164 replies to this topic

[Cybergoldfish]

Don't forget the leopard and guitar sharks around the alternatives, and the eagleray around the outside of the small passage too. Eagles look great in Natural lighting as Eric has pointed out.

[CeeDave]

Hello,

I am just trying to think through the (for lack of a better word) physics behind the filter/CCD/digital editing workflow. I am new to this, so I would appreciate it if you might consider commenting on which bits I still haven't grokked even after reading the fine articles by Alex and Craig.

Would it be right to say (for example) that

"When there is excess blue-green light, the blue ( B ) and green (G ) pixels in the CCD array will tend to limit the exposure time (to avoid blooming), whereas the "not blue-green" ( R ) pixels will have low input and thus will be subject to noise. Inserting a red filter will block portions of the blue and green light. Thus, a longer exposure (or lower f-stop) can be used. This allows the red pixels to "fill up" better and avoids noise in the R without blowing pixels in BG." Now, here's the important part, to me: "This is fundamentally different from postprocessing the image, because if BG is blown/clipped or R is noisy, these are hard to repair without introducing artifacts. 'Precorrecting' the light balance with the filter better uses the dynamic range of all pixels, which can theoretically lead to a superior image." ???

Shred away!

[craig]

No, you got it exactly right. All the digital mods in the world won't do any good if there's no signal to begin with.

[rstark]

A question. Instead of using a physical filter on the lens why not use the WB as a virtual filter? If you take a small color chart with you (laminated) and wb off the exact OPPOSITE color you want. That will have the same effect of using a physical filter. Example, if I want to "use" a red filter I white balance off of a cyan patch. You could have different shades of cyan to represent the different strengths of red CC filters. This does work. Can anyone think of a reason why it's not as good as a physical filter?

[herbko]

A question. Instead of using a physical filter on the lens why not use the WB as a virtual filter? If you take a small color chart with you (laminated) and wb off the exact OPPOSITE color you want. That will have the same effect of using a physical filter. Example, if I want to "use" a red filter I white balance off of a cyan patch. You could have different shades of cyan to represent the different strengths of red CC filters. This does work. Can anyone think of a reason why it's not as good as a physical filter?

It won't work. The comments above applies.

Underwater there's lots more blue light than red. If you choose an exposure that do not saturate your sensors with blue light, very little red would be captured. Any WB operation you may try, even if you have independently measured the lighting perfectly, will fail because you'll just be amplifying mostly noise in the red channel.

It's actually worst than that because the color filters on the RGB sensor have large overlaps. They're really very broadband filters centered around RGB. So a strong blue light will be picked up by the "Red" sensors, and even if your WB can substract off the right average value, it'll still leave behind lots of noise.

[craig]

A question. Instead of using a physical filter on the lens why not use the WB as a virtual filter? If you take a small color chart with you (laminated) and wb off the exact OPPOSITE color you want. That will have the same effect of using a physical filter. Example, if I want to "use" a red filter I white balance off of a cyan patch. You could have different shades of cyan to represent the different strengths of red CC filters. This does work. Can anyone think of a reason why it's not as good as a physical filter?

WB is not a substitute because it works after the CCD exposure and CCD's have limited range. If the correction is too extreme you will get nothing. If it is mild you can do it in Photoshop. When shooting RAW, WB has no effect.

Looking at it another way, filters modify the image before it is recorded. Color corrections, including WB, only change the interpretation of the image that already exists. WB only changes how we perceive the image. It does not fix any errors in the recording itself.

Since herb slipped in while I was composing, I'll add that what he said was true, too. Sometimes greens are equally strong or stronger than the blues as well.

[rstark]

Thanks. I understand.

[SharpDiver]

What rstark is proposing is somewhat like the result of a product called "Warm Cards", which Craig mentioned in this thread or an associated one. They come in a set and range in different intensities of pale blue. When used to white balance a camcorder or camera, they will introduce a slightly warmer tone. The bluer the card, the warmer the result. I bought a set to use with my video and some available light shots with my C-5050. I was not thrilled with the results. Luckily, the back side of the card is a 100 reflectance white, which worked nicely and the the lamination of the cards is completely waterproof. It may warrant more experimentation, but my initial results of trying to bias the white balance were not particularly successful. It tended to go too far, even when using the lowest card level (+.5).

[Swordfish]

I have a PT-015 housing. What does everyone think of the Cokin square filters (with the holder)? Will not saltwater damage the filters (I think they're plastic)? I'm from Manila and the local Cokin distributor does not carry threaded filters. There's an A and P series (which is bigger). I checked the Cokin website and I understand I need the P series for the 67mm diameter of the PT-015.

The following square filters are available here:

1. Red (003) filter [no CC*R] -- the guy I talked to says red is not a color compensating filter?; I also checked the Cokin website and they only have one type of Red filter.
2. FLD (046) filter [no CC*M filters available]
3. 81EF, 85a, 85b and 85c warming filters

Do I buy the red (003) filter or the FLD filter? Which warming filter should go with which?

I've read Craig's article and am I correct in assuming that if I use the FLD filter, I won't need a warming filter such as the 81 and 85 series filters? If I should still need one, which is suitable for the FLD? By the way, I dive mostly in green water but it would be nice to be able to use it as well in blue water.

Finally, where can I find out about the amount of mireds and magenta of the Cokin square FLD filter?

Thanks a lot and I really appreciate your help! =B)

Regards,
Norman/Swordfish

[craig]

Cokin filters should be fine in salt water. I have tried them but haven't used them long-term. Salt water dissolves gels, though.

The red filter is probably not suitable. I would suspect it is for special effects or B&W photography.

The FL-D filter is a combination of a warming filter and green remover (magenta CC). You would not normally combine a warming filter with this, but if you added and 81EF or 85C it would act more like an FL-B filter. A few manufacturers, including Cokin, offer a FL-W filter. FL-W's (Hoya is the only one I've tested) have almost no warming and a strong magenta component. The Cokin FL-W is likely to be similar to a slightly warm (81A) CC50M if it's anything like the Hoya. It would be very good in green water.

FL-D's as a group have modest warming and are suitable for green water where the blue is filtered down already. They are easier to balance with mixed lighting since they lack the strong warming influence that ruins strobes and lights. Since you intend to mount it externally this wouldn't be important to you.

Filter manufacturers don't publish their flourescent filter recipes. My data comes from purchasing a sample of each filter and measuring it with a Gossen Color Pro meter. I also have a spectrophotometer that I measure filters with but I haven't published data from it yet. I could measure the Cokin for you but I'll be leaving for Indonesia soon and it would have to wait until I got back (about 3 1/2 weeks).

[Swordfish]

Craig, thanks for the reply. I think I'd settle with the Cokin (no choice; and the square filters are cheaper [around $7 apiece here].

Here are the combinations I've thought of based on your suggestions:

8-15 meters (green water), no strobe:
FL-W + 81EF

8-10 meters, w/ strobe (green water):
FL-D

16-20 meters, no strobe:
FL-W + 85c [is 85 too strong?]

16-20 meters, w/ strobe:
FL-W

Any comments (sorry if I'm too persistent) before I go ahead with the purchase?

One last thing. I'm planning on getting a wide-angle lens in the future (not too near -- I'm so broke now because of the DS-125 purchase) This means I won't be able to use the Cokin filters with it, right? Is that one of the reasons why one should go with round screw-on filters?

Thanks again!


Regards,
Norman/Swordfish

[craig]

The beauty is that you can experiment. In shallower water and ambient light the FL-D might be better than the FL-W + 81EF.

If I were getting the 81EF, I'd skip the 85C and get the 85 instead. I think the FL-W + 85 is not too strong for 16-20 meters in ambient light. If it turns out to be you can switch to the 81EF. I wish I had tested the Cokin's so I could say for sure.

When you try these filters with strobes you may need to adjust your strobe output down (or close your aperture) compared to what you are use to. You may also need to adjust colors afterward so play with it a don't be discouraged! It's much easier to do when you don't have so much warming effect and that's what you've selected.

[Swordfish]

Craig,

If tweaking white balance is not good enough to achieve the effects of real world filters, how about this?

http://www.chromasof...hic_filters.htm

There's a review here:
http://www.outbackph...f_11/essay.html

You can use the demo 20 times.

What does everyone think of the software filters? What do you think is the catch?

Thanks again.

Regards,
Norman/Swordfish

[craig]

The catch is that they can't add back what's missing, and what's missing is the lack of exposure in reds that occurs because blue and green are too bright. No electronic adjust can ever be the equivalent of a filter.

Using moderate ranges of adjustment, software can appear to provide equivalent function to filters at the expense of noise you probably won't notice. The problem is that underwater, the adjustments needed are extreme. There can easily be a 6-7 stop difference in exposure between reds and greens. You won't see a solution to that problem used as an example. Take an image like that and try their software on it.

A more subtle difference is that filters work on full spectrum images whereas software works on tristimulus representations of an image. Since colors are aliased in any tristimulus image, electronic adjustments can't possibly duplicate the function of any arbitrary filter. It is completely impossible. Not all images will be effected, but some images will have subtle color differences or contrast differences when used with a real filter rather than a simulation of one. There's no way around it since what's gone is gone.

The software may be fun and useful for manipulating images but it's not a replacement for filters.

[rstark]

After reading a bit more about it I understand what is happening when I do color correcting in PS and how that introduces noise at different amounts depending on how extreme the changes are. I am in no position to write a tutorial on this because of my limited knowledge on it. But, may I suggest that someone (Craig) do a tutorial on this to better help those that assume it can all be done with post processing. It might even be a good introduction on your Filters and Ambient Light Photography tutorial.

[craig]

There's a good article being written (not be me) that hopefully will be posted soon. It will address correction of extreme color problems and should be excellent. I agree that a basic PS color tutorial for UW images would be a great asset.

[Helge Suess]

Hi!

There's a nice PS tutorial for colour correction of UW images written by Snoack.

Helge ;-)=)

[rstark]

None of the tutorials I've seen have mentioned what is being dicussed in this thread and how trying to color correct using software is just increasing "gain" to those channels which results in noise eventually. I just thought a breif discussion on this before the filter article will answer a lot of "but why don't you just....".

[Arnon_Ayal]

Few question about the gelatin filters:
1. How and where do you put the filters? Do you cut them and put it in the rear part of the lens and then close it?
2. Do you see any lowering of quality of the shot because of the usage of that filters?
3. What about scratches on the gelatin?

[DaNemesis]

Did anyone use this technique and post the resulting photos? I'm going on a trip to Grand Cayman and Cozumel and will be trying this out for myself, but would love to see how it worked out for one of you experenced people and hear any problems/suggestions.