Strobe Test Results

I would like to add my thanks for sharing your report. Very useful and I have downloaded it and added it to my travel library of PDFs! Excellent.

I would be interested to see a more qualitative assessment of coverage, exposure and light fall off. Strobe performance is not just about the factors that can be quantified. Other factors such as the gradient of fall off of light away from the centre of the beam are just as important as beam coverage when it comes to producing pleasing lighting in wide angle images. A smooth fall off is much better than a sharp one.

So I guess what I am woffling about is that I would actually like to see the images exposed by the strobes - so that I can see side by side the coverage and exposure (which could just as easily be assessed from the f-stop required for a correct exposure - although in a reflective pool this might not be a very large f-stop).

Alex

I agree with Alex. It would be good to have more details about the charateristics of the strobes. Since reading your report, I thought a little about how to best get the most information. Here what I think would be a good test.

Sensor:

Use a camera as the light meter. The sensor are Si diodes not too different from the one in the light meter; you have millions of them and you can get out very detailed output. It'll have to be a camera with raw output. Something like a Olympus 5050 would work very well.

Lens:

Wideangle. It'll have to cover more than 100 degrees. Fisheyes are ok. Something like the Inon dome lens should work.

Target:

Use a large grey card, ~ 3x3 meters for the target. If you can't find one that big I think grey cloth from a local sewing shop will do.

Test:

Mount the strobe under test directly above the camera port and take shots going through the full range of aperture and shutter speeds. Also take a photo of a tape measure stretched across the target for reference. Use raw output.

Analysis:

Convert the raw to 16 bit TIFF using a linear converter. This will make your photos an intensity map. It's important that it's linear and not have a gamma curve applied. I know how to do this for Olympus and Canon raw files. Don't know about others.

With this data, you should be able to compare the strobes directly with the historgram, and by superimposing the tape measure pic in photoshop, you should be able to read off the falloff with the eyedropper tool.

What do you think?

Herb: you have almost exactly described the test setup and procedure that Craig is using.

Now if only he'd quit loaning out his camera gear... so he could finish the tests.

Cheers

James

Herb: you have almost exactly described the test setup and procedure that Craig is using.

 

 

James

Great minds think alike.

I would like to add my thanks for sharing your report. Very useful and I have downloaded it and added it to my travel library of PDFs! Excellent.

 

I would be interested to see a more qualitative assessment of coverage, exposure and light fall off. Strobe performance is not just about the factors that can be quantified. Other factors such as the gradient of fall off of light away from the centre of the beam are just as important as beam coverage when it comes to producing pleasing lighting in wide angle images. A smooth fall off is much better than a sharp one.

 

So I guess what I am woffling about is that I would actually like to see the images exposed by the strobes - so that I can see side by side the coverage and exposure (which could just as easily be assessed from the f-stop required for a correct exposure - although in a reflective pool this might not be a very large f-stop).

 

Alex

Alex, Yes understand. Actually the test that was done by Jeff http://www.digitaldiver.net/images/strobearticle/strobearticlescrn2.pdf ://http://www.digitaldiver.net/images/...ticlescrn2.pdf where he fired the strobes at a grey carpet is basically what we need but conducted underwater? Should not be a problem to do!

I agree with Alex. It would be good to have more details about the charateristics of the strobes. Since reading your report, I thought a little about how to best get the most information. Here what I think would be a good test.

 

Sensor:

 

Use a camera as the light meter. The sensor are Si diodes not too different from the one in the light meter; you have millions of them and you can get out very detailed output. It'll have to be a camera with raw output. Something like a Olympus 5050 would work very well.

 

Lens:

 

Wideangle. It'll have to cover more than 100 degrees. Fisheyes are ok. Something like the Inon dome lens should work.

 

Target:

 

Use a large grey card, ~ 3x3 meters for the target. If you can't find one that big I think grey cloth from a local sewing shop will do.

 

Test:

 

Mount the strobe under test directly above the camera port and take shots going through the full range of aperture and shutter speeds. Also take a photo of a tape measure stretched across the target for reference. Use raw output.

 

Analysis:

 

Convert the raw to 16 bit TIFF using a linear converter. This will make your photos an intensity map. It's important that it's linear and not have a gamma curve applied. I know how to do this for Olympus and Canon raw files. Don't know about others.

 

With this data, you should be able to compare the strobes directly with the historgram, and by superimposing the tape measure pic in photoshop, you should be able to read off the falloff with the eyedropper tool.

 

 

What do you think?

Herbko, Yes I understand the proceedure except the bit on the RAW 16bit linear convertor. Why is RAW so important for the test? Also regarding the histogram. How would this help the comparisson? The histogram simply displays the quantity of each colour, going from black to white.

I do not see a problem in conducting such a test, David what do you think?

Gary

RAW is uncompressed (at least that's what all the manufacturers' claimed) so it allows you to process the data the way you want it. As for the histogram I might try it out on terra firma using the same strobe manually fired from full blast to lowest F-stop to obtain a chart of the falloff.

Herb if you could post a sample of what you suggested it might help us visualise what you had in mind.

I also has an idea of a video of the test. Hang a large grey cloth and take a video of the strobe firing. This could be done above water since we are talking about the quality of the light falloff. In pitch dark free from other interference. The strobe will fire into the front of the cloth and the camcorder located at the back.

This way we will have thousands of frames we can analyse including the strobe duration. Well time to do it would be the problem.

Cheers

David

Herbko, Yes I understand the proceedure except the bit on the RAW 16bit linear convertor. Why is RAW so important for the test? Also regarding the histogram. How would this help the comparisson? The histogram simply displays the quantity of each colour, going from black to white.

 

I do not see a problem in conducting such a test, David what do you think?

Linear conversion of RAW format will give you an output that you can use to calculate relative intensities of the strobes. You can just read the numbers on the historgram. Doubling the exposure will give numbers that are twice as big.

Interesting test.

One has to remember though that power is second to quality of light.

Putting a diffuser on the strobe and using it properly brings in the best results.

Small strobes like Ike's 50's and the like, once fitted with a diffuser (to soften the light not to widen the beam), provide nice soft light. Main drawback on the small units is angle of coverage+single power level.

Larger units (125, 200) are versatile and great for wide angle because of their circular reflector. When combined with diffusers they remain our favourites.

Mastering lighting is key to great u/w photography and it takes time, practice and dedication.

Thanks for the test.

Michel

Thanks for this test. I'm looking at strobes and wanted this exact information. Great site for info, thanks everyone.

This is a great report. I often was asking myself how one strobe does compare to the other in real underwater practice. Thank you very much.

For an eventually upcoming further test I personally would enjoy to see the uw guide number work out in a more common / standardized way:

1m fixed distance and the resulting F-stop. The light fall off is not linear over the distance.

Even better: 1m distance with 2m light travel (flash – target – the way back, the Subtronic way for uw guide numbers).

This would meet the common formula topside divided by 2 to 3 (depending on the water) to give the uw guide number.

However the test gives great information how the strobes do in relation to each other. Very nice, thank you!

Regards,

Julian

Gazzer,

Obviously great effort went into this strobe test and I commend you for your effort. Some how these results however just don't seem right. Guide Number divided by strobe to subject distance equals Fstop. Although Ikelite claims a GN 32 u/w (GN32 divided by F8 = 4' which is very close to my results) you claim a GN of 49, or 6' stobe to subject. And this is low? If Inon claims a air GN of 70, there is no way you can end up with an u/w GN of 59. I think that either your flash meter was on drugs or there was something else skewing your results. 2nd point: light fall off of 1 Fstop from center to edge is maybe acceptable to some, but for me I'd be throwing away those images. 3rd: Most strobe tests (light falloff) are more easily visualized if you actually photograph the results ie: place the strobe at a fixed distance between the camera and wall, with a grid pattern for comparison.

If we are going to publish a test, lets make it scientific and acurate.

Doug

EDITED: sorry guys I posted my response after reading page 1 not knowing another full page of comments existed. I agree with the comments of Alex and Herb. This is a good starting point, now lets fill in the blanks.

Doug

Gazzer,

 

Obviously great effort went into this strobe test and I commend you for your effort. Some how these results however just don't seem right. Guide Number divided by strobe to subject distance equals Fstop. Although Ikelite claims a GN 32 u/w (GN32 divided by F8 = 4' which is very close to my results) you claim a GN of 49, or 6' stobe to subject. And this is low? If Inon claims a air GN of 70, there is no way you can end up with an u/w GN of 59. I think that either your flash meter was on drugs or there was something else skewing your results. 2nd point: light fall off of 1 Fstop from center to edge is maybe acceptable to some, but for me I'd be throwing away those images. 3rd: Most strobe tests (light falloff) are more easily visualized if you actually photograph the results ie: place the strobe at a fixed distance between the camera and wall, with a grid pattern for comparison.

 

If we are going to publish a test, lets make it scientific and acurate.

 

Doug

 

EDITED: sorry guys I posted my response after reading page 1 not knowing another full page of comments existed. I agree with the comments of Alex and Herb. This is a good starting point, now lets fill in the blanks.

 

Doug

Maybe they should not have used "Guide Number" in the reporting, but the results are consistant with what I would expect. Remember they are measuring incident, not reflected, light. Have the meter read F/8 at around 7 feet is about right.

Gazzer,

 

Obviously great effort went into this strobe test and I commend you for your effort. Some how these results however just don't seem right. Guide Number divided by strobe to subject distance equals Fstop. Although Ikelite claims a GN 32 u/w (GN32 divided by F8 = 4' which is very close to my results) you claim a GN of 49, or 6' stobe to subject. And this is low? If Inon claims a air GN of 70, there is no way you can end up with an u/w GN of 59. I think that either your flash meter was on drugs or there was something else skewing your results. 2nd point: light fall off of 1 Fstop from center to edge is maybe acceptable to some, but for me I'd be throwing away those images. 3rd: Most strobe tests (light falloff) are more easily visualized if you actually photograph the results ie: place the strobe at a fixed distance between the camera and wall, with a grid pattern for comparison.

 

If we are going to publish a test, lets make it scientific and acurate.

 

Doug

 

EDITED: sorry guys I posted my response after reading page 1 not knowing another full page of comments existed. I agree with the comments of Alex and Herb. This is a good starting point, now lets fill in the blanks.

 

Doug

Doug,

a light fall off of 1 stop is more than acceptable to me. If you take 0,5 stops or anything smaller you really would have to buy about 4 strobes to lighten your wideangles

1 stop fall off is also topside standard. Subtronic known as very honest with there given uw values bases them also on 1 stop fall off to the edges.

But Doug, it is more worse, your wideangles add some more light fall off to the edges. So you might end up with 2 stops less in the corners with flashed images

The guide numbers in this test are not representive (no fixed distance resulting in f-stop numbers). So you can not compare them directly to manufactures uw guide numbers (which on the other side don't fit either due to marketing purposes :wink: ). BUT however the given numbers set the strobes in comparison and give a true relation.

It is no big surprise to me that a ring type flash (Ikelite DS125) gives a "relative" low guide number in the center. This is tipycal for good flashes spreading their light within a big angle. It is very easy to build a small flash with a higher guide number in the center. You may call them spot lights or white laser pointers .

However, I am happy about this test. I could not find a better reference to these strobes until now.

That said, I went for Inons Z220 (1 DS125 = over 1300$ in germany, more than two Inons Z220).

regards,

Julian

Julian,

What I was trying to point out (and what Herb and others also suggested) is that although these tests are a good point of reference, they are not complete nor scientificly valid. Light fall off along a horizontal and vertical reference is important, but what about hot spots, what about the quality of illumination, what about the color tempreture? This is why imaging the result is important. As a serious but amatuer u/w fine art photographer I do not have the ability to test every strobe that comes out to see if it will do something that my other gear can not, so these independent tests are important to me. Asking for more thorough and accurate information benefits us all.

Many people get very opinionated about why their camera, lense, strobe etc is better than every other brand on the market but in my opinion it's all just bs. So long as the gear does what you need it to and is reliable - brand is insignificant (ask James about why he loves his S2pro and he'll probably say "because it works and does what I tell it to do". I'm glad you like your Z220's, I own and use (2) Ds125's, (2) ds-50's a SB103, and a slew of MCD slave strobes. But if something better came along that did something I couldn't otherwise do with my existing strobes I'd change in a heartbeat.

Doug

Hi Doug,

you are right, I agree with you. In general there are almost no real strobe tests around. All the last points you have mentioned (color temp, ...) are mostly secrets. In spite of not having a curve from center to the edge I was quite happy that the 1 stop fall off was taken. And it is nice to have this test as there is almost nothing else available.

I am not happy with my Inon Z220 because I didn't take them underwater until know. I have two SB-105 which have a about 1000K warmer color temp. . The Inon data say it's cold as a landgun which I also have housed. So I will see. In worst case I will have to exchange them for 2 Subtronic Mini TTLs (ring type, warm) as they are also capable of pre and mainfalsh in a very short time which is required when the E-TTL conversion will work. The big Subtronics are not compatible for sure like my SB-105s -> slow circuit. Therefore I now have just one Z220 in order to see how it works for me.

regards,

Julian

Hi guys and sorry for not having visited this thread in some time. Would just like to comment on some of the recent posts.

Guide Number - My understanding of guide number is f stop x distance. Is that not how a guide number is calculated? There may be some recognized standard for what distance to measure at or what aperture to use but I can't see why in our tests that really matters. What we are looking for is comparison. No doubt a strobes underwater guide number is not critical as long as it can reach a certain minimum level. We make that clear in the report, I think! Far more important is the angle of coverage etc. All strobes were measured with the same flash meter. We can't state that these are definitive results as the meter was never calibrated however we can be certain, due to multiple repeats of the same strobe test and obtaining the same results, that they are as accurate as the meter would allow and certainly reflect a true comparison of the strobes. The whole point of the test! I will state again, for us, guide number is not a major issue but since manufacturers insist on quoting it in the specifications we thought it would be useful to try and get a real life comparison.

Beam Angle - Really my comments reflect what I said above. We took the 1 f stop drop off in light as one that I believe Ikelite state as the desired point at which angle should be calculated. Again this is only in an attempt to get a comparison between strobes. Our test results do not of course show the drop off in light and how quickly it occurs, the real shape of the beam and evenness. An earlier strobe test, taking pictures of strobes fired at a grey carpet produced interesting results. Our intention would be to repeat this test but underwater. Use a grey sheet marked with a grid pattern and take pictures using the test strobes. The next step of the test! Again our report was done to compare the beam angles quoted by manufacturers as it's something they always list in strobe specifications.

Strobe Light Quality - The actual strobe color temperature is something that we need to measure. We did try shooting a colorful picture underwater but the results proved nothing. We had to take a series of shots at different settings and or vary the strobe power to obtain a properly exposed shot. The varying in settings causes a big difference to the light quality and colors reproduced in the image. So we ended up with a series of shots from each strobe which frankly all looked similar when comparing the strobes. Admittedly this was done at the end of our pool session and we did not take the time over it that we should have. We are still thinking of a test that would most accurately reflect the strobes underwater performance as our initial attempts proved to us that it is not an easy thing to measure and quantify! Any suggestions?

Gazzer,

first things first. As I stated before your strobe test is great in my opinion!

For determining the coverage angle one f-stop light fall off to the corners is absolutely correct. Some have stated an interest of having a ‘fall off curve’ from centre to the edges. So maybe some sample photos would show this. But the measuring method for the coverage angle is perfect and therefore nothing to improve/change from my point of view.

Guide number underwater estimation is not easy as there is no prescribed industrial standard

for the manufactures, as on land. In my opinion they just can tell you what they want.

The topside rule ‘GN = f-stop x subject distance’ where subject distance is prescribed 1m for measuring is not valid for underwater.

Underwater you have to take into account the same light absorption with growing distance as we know it with sunlight.

The guide number is not linear with growing distance.

Therefore underwater GN has to include the light travel distance (subject distance x 2). This is different to topside where just the subject / flash distance counts.

As you can see the guide number decreases with growing distance underwater in a not linear manner. With very short distances you almost reach the topside value.

Unfortunately there is no prescribed distance when manufactures give their underwater guide number as it exists for landguns. So they can mean just anything.

I don’t know on what Ikelite’s given values are based on. Inon just gives topside values.

But Subtronic has a very nice and standardized way: 2m light travel distance. That means 1m subject distance. It is a fair value and also comes close to real world underwater strobe use.

My suggestion is to take this fixed 2m light travel distance and determine the resulting f-stop.

By the way, you can check the measured values afterwards.

With subject distance 1m (2m light travel distance), the underwater guide number can be calculated within an exactness of ½ f-stop:

clear water (visibility ~25m, test pool for example): topside GN divided by 2.236

muddy water (visibility ~2m): topside GN divided by 2.824

Again, I enjoy your strobe test, especially the coverage angle table is great.

The test sets the different strobes into an interesting comparison which is seldom given somewhere else. And you clearly stated and correctly called the GN table “our underwater guide number”.

If you want to do one more test, the above is just my suggestion on how to approach the item of non-standardized underwater guide numbers.

Personally I don’t see an important reason to do a colour temperature test as requested. The manufacture’s values are measured topside. So the Kelvin value should be quiet right. And everybody knows that a decent ring type flash is about 1000K warmer than a small simple tube. The Inon tubes for example look pretty similar to landguns. And Inon confirms this with 5500K.

Regards,

Julian

calculated uw GN, 2m light travel.

Could not include Ikelite strobes. No topside data available. And don't know how they estimated their uw guide numbers.

Ikelite gives the power in W/s. That's very nice. But this doesn't lead to any uw GN as it depends on how the electrical energy is converted to light output (reflector characteristic, colored glass for warmer temp, dome glass for spreading the light, and so on).

I think the UW guide numbers do not differ to much with some midrange strobes. the more important question is on what coverage the GN works. So it is no surprise that a small and cheapish flash beats a big one with real good coverage in terms of GN. just think of a laser pointer :wink:

calculated uw GN, 2m light travel.

 

 

Could not include Ikelite strobes. No topside data available. And don't know how they estimated their uw guide numbers.

Ikelite gives the power in W/s. That's very nice. But this doesn't lead to any uw GN as it depends on how the electrical energy is converted to light output (reflector characteristic, colored glass for warmer temp, dome glass for spreading the light, and so on).

 

I think the UW guide numbers do not differ to much with some midrange strobes. the more important question is on what coverage the GN works. So it is no surprise that a small and cheapish flash beats a big one with real good coverage in terms of GN. just think of a laser pointer :wink:

Hi all,

Newbie looking for strobes here. I was trying to download the pdf report but was not able to due to the page not being there anymore. Can someone assist me in getting a copy of the report?

Much appreciated.

Thanks.

Barttrigger