Center performance is a little better with the larger dome, especially the contrast. But it's not bad with the smaller dome. Corner performance is improved more with the larger dome, but it's usable in my view even with the 140mm dome. With the shade removed, the 140mm dome doesn't vignette at all. There is slight vignetting at the top and bottom edges with the shade on. The Laowa 10mm is much more usable with the smaller 140mm dome than the Canon RF 10-20L, which probably needs a 230mm dome or larger to perform well. Here is a comparison of the Laowa 10mm vs the RF 10-20L with the 140mm dome. Laowa on the left, Canon 10-20L on the right. The corners especially suffer, probably because the Canon has a MFD of 250mm compared to 120mm on the Laowa. If you're curious, the two lenses perform much more similar to each other behind a 427mm dome. Head to head tests in this thread.

I call second dibs! 😄

I think you're being a little bit of a fanboy here and creating imaginary strawman strobes. Which large strobe is 600g heavier? With the exception of the OneUW 160s (which is heavy because it's rated to withstand 200m of pressure), every other large strobe (Seacam 160d, Ikelite DS230, Marelux Apollo 3s, Backscatter HF-1, Supe D-Pro) is only 100-300 grams heavier than the Retras with this battery pack, and they're all more compact if you look at the diameter and length. Most of the strobes I've listed use simple lithium cells and don't require any proprietary charger. But I suppose you're ignoring that USB-C charging means you also need.. a USB-C charger of some sort? Those (and the cables) also weigh something and aren't much smaller than for example the charger that comes with the DS230 or Supe D-Pro. Not to mention the USB-C charger and cables are extra $$$ on what is already a $2100 stobe once you include the Vault. The biggest complaint though is that the battery pack doesn't do much to improve the recycle time of these strobes, which is left still trailing the pack..

The MF-2 is 200mm in length. But I don't think they're the same length -- they look the same in that shot, but keep in mind perspective here. The foreground object is smaller than the background object. And there's the knob on the end So.. I still stand by my 220mm estimate. As to whether the weight or bulk is similar to the larger strobes -- well, Retra + Vault is 1020 grams, Seacam 160d is 1320g. Is that a world of difference in your book? If it's 7.2v, then I guess it's probably 2 21700 cells in there.

Hey Davide - standard JPEGs. Not sure why they're not linking. 10mm Laowa F2.8 prime, no extension, 17 10mm Laowa F2.8 prime, no extension, 17 10mm Laowa F2.8 prime, no extension, 17

Good question. Since the RF-mount version is manual only, I can't confirm from the EXIF data for the test with the 140mm dome, so I'm going to re-do it. For comparison, with the Matty Smith 427mm custom dome, I used F11, and here are the results: 10mm Laowa F2.8 prime, no extension, 17 10mm Laowa F2.8 prime, no extension, 17 10mm Laowa F2.8 prime, no extension, 17

Looks like the issue here is that the internal diameter of Retra's battery compartment, having been designed for 4 AA batteries, is too small to fit the LI cells they opted to use. So probably the battery pack has an internal portion that is used only as an extension to connect, and the LI cells sit entirely outside the main strobe body in this new extension. Which, if using 18650 batteries, would mean that the battery pack has to protrude at minimum 70mm past the back of the strobe, assuming some 5mm for the aluminum housing and USB-C charging circuitry. The 40wh rating suggests 3x18650 cells or 2x21700 cells if rated conservatively. That would mean the strobes go from being 148mm long with no booster to ~220mm long with the Vault. That's longer than any other strobe out there (at least the ones I've researched and added to the spreadsheet), including the Seacam 160Ds (190mm), and the OneUW 160s (200mm). Can we really call them a medium strobe anymore with those dimensions? Could anyone confirm if my back-of-the envelope calculations are correct? Retra AFAIK hasn't published the new length of the strobe or the details of of the voltage/AHs on the Vault.

Might work, but it's probably too long and may vignette. No extension is almost perfect on the N120 system, I don't think 10mm extension would improve the image quality. Likely, it would degrade it vs. no extension.

What if you were to take the EF 8-15 L fisheye and de-fish it? How would that compare with a rectilinear lens? First, in terms of field of view, The 15mm fisheye de-fished looks similar to a 12mm rectilinear lens. Like with the Nauticam wet optics, the de-fished fisheye has slightly less horizontal field of view when the vertical field of view is matched. But what about resolution and microcontrast? I'd say the de-fished fish-eye 15mm lens mostly holds its own against the 10-20mm at 12mm. The rectilinear lens has slightly better resolution and microcontrast in the corner, but it's not night and day. You totally can use the fish-eye de-fished in a pinch to produce respectable rectilinear images equivalent to a 12mm lens -- if you don't need the even wider field of view of a 10mm rectilinear.

Finally, are you wondering how the RF 24-50 + Nauticam WWL-C combination performs? Sorry to say.. but, pretty abysmally. Here it is at 24mm compared to the Loawa 10mm prime: The field of view may be similar horizontally, but it's narrower on the wet optic vertically. And it shows mild fisheye distortion, of course. And the image quality? Not even in the same league. The rectilinear lens has much better resolution and especially microcontrast. Both in the center and in the corners. About the only thing the WWL-C has going for it is the smaller size. But I'll compare it soon with the Laowa and RF10-20 behind a 140mm dome. I have a feeling the rectilinear lens will still do better.

Comparing the zoom options, here is the RF 10-20mm L at 10mm (20mm extension) vs. the RF 14-35L at 14mm (40mm extension). There's not much between them in the center. Even in the corner, it's hard to tell them apart. The RF 14-35 might have a touch more microcontrast, but it's hard to say. Impressive performance for the RF 10-20mm! If we normalize and take the RF 10-20mm to 14mm, then the comparison skews more toward the RF 10-20's favor, especially in the corners. Resolution is similar, but the RF 10-20 shows more microcontrast. What about the RF 15-30? Here it is at 15mm, with a 40mm extension, compared to the RF 14-35L lens, with a 40mm extension, at 14mm. Surprisingly, the non-L RF 15-30 looks a little better than the L lens, but there's not a whole lot between them. How about if we zoom in the 14-35 to 15mm? Surprisingly, it doesn't help. The RF 15-30 still has slightly better resolution and microcontrast, both in the center and at the edges. The RF 14-35 is taking a beating. How does the RF 15-30 compare to the RF 10-20L if both are zoomed to 15mm? Pretty similar, but I'd give the slight edge to the RF 10-20L. In conclusion, the RF 10-20L is the best option for a wide angle rectilinear zoom on the Canon RF system. It performs better at 14mm than the RF 14-35L, and better at 15mm than the RF 14-35L or RF 15-30. Not to mention it can go as wide as 10mm, where it almost matches the Laowa 10mm prime (which is a manual-only lens on RF mount anyway, so not really an option). All while needing the shortest extension of the lot at 20mm. Of course, it will probably only work well with the 230mm dome because of its MFD of 250mm, but I intend to test it with the 140mm dome as well to see if it's still acceptable. The RF 15-30 also performs very well, besting the RF 14-35L at 15mm. If you don't need 14mm or 10mm, it's a very cost effective option (I got mine refurbished for $359.99 directly from Canon.) The RF 14-35L is the most disappointing of the lot. Unless you really need the 35mm reach at the long end, it's hard to recommend it over the RF 15-30 or RF 10-20mm options.

Now, the Canon RF 10-20L lens vs the Laowa 10mm prime. The 10-20 uses a 20mm extension, and no extension for the Laowa. RF 10-20L is at F13 like the other shots in the series, but the Laowa is set to F11, since this is a manual lens and only options were F11 or F16. The Laowa could use a 5mm extension (not possible on N120 ports) and the Canon RF 10-20mm L could use probably a 25mm extension, which I don't yet have to test. Nauticam recommends 20mm for the RF 10-20L. Now for image quality: There isn't much resolution or contrast difference in the center, though the color rendering is quite different. In the corner, the Laowa is a bit sharper. I will redo the test with these two using the Nauticam 140mm dome to see how much improvement you get going with a larger dome. The RF 10-20mm has a MFD of 250mm, while the Laowa's MFD is 120mm. This should show more of a difference behind the small dome.

First, let's compare the EF 8-15mm fisheye, bare at 15mm with a 30mm extension and then with the 2x Kenko teleconverter at 16mm with a 70mm extension port. First -- the 2x TC definitely leads to significant resolution loss (not surprising), but also light transmission loss. Shots are all at F13, ISO 100. It's probably still good enough though with the 2x TC, especially given the versatility you gain. I plan to 3d print a zoom gear for this combination and test it on the reef. It'd be interesting to compare the image quality vs Nauticam's FCP solution. The IQ difference going from the 140mm dome to the custom 427mm Matty Smith dome is surprisingly minimal though. Almost no difference in the center, but a bit more noticeable in the corner.

Ok, did more tests today. I now have the Canon RF 10-20mm L lens as well, which I thought would be interesting to test against the Laowa 10mm. Second, I switched to the 17" (427mm) Matty Smith custom "split-shot" dome to see how much of an image quality difference the larger dome made against the 140mm and 180mm options. The 17" dome is a true half hemisphere (fisheye dome). Matty Smith has confirmed that the radius of the 12" (300mm) dome is 6" (150mm) and the radius of the 16.8" (427mm) dome is 8.4" (213.5mm). Finally, I threw in the Nauticam WWL-C with the Canon RF 24-50mm F4.5-5.6 lens to see how Nauticam's wet contact optics compare. We can assume this one is the worst of the lot, especially given the mediocre lens behind it, but it should at least give a reference point. I would expect the WACP-C, WACP-1 and WACP-2 to perform better, especially with better lenses behind them, but don't have either to test. The test set-up looks like this. You can see the virtual image inside the dome.. surprisingly still significantly curved, even with a 427mm dome!

Correct. I've tested the Laowla with the 140mm nauticam dome and the Matty Smith 17" custom dome for nauticam n120. This is on RF system. In both cases, no extension is needed. Actually, my estimate is that around a 2-3mm extension would be optimal for putting the center of curvature of the lens in the middle of the dome, but since that doesn't exist, and 10mm doesn't work, and 20mm is too much... you're stuck with no extension by default. Here are shots with the 140mm dome and no extension.

Is there a link to the full specs? The recycling time improvement is a bit disappointing. Has anyone measured what the new Retras are capable of in terms of high speed recycling for action shots? What power do you need to be at to get 3, 5, 6, 10fps?

On the lists of things to test when I have time again, probably in 2 weeks when I'm back from Mexico. Might. be a bit tricky to measure. I believe Interceptor 121 did this for the Nauticam Sony housings at some point? If so, it would be easy to estimate adding 35mm (for the n100-n120 port adapter) to that. How else would I measure that? I have a caliper.. I guess I could do lens mount to front of housing with that or a ruler and add the manufacturer's flange distance. Do housing manufacturers share this information anywhere? It's not on OpticalBench yet. This is what I've pulled so far from OpticalBench. And from housing manufacturers documentation. All tests were done at F13 for consistency and expediency (otherwise I had to adjust the strobe power to take a series at different apertures. Took too long 🙂. I may do some expanded testing later at different apertures. For instance, the RF14-35 did clean up a bit in the corners at F16, but lost center resolution. I have the WWL-C and RF 24-50 I will eventually test as well. Yes, corrected.

Finally, the RF 15-30. 15mm, F13, ISO 100. With the 180mm dome, it needs a similar extension to the RF14-35, just slightly shorter. Here it is at 30mm, 40mm, and 50mm. At 50mm it already vignettes, but the entrance pupil has already been moved forward past the center of curvature. 40mm puts the entrance pupil behind center of curvature. At 40mm it is a bit closer to the correct placement than the RF14-35. This lens probably also needs a 230mm dome, but 45mm should be about right with the 180mm dome and shouldn't vignette much. (Incidently, Nauticam recommends 30mm, which is clearly wrong. 40mm is better) Center and corner crops at 40mm. Center resolution is not as good as the RF14-35 or Laowa 10mm, probably because the lens itself is not as good, but on par with the RF16.. Corner is better than the Laowa 10mm, RF16 and RF 14-35 though.

Now the RF 14-35 F4 L zoom, at 14mm. All pictures again at F13, ISO 100. The lens sits too far back from the center of curvature with the 180mm dome. Here it is with a 30mm, 40mm (what Nauticam recommends for 180mm dome), and 50mm extension. Even 50mm is slightly too short, but already at this point it's vignetting heavily. I think this lens needs to be used with a 230mm dome, the 180mm doesn't have enough field of view. Best compromise if you're set on using it with the 180mm dome would be a 45mm extension I guess, but you might get vignetting even then. 100% crops from the center and corner with 40mm extension. The center is a bit better than the RF16 prime but not as good as the Laowa. The corner is much better than the Laowa but kinda on par with the RF16 prime (which is behind a 140mm dome)

Contrast with the Laowa 10. With no extension, it vignets slightly, but it's only the dome shade (which was designed with the more limited vertical field of view of a fisheye lens in mind). If you remove the shade, you should be able to get the full field of view. I tried also with a 20mm extension, but that produced very significant vignetting inside the port. Notice the lines are almost straight above the water. Just sligtly smaller. Basically, you would get perfect placement with a 2mm extension. Maybe a 10mm extension could work, but then the entry pupil would be too far back of the center of curvature of the dome, so image quality would likely not improve. Center resolution is better than the RF 16mm prime or the EF 8-15, but the corners.. well, they're pretty bad.

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.

I wanted to test to see the correct port extension for various wide angle lenses with the 140mm fisheye dome and 180mm wide angle dome for the Nauticam N120 full frame system. I also wanted to compare resolution of the different lenses and how the position of the lens and the size of the dome impacted the resolution. Lenses tested: Canon EF 8-15mm Fisheye Zoom Canon EF 8-15mm Fisheye Zoom + 2x TC Canon RF 14-35F4 L Zoom Canon RF 15-30 Zoom Canon RF 16mm Prime Laowa 10mm Prime All tests done with the Canon R5C. Here's the test rig setup -- a chessboard to see if the straight lines continue straight above the water (to determine if port extension is too long or too short), with some paper bills glued to it to eyeball resolution. I haven't gone through all the photos yet, but already some interesting findings. 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. The incorrect placement of the entry pupil doesn't have much of an effect on the center sharpness (30mm on left, 40mm on right). Microcontrast might be a bit better though with the 40mm extension. But there is a noticable improvement with the longer port when it comes to the corners. More to come.

Can’t fit the R5 body in an R6Ii housing unfortunately. Not even close.

Thanks Chris!