DreiFish

I've measured strobe color temperature always on land, because water introduces an unknown filtration variable. I've used two methods -- a Sekonic spectrometer (which is what Backscatter also uses) and taking shots of an 18% grey card illuminated with the strobe using a macro lens at F32 ISO 100 1/250th to minimize impact of ambient light. With the grey card method, I then use the WB tool in Lightroom on the grey card (shoot in RAW) to see what color temperature Lightroom reads to render neutral grey. Both these methods produce quite similar results, within a couple of hundred degrees kelvin. For example, the Marelux Apollo 3s without diffusers measure as 6500k with grey card method or 6800k with Sekonic spectrometer. With the diffusers, grey card gives 5800k and Sekonic 6000k. So I'd say you get to within 200k margin of error with either method. My results also tend to closely match Backscatter's, which also uses a Sekonic spectrometer. . For example, I've measured the Ikelite DS230s as 5200k with grey card method. Backscatter got 5400k with spectrometer. This strobe, by the way, comes closest to the advertised specs of 5000k. Every other strobe I've tested is somewhere between 500k-1000k cooler than advertised. Why is that? All I can conclude is.. marketing. Manufacturers know that warming up a strobe means inevitably decreasing the power, and they'd rather have higher GNs. Power is fairly easy for a user to 'see' and complain about. Color temperature.. not so much unless you're a nerd and into testing these things. As Massimo says, zenon flash tubes have a native color temperature between 5000k-6000k. So to get a warmer strobe, what manufacturers like Seacam/OneUW and perhaps Retra and Ikelite do is apply a orange-yellow coating directly to the flash tube. If you have one of these strobes, check the flash tube -- does it seem a bit yellow-orange? Others like Inon and Backscatter offer optional snap-on filters that accomplish the same. I wonder if the coating method actually wears down over time as the strobes fire and age. This would lead to color temperature becoming cooler over time with no way to fix it? That'd be very bad. Perhaps the click-on filters are a better method.

Backscatter tests it using a Sekonic spectrometer. I've tested it both with a Sekonic and by shooting an 18% grey card in macro conditions ON LAND (F32, ISO100, 1/250 to limit ambient light) and then checked with the WB tool in lightroom. The results of these methods match closely (within 100-200 degrees Kelvin). Testing color temperature should be done on land as backscatter does it because any amount of water will introduce filtration, and the specific amount of filtration will depend on the number of particulates in the water and the distance. Air has much less of a filtration effect at close distances. Incidentily, my results also generally match the Backscatter results. I measured the Ikelite DS230 at 5200k (via 18% grey card method) and Backscatter got 5400k. (Advertised specs are 5000k). Of all the strobes I've tested myself or that Backscatter has tested, the Ikelite DS230s were the closest to the manufacturer's specs. Everything else is anywhere from 500-1000k cooler. Why is it? Probably.. marketing. As Massimo explained, Zenon flash tubes produce somewhere between 5000-6000k natively. To get it warmer, you have to apply a warming coating to the flash tube, which is what Seacam/OneUW and probably Retra / Ikelite do. This of course absorbs some part of the spectrum of light, leading to lower overall light output. You can check this out by having a close look at the flash tubes on your strobe -- do they have a orange-yellow tint or coating? What I'm curious about is whether this method of coating the flash tube degrades over time and leads to strobes becoming cooler as they get older. That would be a big downside vs alternative methods of warming up the strobe temperature like click-on filters.

Since I've gone thourgh all the work of gathering this data from various sources and personal tests over the past few months, I might as well share it here in case other find it helpful. Corrections and additions welcome. Update on 24/06: I've updated the spreadsheet with three new sheets, one showing my detailed Power, Color Temperature and Flash Duration tests,other showing the recycling times / high speed shooting tests, and a final one with images comparing the beam cover with and without diffusers for the strobes I've tested. https://docs.google.com/spreadsheets/d/1hWG6Enh-ukiQXjN5y_3NMTFF8b4dhJ64YfZqyIdpFew/edit?usp=sharing

DivePro G18 lights -- you can probably get them for around $600 if buying in 'bulk' at least.

Not following -- why wouldn't the battery capacity in Jules divided by the number of shots at full power give you a fair approximation at least of the power used for each shot? Typo.. or dyslexia 🙂 Eh.. it's not about auto white balance only. I have my in-camera white balance set at 5000k. But I do a global white balance adjustment in post to get the color temperature correct for the foreground subject. In the process, you get a different color for the background water column. Well, I promised in-water tests, and here they are. Subject is the side of a wreck, so fairly flat, distance is approximately 1m. All images white balanced to 5000k, so you can also see the color temperature. First, Marelux Apollo 3s at full power, without diffuser (first image) and with diffuser (second image). I judged good exposure to be ISO 100 F16. Second, the Inon S-220s at full power, without diffuser (first image) and with the 4600k filter (second image) Best exposure was at ISO 100 F11. And now the Apollo 3 first and the S220 2nd. Basically, the Apollo 3 is one stop brighter, but the biggest difference is coverage. These are both shot with a fisheye. I think the real weakness of the S-220 is not power, it's coverage.

By pushed back you mean use a shorter extension so the lens protrudes more into the dome? With 30mm extension, the 14-35 did not vignette with the 180mm dome. I just don't really love the look.

The real comparison I'd like to see is the Seacam Seaflash 60d vs the Inon s-220. Seacam rates their strobe at 60 jules, GN 8 (underwater). Inon rates it at GN22, but this is probably in air. They don't publish jules output, but calculating backwards from rated number of shots and battery capacity, -- both use 4xAAs -- , the Seacam has 200 shots at full power, and the Inon 500. Something here doesn't add up. If the Seacam is using 60 jules for a full flash, it should be closer to 600 shots. Unless something about the design is really inefficient. Working backwards from the energy density of 4x Enerloop Pro AAs, the Inon s-220s are probably maxing out around 75 jules at full power.

I know comparing across brands is pointless as they all test the GN under different conditions (in water or not). Beam spread and hotspot also impact the GN number if all you take is a central head-on reading. I was comparing within the same family.. and there, as your findings also corroborate, if the Z330 is rated as GN32 by Inon, and the Z240 was 24 and the D200 was 20, GN 22 for the S-220 seems right in comparison. I'm testing the S-220s side by side with the Marilux Apollo 3s and will have some side-by-side pictures to share shortly. I found a perfect 'test environment' with an underwater subject at 1m distance at around 12 m depth to try out different F stops at to see coverage, illumination and color.

I have tried the Canon RF 14-35 with Nauticam 180mm dome and 30mm extension* (*Nauticam recommends 40mm, but I don't have one, only 30 or 50). The results were.. not really great at 14mm in my view, even at F16. Nauticam recommends the 230mm dome for this lens as the best optical option, so I'm not sure how much of an improvement the 40mm extension would make. For me, 180mm dome already starts to be too big and floaty for good ergonomics. I can't imagine what the 230mm would be like. Certainly a pain to travel with, push through the water and hold horizontal without extra trim weights I would think . The other issue is that by 14mm you already have extreme perspective distortion stretching out everything near the edges of the frame a lot. I find that fairly difficult to work with. So 16 or 17mm might be better. For me, the holy grail would be a rectilinear zoom lens that works best with an 180mm dome. Haven't found that yet for the RF system -- but perhaps the 16mm F2.8 or the RF 15-30? Both have very short MFD, so, probably the things to try. Does nauticam recommend the 180mm dome as the best option for any of the 14mm lenses listed? Or has anyone actually confirmed that one of them works well with a 180mm dome? I think there's some 16 and 17mm lenses that do work fine with the 180mm dome, so those might be the best rectilinear options.

Yes, correct. The concept is not exactly new -- it's been used with video lights for a long time. Keldan is the biggest proponent, but on the old Wetpixel forum, there's a long thread on "ambient light filters" using Rosco gels to match the torch color temperature to that of the ambient light at different depths (works best down to around 12m). I did some experiments 2 years ago and came up with a pretty decent formula using cheap gels to get to similar end result as you get from the $100 Keldan ambient light filters. No reason same couldn't be used with strobes instead of video lights.

Thanks for the head's up. I really would like to try the 60Ds before committing that amount of money. Unfortunately, I don't know anyone using them or that would have a demo or rental set on hand. Only review I've been able to find was from Adam Handlen, and that one is fairly subjective. Based on everything I've read, I'd assume they're comparable in power output to the Inon S-220s. Perhaps even weaker. That wouldn't be much of an improvement.

Agree with this. Also, I think this thread (despite veering towards personal fights) did highlight that what's needed for close, colorful reef shots is not the same as what's needed for monochrome subjects at further distances. Scenario 1 needs less power but better color temperature. Scenario 2 needs maximum power and color temperature isn't very critical since.. grey is grey.

Why do you say the S-220 is rated conservatively? Sea & Sea rates it as GN22, vs GN20 for the S-2000 and GN 16 for the Backscatter MF-2. This seems about right. In practice, I can use it (with the 4600k diffuser and a 1/4 CTO gel, which rob about 2/3 of a stop of light) fine for wide angle on the full power setting. Here are a couple of shots, at 1/125, F8 ISO100 and 1/125, F13, ISO200. Based on these results, I think it's pretty ok for most general wide angle usage, even in bright water. I'd happily shoot it at F13, ISO 400 and crank up shutter speed to 1/250th. This would be true if what we cared about was illuminating the subject, not restoring color. But in most cases, we don't need to illuminate the subject -- the sun does a plenty good enough job of that. What we need to do is restore color, i.e. get more red and yellow light on the subject. Cooler wavelengths of light might help the light penetrate further through the water, but they don't help with restoring color. This seems pretty accurate. In my experience, 4800k produces aesthetically pleasing results, especially with skin tones on subjects. (The Inon S-220 + 4600k diffuser + 1/4 CTO gets me to 4800k, which is what the above images were globally whitebalanced to). So I think I could live just fine with a strobe in the 4800k-5200k range). I think in practice F11-13, ISO 400 1/250th is sufficient for even shallow shots in tropical seas. And the S-220s are able to manage that (just barely) with decent color temperature. So.. unless shooting heavily backlit scenes or sunballs, they're 'good enough' for wide angle, even with full frame cameras. Those limitations come up infrequently enough for me not to matter much. Even if you need to shoot a rectilinear lens at F16, you can simply up the ISO to 800 to compensate keeping all else equal. Agree.. the S-220s seem to be.. 90% of the way there, and the size, weight and price are hard to beat. It would be nice to have another stop of light to play with though. I shoot video too. It all comes down to what you're trying to achieve. Blue filters on your lights are great if you want a consistent foreground to background color temperature, but (a) they don't match the color contrast you can get to separate the subject from the background with warm strobes and (b) they're not exactly ideal when shooting models where you want to get natural skintones (not 50 shades of cadaver grey...) Useful in some circumstances. I'm happy to see Backscatter offer the option, and wish other strobe manufacturers would do as well. Of course, you can achieve same result with the right combination of cheap gels. I think your options are basically Sea & Sea YS-D3 or the new Backscatter HF-1. The HF-1 however is already 1.1kg, so.. not light. The YS-D3 is a much more appealing 730g, but was measured by Backscatter as 6600k. So you probably need at leas 1/4 CTO gel to make it reasonable color temperature. Which makes it ~GN24, not 32. I think that conclusion may be influenced by the fact that the Z240s weren't truly GN24 as advertised? To me, they seemed almost a stop weaker than the YS-D2s at the time when I tried them. A true GN24 strobe should be more then enough, if it's GN24 underwater at one meter -- when do you ever shoot wide angle at F22 ISO100? The only example I can think of is sunballs. If a strobe produces GN24 underwater with a color temperature between 4300k-5200k, I'd be quite happy. Agree that GN32 (true, in water) would be plenty enough. Seeing as the Inon S-220s doesn't get you there (but are probably close in power to the old Z240s), the only options I see is to step up to the YS-D3 (730g) or Retra Pro Max (878g w/ 4 batteries). I wonder how the Seacam 60D compares here. It's only 550g, so marginally heavier than the Inon S-220s. Seacam rates it as GN8 in water (which probably is accurate/conservative). That would put it more or less in the same ballpark as my test results for the S-220 in water. Which is to say.. good enough for 90% of use cases. GN8 gets you F8/ISO 100 or F11/ISO 200 or F16/ISO 400. It's good enough if not shooting sunballs.

Main problem is weight. Both carrying weight and the weight of the strobes on long arms. My R5C rig (with 8-15 and a 140mm dome, about as light as it gets) with 2x Marelux Apollo 3 (1.3kg each) strobes is 11.5kg. This is too heavy.. I hate carrying it any distance, and even handing it up to the boat is a challenge. It's hard for any arm system to handle that much weight. Certainly on land, but even in water. When I was using the WACP-1 before and arms with clamps (rather than locline), it was closer to 15kg. That's.. cumbersome. By comparison, handling my wife's rig (R6 Mark II) with 2 S220 strobes is lovely. It weighs 6.5kg. So.. reducing weight (meaningfully, we're talking almost 2kg here from the strobes) is a big plus. Not just for travel, but also for handling above water. Exactly this.

For objective testing (in-water) of the effect of dome diffusers: https://www.backscatter.com/reviews/post/Sea-&-Sea-YS-D3-Duo-Underwater-Strobe-Review https://www.backscatter.com/reviews/post/Retra-Flash-Pro-Max-Underwater-Strobe-Review https://www.backscatter.com/reviews/post/Ikelite-DS230-DS232-Underwater-Strobe-Review https://www.backscatter.com/reviews/post/How-To-Underwater-Strobe-Flash-Placement

The reason for using warmer strobes is that you can adjust the white balance (for the whole image) to the strobe illuminated subject (e.g. to 4800k), so the subject looks natural. This overall white balance then has the effect of rendering the background, especially water column a more rich and pleasing shade of blue. So it's not exactly correct that you can just fix it all with WB. If you have a cooler strobe, you must wb the whole image (or, selectively the subject at least) to a cooler temperature like 6000k to look natural. That leaves the blues as less pleasing and muddy. Of course, you can achieve everything that warm strobes achieve in post-processing by indepently adjusting the color temperature for the strobe-lit foreground subject and ambient-light-lit background, but this requires more work and time in post-processing. Many prefer to just use warmer strobes or filters and achieve the same outcome without post-processing. There's an old Alex Mustard post that explains this in detail.

Hey Chris, You might be mixing up two different things. If you think of the cone of light, wider strobes have a wider cone of light, but the principles are the same -- moving the strobes back (without moving them out) doesn't actually reduce backscatter -- if anything, it increases the area between the dome and the subject that is illuminated. Only moving the strobes out to the sides will decrease the backscatter. What moving the strobes back does (especially with a wide lens like a fisheye) is ensuring you don't get light hitting the dome itself (which produces reflections and artifacts way worse than backscatter). Anyway, it's a fine distinction, but from a physics perspective there's no way moving the strobes back can actually decrease the amount of backscatter in terms of lighting up particles in the water column between the dome and the subject. What it does do is mean that less light hits those particles because of the distance between them and the strobes (inverse square law). But.. that also means that less light hits your subject for the same reason, so you end up having to increase light power to compensate, which gets you back in the same place. Basically, move the strobes back just far enough to avoid direct light hitting the sides of your dome and then move them outwards proportionally to the subject distance (or inwards for close subjects) to control backscatter.

I haven't done that with the light meter (since travelling currently), but I have test shots I can examine in Lightroom. 1/4 CTO filter reduced light output by 1/2 stop (6200k -> 4800k on Marilux Apollo 3s w/ diffuser) 1/2 CTO filter reduced light output by 2/3 stop (6200k -> 4000k on Marilux Apollo 3s w/ diffuser) The Inon 4600k diffuser on the S-220s is.. pretty ineffective. It reduced light output by about .2 stops, but also only took the strobe from 6100k to 5800k. You definitely lose light output... more or less proportionally to what the strobes gained in the first place by having a cooler color temperature. So I'd say, power-output wise, it's basically a wash. You can get more powerful cooler strobes, or warmer less powerful strobes. But if you use a filter on the more powerful strobes to improve the color temperature, you end up with similar light intensity as the warmer strobes. (In my testing, the Retra Pro Max was only about 1/2 stop less bright than the OneUW 160/Marilux Apollo 3) I do think recycling times are a differentiating factor for certain use cases (fast action, pelagics), but less so for general reef wide angle uses. That's the main reason I got the Marilux Apollo 3s -- for sharks and fast action bursts. But that seems to be primarily a factor of power output vs. how quickly the capacitor can be charged based on the current the batteries put out? Is there a reason the newer strobes with LI-ION batteries like the Marilux Apollo 3 and the Backscatter Hybrid Flash recycle so much faster than older NiMH strobes? The power output, if anything, appears to be higher (~175 jules for the Apollo 3, 350!! jules for the Backscatter HF-1).

I'm trying to downscale (weight and size) the camera rigs for me and my wife while maintaining full-frame picture quality. We both enjoy wide angle photography. Are smaller strobes [e.g. Inon S-220 (463g) or Seacam 60d (550g)] good enough for that, or do you need something medium-sized [Retras (878g0, Backscatter MF-2s (1130g)] or large size [Marilux Apollo 3s (1375g), OneUW 160 (1550g) Seacam 160d (1320g)]? For reference, I've recently tested a whole bunch of strobes, and own/have owned Inon S-220s, Backscatter MF-2s, Retra Pro Max, OneUW 160, Marilux Apollo 3, Ikelite DS230 and Supe D-Pro. I have 2 Backscatter HF-1s on preorder. I'm looking for that goldilocks intersection between size, power, and color temperature. And maybe also price. But more on that later. For reference, we're currently diving in the Red Sea and here are the rigs we're using: My rig: Nauticam Canon R5C + 8-15 Fisheye in 140mm dome + 2x Marilux Apollo 3s w/ 1/2 CTO filters My wife's rig: Nauticam Canon R6 II + 24-50 + WWL-C + 2x Inon S-220 w/ 4600k filters + 1/4 CTO filters (I've bought some 1/4, 1/2 and full CTO filters to experiment with to adjust the color temperature just right) Red sea environment is great for testing as it's in some ways the most challenging for strobes. Very strong ambient light, especially in the shallows at midday. Goal -- get the smallest lights possible that still achieve good coverage and color (rich blues) in this environment. Thought I'd share my findings so far and invite further discussion: 1. Strobe Power Isn't that different between Small and Large Strobes On land I've measured the Backscatter MF-2 and the Marilux Apollo 3 based on lux output at 30cm, center, using a light meter. Backscatter MF-2 (advertised as guide number 16) : 381 lux Marilux Apollo 3 (advertised as guide number 44): 1179 lux OneUW 160 (advertised 😞 1060 lux Surprisingly, the difference wasn't so huge. We're talking 1.5 stops. I would expect the S-220s to test even a little bit brighter than the Backscatter MF-2, as the advertised GN is 22. Hypothesis: The power difference between the biggest strobes and the smallest ones is 1-2 stops, perhaps 3 stops at most. Even comparable figures from the same manufacturer seem to back up this hypothesis. For instance, Seacam rates the 160D as GN 16 in water, and the 60Ds as GN 8. That's only a 2 stop difference. If my hypothesis is true, it's hard to think of many situations where you really need that extra 2 stops of light output (all other factors like coverage, quality of light and color temperature being equal). I plan to put that to the test over the next week and see where the S-220s fall short for wide angle in the red sea. But so far.. not bad. Some preliminary results from the S220s (first image is at ~16m, with only the 4600k filters, second at ~10m depth, midday, but with 4600k filter + 1/4 CTO gel) Of course, the Marilux Apollo 3s may produce richer, warmer colors. I'll try to test them side by side on the same subject. 2. CTO Filters are a Cheap and Customizable Way to Adjust Color Temperature My testing so far shows that most manufacturers consistently advertise their strobes as being warmer than they really are. The S-220 for example measure as 6650k without diffusers and 5400k with the Inon '4600k' filter. The Apollo 3s measure as 6500k without diffusers and 5800k with diffusers. (The only 2 strobes I've tested that are close to advertised specs are the Ikelite DS230 -- 5200k measured, 5000k advertised, and the Retra Pro Max -- 5000k measured, 4800k advertised). The good news is that it's easy to adjust with some CTO gel you can order for $8 from B&H Photo Video. For instance, the Apollo 3s + Diffuser measure 5800k. Add a 1/4 CTO filter and that's down to 4800k. With a 1/2 CTO filter, you get 4000k. This may be a better option than buying expensive strobes or relying on manufacturer's warming filters. 3. Coverage Angle and 'Quality' of Light This is probably the hardest factor to measure objectively and test, without more specialized testing environment like Backscatter uses. What are your thoughts and experiences? Are there small strobes that are really good enough, if recycling times and battery capacity is a non-issue? The Seacam 60Ds seem like an interesting option, given their advertised 130 degree coverage, circular flash tubes and 4400k color temperature. Anyone have any experience with those? They're light, but mighty expensive. But if they're really only 2 stops weaker than the 160Ds, with all the same features...

Following. I already have the 180mm and 140mm nauticam domes. The 15-30 is a relatively cheap buy if it turns out to work well with those smaller domes. Not really that interested in rectilinear options that require a 230mm dome -- the ergonomics of that (makes the camera floaty on the front end, acts like a sail) are not very desirable in practice.

Don't forget the Canon R5C. It's.. unlikely that the R5II will have significantly better video specs. Yes, the housing is lange, but it comes with the benefit of a external battery pack that gets you over 3 hours of filming. You really can charge it once and use it all day. I only really see the R5II as a potential upgrade for photography or if you want to downsize from the R5C (I kinda do... but it doesn't make much financial sense). Fortunately, the wife now has an R6 II, and that's pretty compact in the Nauticam housing. So I just borrow that. For testing. 😄

Regarding the color temperature, I tested my Apollo III with a Seikonic C-800 spectrometer and got 6800k without diffuser and 6000k with. For comparison, I tested a few other strobes: 1. OneUW 160 - 5450k without diffuser, 5000k with diffuser 2. Supe D-Pro - 6500k without diffuser 3. Backscatter MF-2 - 5858k without diffuser, 5260k with diffuser 4. Inon S-220 - 7000k without diffuser, 6700k with diffuser 5. Apollo III - 6800k without diffuser, 6000k with diffuser 6. Retra Pro - 5100k without diffuser (Backscatter measured it at 5800k) 7. Sea & Sea YS-D3 (Backscatter measured at 6600k) 8. Ikelite DS230 5200k

Thanks!

Hi @UltralightCameraSolutions Do you guys have the weights of your warious arms and clamps listed anywhere? Or could you share here at least the weight of your standard clamps? I'm currently deep down the rabbit hole of optimizing the overall weight of my rig, and discovered that there's a significant difference in weight between clamps and arms from various manufacturers. Would be nice to understand how your offerings stack up in this area. Thanks! Andrei

Interesting findings, Phil. Did you run the test in a fully dark room, or one with ambient sunlight? I wonder to what extent the ambient light is influencing the results. That could explain why the color temperature changed at lower power. Since testing the OneUW strobes, I've been in contact with the owner and learned a few things that are worth mentioning. So please take the above results with these caveats: The strobes I have are some of the earliest models made, from late 2018. OneUW subsequently changed the design of the aluminium housing, and the new design is about 115g lighter without compromising on Also, the battery packs are old, which may be impacting the recycling times. I re-did the testing and was able to get 3fps at -4 stops and 6fps at -5 stops (with light output decreasing slightly over time). With new battery packs, the results may be better. It's worth noting that these results are on par with the DS230s, so not bad at all for NiMH strobes. 3. The color temperature remains unresolved. I will try to get my hands on a color temperature meter to supplement the testing done using a grey card and Lightroom.