Magic-Filter vs. PhotoShop

Example Four: When I jumped in without the flashguns (strobes) switched on and found myself face-to-face with a good subject - published several times (Nikon D200).

John, are you sure you posted the right pic? The illumination of the face and underbelly of the longimanus, is that from another source of light?

John, are you sure you posted the right pic? The illumination of the face and underbelly of the longimanus, is that from another source of light?

Oh oops! Is my face red? This particular picture has nothing to do with colour values. In fact it was an example from a talk I gave at the British Dive Shows on using RAW files and actually I jumped in without my strobes rigged and on minimum power setting so the exposure set on the camera meant that had it been on film it would not have 'come out'. Here's the original 'as shot' and the other is how it looked after adjusting the RAW file. Yes, there was a small amount of strobe lighting and the white hull of the boat probably helped too. Not in the other examples though! Sorry everyone. It was a mistake. Put it down to over-enthusiasm, grabbing pictures off divernet - and senility.

(...and how do you delete an attachment when you've put it on here twice at the wrong size?)

But in the end RAW + Photoshop will give you perfectly useable results and without the need to dedicate a dive to filter photography.

Alex, there's no need to dedicate a dive to filter photography, if you use a screw on filter, but it sounds like you are conceding that if you shoot raw and have good post processing skills, then there is no need for a filter anyway.

If I understand correctly the other points being made here. Then using a filter or a manual WB adjustment in the water, reduces the level of blue & green light to a level similar to that of the red and as a consequence the exposure has to be bumped up to deal with the lower light, so making noise more likely. I'm sure someone will correct me if I've got that wrong.

Though perhaps it's time to confess that I'm not a great fan of RAW, since in my post processing I have first to convert them to DNG then open CS2 (btw. I know I should have CS3) then get bewildered by all the available adjustments that I have to do in trial and error mode, until I get something just about adequate.

I sometimes wonder if all I'm really doing in post processing is trying to perform the same steps that the manufacturer's in camera software does to get a jpg image. Isn't the manufacturer likely to have better noise reduction than I can achieve in post processing, after all they know (or should know) the characteristics of the sensor better than me.

Now, I have done the same using a variety of filters as well as without. If I shoot with no filter, there's basically information in only the green channel in the picture. This in my eye tends to lead to a very low contrast image and trying to correct this using the RAW converter or channel mixer results in lots of colour noise. Even with the 70cc magenta filtration I get almost no blue! Hence, I'm currently looking for a stack of a cc50M and something bluish. I suppose what I would need is something like a Lee v98 that looks like this. Unfortunately, you cannot get it in a material suited for backmounting to the 10.5mm.

 

timo

Hej Timo,

Using 3500 K transparency film and shooting old Art Deco interiors illuminate with fluorescence light, I used so many cc gels on the back of the lens it turned my view camera into a point-n-shoot. I pick up a collection of these thin gels at used photographic equipment shows. Alas these shows and the wide use of cc filters are going the way of buggy whips (except for u/w use). Surely you can find some of these on your side of the pond.

Regards,

Bob

[see below]

" .... available in the three primary colors: Red, Green, and Blue; and in the three secondary colors: Cyan, Magenta, and Yellow. Kodak color compensating filters are made in the following increasing densities: 0.025(CC025), 0.05(CC05), 0.10(CC10), 0.20(CC20), 0.30(CC30), 0.40(CC40), 0.50(CC50). "

http://www.bhphotovideo.com/c/product/2686..._.html#features

Even with the 70cc magenta filtration I get almost no blue! Hence, I'm currently looking for a stack of a cc50M and something bluish. I suppose what I would need is something like a Lee v98 that looks like this. Unfortunately, you cannot get it in a material suited for backmounting to the 10.5mm. here.

 

timo

Send me your address and I'll send you all the Wratten ccM filters (75x75 and 100x100mm) I have when I get back to the UK in mid January.

JB

Well,

Finding the gels is not a problem as such. More finding the right combination of gels. The ccM stuff as well as standard Lee stuff is widely available both sides of the Atlantic. Kodak gels are even reasonably cheap these days. The problem is the more exotic colours like the lavender quoted above.

I suppose if I had unlimited time in my hands I could start from first principles, but as I do not, I'm trying to shortcut a bit and use something I already know sort of works as a basis. Also, I'm starting to get to a point of rapidly diminishing returns and I think I'll concentrate on some actual shooting as opposed to experimenting this coming season.

Which is not to say I'd not put John up for his offer...

timo

Let me say this again as some people seem to be missing the point...

White balance settings in the camera will not change the information stored for each pixel in a raw image.

As the name suggests, the raw image is exactly what the photosites record. The white balance setting will not change what is recorded for red, blue or green at all. Instead the white balance is used by the raw processor when it is converting from the raw information to something that can be displayed. So, the red, blue and green channels will record the actual levels of those (not withstanding Craig's comment about the spectra being continuous).

It is also worth noting that what is stored by the sensor is linear, basically counting the number of photons. Double the number of photons for double the amount of light. The problem here is where there isn't much light, there aren't many photons so the amount of information to store is quite low. In low light the difference in shades is quite small and won't be very accurate. The sensors work much better at the other end when you go close to blowing the highlights.

You can see it doing some simple arbitrary numbers. Lets say at one light level there are 10 photons. Double the amount of light and there are 20. Increase another stop and there are 40, and so on. When you consider that the sensors have a range of around 5 stops, that's 10 photons at one end 320 photons at the other end. Now, if you have a 5% difference in amount of light at the dark end it would be half a photon (which can't be recorded so you'd either miss it, or it would be recorded as a 10% difference. At the bright end, 5% is 16 photons so it could be recorded quite accurately.

The only settings that change what the sensor records are:

* shutter speed - a longer shutter speed will mean more photons

* aperture - a wider aperture will mean more photons

* ISO - I think. I don't really know how changing the ISO changes how the sensor records information

No other setting should have any effect on the raw image at all, other than the information about those settings being recorded in the file along with the raw image and the raw processor may (or may not) use that information when the image is converted on the computer.

Putting a filter somewhere between the subject and the sensor WILL change what the sensor records and has much the same effect as adjusting the aperture, only the filter may change the relative amount of the colours. A neutral density filter doesn't change the colours and has the same effect as adjusting the aperture (other than DOF, etc.). A magic filter will change the colours of what is stored in the raw image.

Adjustments in Photoshop (or equivalent) are essentially the same as white balance settings. All they do is change how the raw image is processed but not what is stored in the raw image. Post processing adjustments are not the same as using a filter. You MAY be able to achieve similar results, but it is definitely not the same thing. You MAY be able to produce acceptable results, but it is definitely not the same thing.

but it sounds like you are conceding that if you shoot raw and have good post processing skills, then there is no need for a filter anyway.

H Brian,

I'm fascinated by this discussion. I didn't get your interpretation anywhere here. I think the key is what one defines as "useable". As far as I can tell the color performance of the Magic filters is "better" than what one can do in PS, especially with regard to the blues of the background water. Alex has some comparison shots on the Magic filter website showing the difference. It doesn't mean that John's images without filters aren't great. By the same token just because they were published doesn't mean they couldn't have been better.

Though perhaps it's time to confess that I'm not a great fan of RAW, since in my post processing I have first to convert them to DNG then open CS2 (btw. I know I should have CS3) then get bewildered by all the available adjustments that I have to do in trial and error mode,

Brian, someone who shoots the quality of photos that you do might not need the advantages that RAW provides. I know that a lot of wedding and sports shooters use JPEG because the light is so stable and known to them in the church or the sports arena that they have it nailed and don't need the versatility shooting in RAW gives you. In a cursory look at your shots from Oman on Flickr I only see 3 of the 165 that I would have tried to improve the exposure on. That's pretty damn good. For me it's clear we have a broad continuum of folks here, especially when it comes to post processing. The spectrum runs from the guy who takes his CF card down to the drugstore and gets prints of his snapshots to the pros who spend hours, weeks or even months working to improve one image. They are both right in that they are doing what they want to do. The overwhelming draw for me is the passion I have for improving and producing higher quality images. Lucky for me I have a lot of room for improvement. Like you, I struggle with PS. Do yourself a favor and download a trial version of Lightroom. Import your RAW shot of the pair of Collared Butterflyfish, click on the Develop module and slide the exposure a little to the right. See what you think. Is it worth the hassle? your call.

Cheers,

Steve

I think you've hit the nail on the head, Steve. It's what works for you that counts.

For example, my knowledge of Photoshop proper as opposed to the RAW converter is almost non existent because I really only use the healing and cloning tools and maybe 'dust 'n scratches'. I work on the basis of my need. (So that's around 8996 opportunities missed with CS3!)

I used filters (Wratten85) when I shot film. I haven't needed them since although I suspect that James' complimentary filter on the strobe is the way to go and when I get a camera underwater that works better with higher ISO than the ones I use currently, I might well go that way (since filters reduce the amount of light). I am not averse to picking up good ideas here. (That's why I had to run out and buy the Tokina 10-17 and the Inon 240 strobes!)

The reason I mention that a shot was published is that the computer screen is very seductive whereas in 4-colour CMYK print you really see the problems if there are any.

So it's what floats your boat. My wife insists on shooting jpegs because she cannot be asked to spend time with a computer.

I've never had a case where a filter forced me to switch to an ISO I didn't want to use, and even if I did in the future I'd rather have my noise evenly distributed in the channels than to have it all lumped into the red channel and have red posterization to boot.

I had that problem in South Africa this year. My attempt to try filters for ambient light shots of baitballs backfired. Bad viz, wrong water color and the filter taking 2 stops away forced me to shoot at ISO800+ to get decent exposure, and that was at the max fstop of the lens (which also was a bad choice!). I didn't dare to turn on the strobes for fill because the water was so dirty. I definitely learned the hardway that filters, like CS4, is not the universal solution but just another tool for another dive. Choose your conditions and tools wisely.

That's why I had to run out and buy the Tokina 10-17 and the Inon 240 strobes!

John, that attitude for spending is what is needed in the economy right now... maybe I can convince the manufacturers to give us more toys to bring out the must have gadgets of '09.

John, that attitude for spending is what is needed in the economy right now... maybe I can convince the manufacturers to give us more toys to bring out the must have gadgets of '09.

That was 2006. This is 2009! I'm now using my obsolete kit (I'm now ashamed to mention what I use here) to try to earn enough to pay the taxes since my savings have gone missing (temporarily I hope).

All any coloured filter is doing is altering the ratio of the different colours in an image relative to each other. So using an orange filter will block blues and greens and boost the amount of reds and yellows present in the final file. White balance is something of a red herring if shooting raw - although it does have a slight effect on exposure. Not using a filter and relying on Photoshop to do the same can work, but as it has to 'amplify' the reds and yellows relative to the blues and greens it will also 'amplify' the noise - and this can result is a noisy image - especially as blue is already the noisiest colour too. No one filter can work in the very variably filtered daylight found underwater, but combining an 'average' filter with accurate white balance/'correct' exposure will usually produce pretty good results if the original contrast is sufficient - if its not then somewhere, something will give and noise will again rear its ugly head. I'd suggest that using an appropriate filter, carefully white balancing and then exposing so that highlights are not quite clipped is probably as far as you can go without adding illumination from strobe or lights. Adding in HDR might be a useful technique sometimes, if and when possible.

And believe it or not all this even works using obsolete kit, thus avoiding spending the taxman's money.

The way I like to think about it is that every shot has a range.

The range being the amount you can change a shot in post processing without significant loss of quality.

The range for people who shoot in RAW being much wider than people who shoot in JPG.

What adding a filter does is to shift the range along a bit.

So, for John Bantin, the end result sought usually falls within the range of the unfiltered RAW shot, and therefore shifting the range

along a bit with a filter does not make much of a difference. The end result being likely to fall in the overlap in the ranges anyway.

The advantage of using a filter with JPG photography is more obvious as the desired end result will fall outside the range much quicker.

There will be circumstances where using a filter will yield a better end result, but certainly not all the time. It is another tool available in the kit bag for when it is needed. There is also a good argument for getting as close as possible to the desired end result prior to post processing as that would minimize the number of times the end result falls out of range, there will be times that that will be with a filter.

Happy new year to all.

I had that problem in South Africa this year. My attempt to try filters for ambient light shots of baitballs backfired. Bad viz, wrong water color and the filter taking 2 stops away forced me to shoot at ISO800+ to get decent exposure, and that was at the max fstop of the lens (which also was a bad choice!). I didn't dare to turn on the strobes for fill because the water was so dirty. I definitely learned the hardway that filters, like CS4, is not the universal solution but just another tool for another dive. Choose your conditions and tools wisely.

Possibly another advantage, in situations like this, of cameras with high ISO ability, allowing more RAW post-processing WB and channel adjustment flexibility, and also significantly more flexibility to tone down or eliminate the strobe if we choose.

There's a lot of points being made here but how cameras work isn't among them. That's too bad.

Digital cameras implement variable ISO using an amplifier (actually 1 to 3 amplifiers) between the sensor and the converter. Programming the gain of the amplifier matches the voltage range of the sensor to the converter's range. There comes a point where increasing amplifier gain adds as much noise as simply scaling the digital values in post so, above that point, higher ISOs aren't useful in RAW shooting. Knowledgeable people spend a lot of time determining what that threshold is on a per-camera basis. This gets to Jeff's point: increasing ISO to compensate for light loss is a net win so long as you are operating below that threshold. Some cameras, like the D3/D700 and the 5D, have very high thresholds. How often do you find yourself needing ISO 1600+ underwater? Arguing that filters don't help because they require higher ISO is losing proposition much of the time. There are occasions when filters are problematic in that regard but those situations will probably result in monochrome images anyway.

The other thing to understand is the color balance of the light AND the natural color balance of the sensor. Ignoring the filtration effects of water, it would surprise many people to learn that there can be a stop of more imbalance between channels in good, outdoor light. Add 2-6 stops of effect due to the water and you have a great deal of imbalance to deal with in post.

Most of the time, the highlights we set our exposure to will be nearly white so we would like our channels to saturate at roughly the same time. With strobe photography the dominant light source is the strobe filtered by 2-6 feet of total water. With ambient and sunball shots the dominant source will be ambient. The trouble with sunballs is that they don't saturate the sensor except in the green channel. Most everyone has seen sunballs with ugly cyan rings and posterization. This is caused by overexposure (sometimes hard to avoid) combined with a poor white balance match in the sensor. Filters help noticably with this. So does removal of the hot mirror.

Anytime you balance the utilization of all the channels you will be better off assuming that hardware required doesn't prevent you from getting the shot in the first place. Drew pointed out an example where filters prevented him from getting the shots he wanted. Filters aren't a universal positive. Nothing ever is underwater.

Thanks for that Craig - as per usual with your posts great and to the point quality information.

Happy new year to you all and please keep posting next year!

Paul C

This subject does bring up alot of arguments it seems from both sides and we are not the only ones discussing (and perhaps fighting ) about this:

Sorry, everytime I see the thread title, I just kept on thinking..well.. there you go. Kind of rough.

Happy New Year everyone

This subject does bring up alot of arguments it seems from both sides and we are not the only ones discussing (and perhaps fighting ) about this:

 

Sorry, everytime I see the thread title, I just kept on thinking..well.. there you go. Kind of rough.

 

Happy New Year everyone

ROFL! Good one. Happy New Year back!

There are occasions when filters are problematic in that regard but those situations will probably result in monochrome images anyway.

 

Most of the time, the highlights we set our exposure to will be nearly white so we would like our channels to saturate at roughly the same time. With strobe photography the dominant light source is the strobe filtered by 2-6 feet of total water. With ambient and sunball shots the dominant source will be ambient. The trouble with sunballs is that they don't saturate the sensor except in the green channel. Most everyone has seen sunballs with ugly cyan rings and posterization. This is caused by overexposure (sometimes hard to avoid) combined with a poor white balance match in the sensor. Filters help noticably with this. So does removal of the hot mirror.

 

Drew pointed out an example where filters prevented him from getting the shots he wanted. Filters aren't a universal positive. Nothing ever is underwater.

I've posted this shot elsewhere but I'll repost it because it illustrates a real problem which is difficult to overcome using filtration - which is contrast. Underwater contrast is lower than on land (or at least in shots such as this it is in the mid-tones and shadows). The solution is to add illuminant to boost the light and thus contrast where needed - but if you are using a filter it means adding appropriately coloured light - ie a filtered strobe. In this shot I've used mostly ambient light but added a sufficient amount of strobe to give detail where needed. IMHO without adding illumination the dugong's face would be way too dark and would lack shadows to empasize it. Of course without a strobe it would be still possible to shoot a good image with filters and whether this would have been acceptable depends on personal preferences and style of image creation. Personally I have never been happy with using filters or just post processing ambient light shots, BUT this is a personal opinion somewhere between the turtle and shark - fighting over this seems a bit extreme - all techniques are valid and having a choice is great.

If you want to use either filters to 'correct' for colour imbalances or software such as Photoshop to post process a Raw file then you need to understand what you are doing before the image is shot and also what you are trying to achieve. It is also very important to appreciate the consequences of such actions one the 'quality' of the file. And on this note, the word 'quality' is oft used but rarely defined - but unwanted detriments may well include noise, coarse (not smooth) tonality, shadow artifacts, and banding/ colour fringing to highlights. As John has illustrated, your definition of 'quality' may well vary and depends on the purpose for which an image was shot - large 'art' prints require far more accuracy and consideration in file origination than do Powerpoint sized images or photos used on the web (although both these uses have their own requirements and their original files may not be as 'rough' as sometimes seems to be considered.

^ Yeah but I think the point here is that if you did not have strobes but only had a filter; the resulting filtered image (if compared to a photo without using filters) would:

- Retain more physical red data (and thus colour) in the photo

- Have less noise after WB adjustments

- Have deeper and richer blues in the background

Or am I missing something here