Thursday, August 10, 2017

Filters in Photography

Brief discussion about optical filters in conventional (film) and digital photography. There's a lot of misconceptions about their effects and even finality of use by beginners and even more advanced users.


Let's start by the simplest filter ever, the UV. It looks like a flat glass disk without any obvious feature, but as it name suggests it blocks, or at least attenuates, the amount of ultra violet light that reaches the film or sensor.

Film is sensitive to UV light and the excess of it can cause a bluish cast and haze on the photo. But why this ?

Let's talk about ultraviolet light. It's the light spectrum with wavelengths between 400nm and 100nm, being 400nm the very edge of the visible spectrum and 100nm a very penetrating and dangerous form of electromagnetic radiation. The Earth's blocks the shorter wavelengths more than the longer ones. 

To keep it simple, accept that the 100-300nm part will be totally blocked from space to the altitude of about 7km. Some of the 315nm may reach the lower altitudes and even the surface and longer ones will reach it, depending on the latitude, season, solar activity, pollution and other factors.

But what causes the bluish cast ? One of the causes are the very small dust particles in suspension. Small particles are able to reflect small wavelengths very efficiently, and thus, why they reflect more blue-violet than red light. That reflection makes the particles itself visible. Blocking the shorter light wavelengths will make the particles less visible, resulting in a clearer image.

Note that the particles also absorbs light from what's behind them, and this absorption will not be reduced or influenced by the UV filter.  

Films are sensitive in some degree to UV light and this is also an important issue.

Digital cameras are less prone to the problem but in a minor degree, they still are.

The camera sensors usually have a filter to block UV and IR (infrared) light and let pass the visible spectrum without being attenuated. This filter is technically a combination of a high pass and a low pass filters, resulting in a bandpass one.

But the filter cut-off limits aren't like a wall. It's a curve with a certain inclination.

Schneider B+W 010 UV Haze transmission curve
(C) Schneider Kreuznach 
The graphic above is the transmission curve of a typical good quality UV filter. The darker curve is for a multicoated filter. Note the higher transmission because of less loss of light due reflections.

The visible light window is, for practical information, between 400 and 700 nanometers. At about 800nm note that the filter also start to block infrared light.

This particular filter blocks 100% of the UV shorter than 350 nm but just 60% at 375 nm. This means that some UV will pass, depending on the wavelength. It's a more sophisticated filter, if compared with the standard UV filter.

But there are lots of makers and filters belonging to the UV family and quality varies a LOT depending on this. Some filters are flawless and other will degrade the image quality.

Keep in mind, for any filter type:

- Good filters aren't cheap but they don't degrade the image sharpness.

- A good anti reflection coating is a must. Ignoring this may result in undesired reflections and flare.

- Good UV filters have almost no light absorption.

- There are lots of counterfeit filters on the market. Be careful with eBay, Amazon and similar sources. If possible always buy from a trustable seller or store.


First of all, a Skylight and UV filters are NOT the same thing. The main difference is the pinkish tint on the Skylight.

The Skylight filter blocks UV light but also attenuate a little the purple-blue part of the spectrum, resulting in a slight warmer image.

Both filters are extremely useful for film photography, specially for mountain/snow/beach areas.

There are different types of filters belonging to this family:

- SKY 1A
- SKY 1B
- Haze 

The difference between them is basically the intensity of the warming they add. Those filters are maybe the most widely used on film photography and have less impact on digital due the possibility of post processing.


Those are supposed to be neutral filters, with no attenuation of the light spectrum and no color cast, at least on theory.

Again, a good coating is always a good idea and of course good filters come for a price.

They are excellent for lens protection against impacts and they saved me more than once. I remember some years ago, I was taking pictures at a mountain place and a small pebble came from somewhere and hit exactly the center of my Leica Apo Telyt and cracked a Hoya HMC Skylight filter. I was using a metal lens hood and it didn't make any difference. If the lens was hit, I'm pretty sure that it would be damaged.

It's also easier to clean a filter than a lens and if by some bad luck you scratched it, just buy another one.

In my view, good quality protection (or UV) filters are a must have. 


Black and White FILM photography can be tricky if we're talking about contrast and tones.

It's easy to imagine a scene in plain colors, but not so easy to have a precise idea about the final look on black and white film.

B&W photography is all about contrast and tones and to make things worse, there are color tones that will be rendered exactly the same way on film, even being quite different colors.

There are four main filter types for B&W photography:

- Yellow
- Orange
- Red
- Green

There are other ones, like blue, lime, but let's talk about the four main ones for now.

Light Spectrum
Take a look at the spectrum image above and let's talk about the work horse of the b&w filters, the yellow one.

Transmission curves for filters (C) Hoya
Y2 = Yellow
YA3 = Orange
R1 = Red
X1 = Green
X0 = Lime

Filters like the yellow, orange and red are lowpass filters. They attenuates higher frequencies and let pass the lower ones with sight or no attenuation.

Take a look at the figure and let's talk about the transmission curve of a yellow (Y2) filter. The cut-off frequency of this filter is about 475nm. It will block all wavelengths below, let's say, 425nm. But the blue is at something like 470nm and will be just partially blocked. Looking at the graph I would guess that the blue light will be reduced by something like 10-15%.

This slight reduction on blue light can result in a slight contrast enhancement between a cloud and the blue sky. The sky will look a bit darker than the clouds.

The orange filter transmission curve "knee" is shifted to a longer wavelength, and will reduce dramatically violet, blue and some green.

And the red filter will cut from green to violet. 

A yellow filter makes colors with shorter wavelengths than yellow to appear darker than the longer wavelengths and all the longer wavelengths will have the same look as without the filter. The orange and the red ones follow the same rule .

Green filters have a different transmission curves. The Green (X1) is a band-pass filter, with the central point at something like 540 nanometers and it will attenuate any wavelengths in both directions but in a not so steep way.

Green filters makes green objects lighter (or brighter if you prefer) than objects of other colors. They're also nice for portraits of people with white skin, giving a more natural tone.

The lime filter is something like a combination of a band pass and a low pass filter, and can be used as a general purpose filter for b&w photography, if you plan to photograph people and vegetation.

The BLUE filter does the opposite of what the Red does. It will lighten the colors at the blue side of the spectrum and darken the ones at the red side. Blue filters are high-pass or band-pass filters, depending on the manufacturing process.

Important to mention is the fact all the above mentioned filters blocks 100% of the UV light.


All filters blocks some light and this quantity is called FILTER FACTOR. It's expressed in how many times the light will be attenuated. It's simple, if you're using a red filter it will reduce de light to 1/8 or -3 stops. Remember to compensate the exposure if you're not using TTL metering or if the filter is not covering the light sensor of your camera.

B&W Filter effect table

The skylight filter is like the Y1 filter, but for color photography.

Do NOT use physical Y/R/O/G filters on digital cameras with Bayer or X-Trans sensors, even if the camera is set to monochrome. This WILL cause interpolation errors during the demosaic process and may give very strange results. In this case you need to use software based filters.

But if you have a Sigma camera with the Foveon sensor and plan to shoot in b&w, you need the filters, like you were using b&w film. By the way, Foveon's pictures shot this way are amazing.

Original image in color
Yellow filter, slight blue attenuation
Orange filter
Red filter
Green filter
Blue filter, severe yellow to red attenuation

Another set with more green.

Color, no filter
B&N no filter

Tuesday, July 25, 2017

Fujifilm X-TRANS processing - Part 1

This is not a review. Just some short comments about the most important programs that can handle X-Trans raw files.

X-Trans raw files are a pain to process. Really.

The complicated nature of its non standard demosaic algorithms are a real nightmare for both users and developers. 

From the user side, there are just a few programs that are able to extract the full potential of this new class of color matrix, and all of them, besides the awfully complicated RawTherapee and Fuji's OEM version of Silkypix are paid and usually expensive.

Things aren't different from the developer's point of view. There are almost no technical information about the mathematical methods for the decoding process, and many of the amazing sharpen and noise reduction algorithms used be well known programs just can't cope with nothing else than the traditional Bayer pattern. That's why, for example, DxO just gave up on any non Bayer sensors.

These are my own opinions based on my experience and needs. Of course opinions can vary from people to people.

The most practical options on market due this day are:


It's a very powerful software and probably the best one in terms of extracting the highest detail from the raw files. It's really good on this and also on noise reduction and film profiles. The price is fair.

The drawbacks are a horrible interface and it's slow. By horrible and slow I meant really horrible and slow. Its features are sometimes hidden in non obvious places and I never managed to find how to apply setting on a group of images in real time. I suggest you to try to do this.

But it's the best program if you want to have extremely detailed images from the X-Trans sensors.


It's a stripped down Silkypix 4. It's free, so don't complain. =)

The interface is way better than the previous program but still have some serious problems with (maybe) the Japanese to English translation. Some terms are just too weird.

The overall results are pleasing but it's slow like hell.

SILKYPIX 5,6,7,8

The full version of Silkypix isn't that good to justify the extremely high price tag. It's still slow, but it has many more features than the OEM version, and can deal with lots of raw formats.

I think, something between $100 and $150 would be a nice price instead of the ripping almost $300 they ask for it.


It's free, fast, well documented BUT it's the most complicated image processing program I've ever used in my life.

The sharpening and noise reduction functions are overwhelming and extremely comprehensive. Curves, color and histogram operations are on pair.

The results can be awesome if you don't go nuts using it.


Well, they work but honestly their raw engines are not a match for neither one of the above mentioned programs and they are very expensive for what they are. You need to do a true olympic marathon in adjustments to get "near" the image output from the other ones, so I won't even comment them. I just gave up on them.


This was a surprise. Its raw engine is still under development but I got some interesting results from it. I have to do some more experiments before giving it an honest rating.

So, how do I cope with the extreme sluggish Iridient ?

I just created a default profile for the used camera, for example a X30, with just the Iridient Reveal sharpening, hot/dead pixel auto removal and auto white balance. Then I batch process all my images this way and save them as TIFF. 

If you use Windows, you can use the Iridient Transformer the same way.

Then I process the TIFF files with DxO, but you may use whatever you want.

It's much faster than any other option I had tried.


If you're a Mac user, be informed that Aperture can handle Fuji's files pretty well, but you need to be running at least Yosemite's version of Mac OS.

Bottom Note:

In the end, the X-Trans matrix doesn't appears to make any real improvement over a Bayer (without the AA filter) besides maybe some higher detail on high frequency zones. But the Fuji's package is very powerful considering the price and final result. Still a very good and solid option. 

Fuji always liked to try some different approaches on sensor design, like the Super CCD and its numerous incarnations and I respect this. At least, like Sigma with the Foveon, they try.



Saturday, July 15, 2017

Cool Cameras 3 - Rollei 35

Rollei 35

This is a very nice camera. It's probably the smallest all mechanical and manual 35mm camera made.

The lens barrel shown at the above image is at the collapsed position. The shutter needs to be armed to retract the lens. Again, odd. I would keep it unarmed to not stress the spring without need.

The control layout is weird, with speed and aperture dials placed at the camera's front. The light meter is at the top and the flash hot shoe is under (yes, under) the body.

It's very well made and rock solid. There are many variations of the Rollei 35: 35, 35S, 35T, 35 LED, 35 SE, 35 B ... The differences are usually the lens type and the lightmeter. For my taste, the best picks are the 35 and 35S and I would avoid the ones with aperture and speed rings around the lens barrel because they're very prone to fail.

The one above is the original one with an uncloupled match needle type light meter on top and a 3.5/40mm Tessar lens.

Variations can have the Triotar, Tessar and Sonnar lenses and even a led based light meter. But all are manual focus and all mechanical cameras. There's no focusing aid like a rangefinder. You need tho estimate the distance and set it properly on the lens distance scale.

Virtually all lens types used on this camera are very, very sharp, even the Triotar triplet. The later models with the HFT Sonnar are multicoated and superb.

The early types were made in Germany and the late ones in Singapore. It's not correct to say that the German made ones are better, both shares the same quality.

This camera is easy to repair and maintain, but the match needle analog light meter is simply horrible and it's rare to have one working in perfect order due wiring oxidation. It's also not very accurate. To make things worse it uses a mercury battery, so you have two options: 1) Use a Zinc-Air battery  2) Use an alkaline or silver oxide one and have the meter properly calibrated. To be honest it's  better to forget about it and have a handheld one.

To take pictures, you need to pull the lens, like in the image below.

The camera back needs to be removed for film insertion. just release the lock and pull down the back.

Note the swinging film pressure plate. You need to position the film UNDER it.

The viewfinder is very spartan. Just a very clear Albada type with framing lines.

Below, some servicing pictures:

Light meter showing oxidized wiring and contacts
Corrosion due bad materials and a light battery leak
Rollei 35 without the light meter
"Good" (haha!) light meter installed on it. I had to change some wiring, not elegant but it works.

You can download the service manual here.

Sunday, June 18, 2017

Poor man's fluid scanning (wet scanning) with Epson V700 scanner

I bet that 90% of the people who bought the Epson V700 scanner to scan film is at least very frustrated with the so-so scan quality from 35mm films and the original Epson film holders.

There are three main reasons for this:

First, (Lack of) Flatness:

The original Epson film holder is, let's say, horrible (to avoid saying any ugly words). It's flimsy and without any structural integrity. The film is always not flat due the lack of compressing bars between the frames.

Second, (Lack of) Focus:

The film plane with the Epson holder is never at the optimal focus. There are just two height adjustments, but it never stays at the optimal height. 

You may try to use a good quality third party film holder like the ones provided by and even use anti newton glass to keep the film strip as flat as possible. Those film holders have several screws for proper focusing adjustment and it's very labour intensive to have all set to the proper height.

The Betterscanning holders aren't cheap but they provide a very ample fine focus adjustment and compared to the Epson holder they're much better. But there are still two problems: It's expensive, specially if you order it with the anti Newton glass plates. And because of the extra glass, there are some contrast loss due to reflexions. The second problem is that the Digital ICE infrared dust removal works very badly with the extra anti Newton rings glass. You mus have pristinely clean films to use the Betterscanning holder (and similar ones).

Third, Too many Air to Glass interfaces:

This will always cause contrast loss. If using the original holder, there's the scanner bed glass between the film and the scanner optics.

Too many air to glass interfaces decreases contrast and may add odd reflections. Usually the reflexions aren't a big problem. Below, that's what happens to a light ray when it hits a glass plate: dispersion and internal and external reflections making it bounce in and between other glass surfaces.

The additional dispersion is almost certain to make the Digital ICE to fail. You may check if it still work.

The dispersion effect is much higher with anti Newton glass because in fact, this kind of glass has one (or both) external surfaces etched to minimize the glass to film contact. The effect of this is the reduction of Newton rings. 

The internal reflectance at an air/glass interface for light rays from a point source in glass. Light rays incident at angles to normal at greater than the critical angle (here, 41° for glass to air) do not leave the material and are reflected at the glass/air interface. source: ASU 

What's fluid mounting (wet mounting) ?

It's when you use a fluid between the film strip and the glass to keep it "glued" to the glass to make it as flat as possible and also to remove one of the air to glass interfaces.

There are two main ways to do this. The first one is to use a dedicated fluid mount holder system, like the one Epson send with the V750. There are some clones on the market.

Epson Fluid Mount
Epson fluid mount for the V750

It's basically a holder with a glass bottom and an alignment support with a printed grid.

The idea is to:

  1. Put some sort of fluid over the glass
  2. Place the negative over the fluid
  3. Allign it using the grid pattern
  4. Cover the film with a transparent mylar or polyester film. You need to apply some fluid over the film, between the film and the transparent film.
  5. Squeeze out any bubbles with a very soft cloth
  6. Place the holder over the scanned flatbed and scan

This works well IF the focus distance between the holder and the scanner sensor is correct. It's the same focusing problem all over again and you still have a damn glass-air-glass path between the film and the scanner sensor.

The fluid mount holder must be at the proper focus distance

The holder has at least two advantages: First, it keeps the fluid out from your scanner glass. Second, it's easier to align the film. But you still have the two problems I mentioned above.

Again, the fluid mount holder is not exactly cheap.

What I'm doing is simply to fluid mount the film directly over the scanner glass. Yes, i know this may sound creepy but believe me, it's completely safe if you know how to do it properly.

The main concern is about damaging the scanner. This is simply the case to not let the fluid leak to the scanner interior. Just use the proper quantity and it will not leak.

Some people use mineral oil as a mounting fluid. Please don't do this. It's a total mess and a real pain to clean both the film and the glass. I've tested many fluids and, believe me, the best one by far is the old good Zippo Premium lighter fluid (black can). It will not attack the film emulsion neither the glass. It's very volatile and will let no residues behind. Just use the minimum quantity you can.

For the transparent film, I use those cheap transparencies used on overhead projectors.

If you're paranoid with the eventual fluid leaking inside your scanner, you can apply a thin silicone rubber sealing on the glass-plastic gap or even using adhesive tape. I use UHU's Patafix for this task.

Mounting directly to the scanner glass removes the last two problems and in my case, a great improvement in image quality.

Wet mount directly over the scanner glass
Epson V700 at 2400 dpi

Saturday, May 6, 2017

Zeiss Nettar 515/2

Zeiss made many cameras under de Nettar name, with various choices of formats, lenses and shutters.

Like all Zeiss cameras, this one is very well made, with an black enameled body with a nice leatherette finish. 

Zeiss Nettar 515/2

Compur Shutter and Nettar Anastigmat

The one I have is a prewar 6x9 with a Nettar Anastigmat and a Compur shutter. It uses 120 film and gives 8 beautiful and large frames. The last Nettars were built in 1955.

The lens, a 105mm f4.5 uncoated triplet is quite simple, with the expected optical aberrations for this kind of lens, specially at large apertures and at the corners, but it can be sharp enough if stopped to f11. Keep in mind that this lens is uncoated and very prone to flare without using a suitable lens hood. This lens works beautifully with black and white film.

Nettars have many shutter flavors: Compur, Prontor, Pronto, Vario, Prontor-S, Prontor-SV, Prontor-SVS, Klio, Telma, Derval and Vero. Yes, is this crazy. There are at least 12 Nettar types: 510, 510/2, 515, 515/2, 516, 516/2, 517/2, 518/2, 515/16, 516/16, 517/16, 518/16

My 515/2 has a nice Compur with 1s to 1/250s + B and is supposed to by one of the "high end" Nettars along with the top model with the Compur-Rapid (1/500s) and a f3.5 Nettar Anastigmat.

There's no rangefinder, just a barebones foldable viewfinder (open type, Galilean) for framing.

Friday, May 5, 2017

Takane Mine Six IIF

Ever heard of the Takane Mine Six IIF ? No ? Don't worry, this is one of those obscure Japanese folding cameras from the '50s.

It was made by Takane Kogaku. You can find some historical and technical information on Camerapedia .

Takane Mine Six IIF

The first thing you put your eyes in is the beautiful Takumar lens. Yes, a Takumar, nothing that you usually see over there. It's an Asahi T-Takumar coated 75mm f3.5 triplet. The diaphragm iris has 10 blades and a beautiful almost circular shape.

The shutter is a COPAL with speeds from 1s to 1/300s plus B.

Contrary to the vast majority of folding cameras with Copal and Compur shutters, the focus is not done by rotating the first element, but by moving back and forward the entire shutter/lens assembly. If you try to rotate the front lens you will unscrew it, be advised.

Takane Mine Six IIF lens and shutter

One interesting thing is the possibility to shoot in 6x6 or 6x4.5 formats. There's a switch near the viewfinder to overlay the 6x4.5 mask over the 6x6 viewfinder. If shooting in 6x4.5 you'll need to use a built in foldable mask for the film, but you can't shoot both formats on the same film roll. 

There's a coupled coincident image rangefinder for focus adjustment. The viewfinder is very basic.

I would rank this camera along with the AGFA Isolette with the Apotar but better than Agnar and Zeiss Nettars. 

Repair tips:

If you see small perforations or even small tears on the bellows, you may try to fix them using black liquid electrical tape. It worked in my case.

If you need to CLA the lens/shutter assembly I have good news. The front group (2-element) goes out simply by unscrewing it. The rear element likewise.

The shutter mechanism is exposed after removing the first element and the retaining screw round the lens thread. Then remove the cover plate and apply lighter fluid all over the gears. Let it dry and then lube the gear axis applying a light mineral oil with a toothpick on the proper places.

If you need to clean and relube the shutter remove the front and rear lens groups first. 

Clean the lenses using your preferred method and reassemble everything. Easy peasy.

The rangefinder adjustment is done by a small screw mounted on an L shaped part, visible on the lens and shutter image above.

The focus adjustment is awkward. Its helicoid is accessed by the rear side and it's a pain to adjust. Just use a spammer to adjust it, but be careful to not damage the bellows or the rear element.

If you need to disassemble the helicoid, I wish you good luck...

Thursday, April 20, 2017

The myth around the Nikon "universal" mount. Not so fast !

Before some people start to go crazy about what I wrote, be advised that I have used Nikon for years. From Nikkormats to D800. So just relax and read before raging against someone who dared to say something against your god. =) 

It's all about Nikon and a huge legion of its users and fans saying about that Nikon has the most backward compatible lens and mount system on earth, and because of this you can use any Nikon lens on any Nikon body. Wrong. Very Wrong.

The real fact is that Nikon has no total backward compatibility and even worse, even lenses and camera bodies from the same year may not be compatible. The nikon mount is a mess by definition. Some lenses can even damage some bodies. What was supposed to be good can turn into a real nightmare.

Lenses from 1959 to 1976 are the original A mount type. They can't be mounted in ANY Nikon body except pre-Ai bodies and the ones that have a foldable aperture index tab on aperture coupling ring, like the flagship Nikon DS. 

If you try, for example, to fit an old Nippon Kogakku 50mm F1.4 from 1970 , not modified to Ai, on ANY camera without the folding tab you may break the body's aperture coupling ring and stuck the lens.

There are two types of autofocus lenses, with the AF motor built in the lens itself and without it. 

Lenses without motor won't AF on camera bodies that don't have motors. 

For example:

The D5XXX has no AF motor on the body neither the aperture coupling ring. You can use AF lenses with motor and Ai/Ais manual focus lenses that have electronic contacts. No contacts = no light metering.

The D7XXX has AF motor and the aperture coupling ring. You can use all lenses on it , but any old "PrĂ© Ai" lenses.

There are even lenses with and without aperture ring. You can't use any lens without aperture ring in any non AF nikon, for example, the FM2. The FM2 will accept just Ai/Ai-S lenses or AF fullframe lenses with aperture ring.

The mess:

- Pre Ai lenses
- Ai and Ai-S manual focus lenses with aperture ring
- AF lenses with and without built in motor
- AF lenses with and without aperture ring
- Full frame and APS size sensor sizes
- Cameras with and without aperture ring
- Cameras with and without built in AF motor
- Coupling "horns" of two types


- If you have an OLD Nikon like a Nikkormat you need a pre Ai lens with the famous "horns" to couple the lens aperture ring with the camera aperture ring. 

Be advised that there are TWO "horn" types and they're not compatible and mounted in opposite ways.  Pre Ai lenses have the aperture coupling "horns" facing to the lens front. Ai and later lenses have it facing backwards, to the lens rear. If you plan to use an Ai lens on a pre-Ai body you need to reverse the horn orientation to have the proper coupling.

Nikon Pre-Ai coupling "horn"
Nikon Ai coupling "horn"

- If you have a camera like the FM, FE, FM2, N2000 you need an Ai / Ai-S or pre-Ai converted to Ai lens, with aperture ring for proper metering.

- For autofocus cameras like the D70 and D5xxx you can use AF lenses with or without aperture ring but you need the AF motor on the lens to have autofocus. You need lenses with CPU for proper light metering.

- If we're talking about high end autofocus cameras like the D600 and D7xxx you are allowed to use all nikon lenses but NOT pre-Ai ones, unless they were converted to Ai. Those cameras should work with all AF lenses, with or without autofocus motor and aperture ring.

- Finally if your camera has a retractable tab on the aperture ring you'll be able to use any kind of Nikon mount lenses on your camera.

If you want a really universal SLR system, go for Pentax. All K lenses work on all K bodies and you have an extra bonus, can also use all M42 screw mount lenses on a Pentax K body with a small and cheap adapter.