Step up filter ring

[leo]

Hi,

I'm new to the forum. I really appreciate digital photography and I had started to learn how to produce what I think it's beautiful from this website. So I thought I would give the forum a try.

I have a pretty basic but specific question, since I have NO experience with lens filters.

Can step up rings compromise image quality, say creating extra vignetting, or flare? I'm not talking about the filter itself; just the ring. If I use a step up ring from, say, 49mm to 52mm, or 52mm to 77mm, for a CPL, ND, GND (with a linear gradient), or a four point cross filter, will that be in any sense not as good as using the right size filter directly attached?

Thanks
-Leo

[Dave Humphries]

Hi Leo,

Welcome to the CiC forums from me, good to have you start with a question.

As long as we are talking about the GND/CPL/etc. filter diameter being bigger than the lens filter thread, which is usually what step up means, then no, you normally* wouldn't see any vignette.

* I might envisage that fitting a stepup ring of just one small step size up (say 49-52) on the front of an already fitted UV filter and then having a fairly thick filter on the front of that just might be an issue because the depth of the ring plus the new filter exceeds the step up in diameter - but I think this is unlikely to happen unless the lens is fairly wide angle in reality.

The only other quality reducing factor would be if this filter assembly is used without a lens hood (or pokes out in front of it) and there is a significant amount of off axis light hitting the filter - that might cause a vieling flare.

Cheers,

[Colin Southern]

Hi Leo,

A big welcome from me too

I agree with Dave, but I'd just like to quickly add a few things ...

- If you're using CPL / ND / GND filters then it's unlikely that you'll be shooting wide-open (eg F2.8 / F4 / F5.6 etc); far more likely that you'll be stopped down (F16 / F22 etc) ... and in which case vignetting won't be an issue regardless. If you're shooting wide-angle though, you may well get light fall-off due to the extreme field of view ... and this in turn may LOOK like vignetting, but it's not.

- If you're going to be using a GND filter then avoid the screw-in types like the plague; in practice you'll seldom want the transition zone right in the middle, and that's the only place they put them with screw-in types. Singh-Ray make the best sliding GND filters, which is what I use exclusively. I've you're interested, I've written a few articles for their blog detailing how I use them.

- With regards to hoods and step-up rings ... unfortunately they're often mutually exclusive because you can't get the hood on once the stepping ring is mounted, and if you mount the hood first then it's usually pretty tricky to mount the stepping ring and adjust the likes of CPL filters (depending on the lens focal length).

Hope this helps

[rpcrowe]

Colin wrote: - "If you're going to be using a GND filter then avoid the screw-in types like the plague; in practice you'll seldom want the transition zone right in the middle, and that's the only place they put them with screw-in types. Singh-Ray make the best sliding GND filters, which is what I use exclusively. I've you're interested, I've written a few articles for their blog detailing how I use them."

I totally agree but I think that I would have a use for a round GND filter IF... If some company produced a round GND with a cut off about 1/3 into the filter. That way, I could still use the round filter (and incorporate my existing lens hoods) but, be able to frame the image roughly according to the rule of thirds.

[leo]

Thanks for the replies! That settles the question. Also thanks Colin for your comment on GND filters, I will take a look at those.

Best,
Leo

[leo]

Hi, I'm kind of lost, I hope someone reads this.

So, I followed the suggestion on rectangular filters from Colin Southern. The concept of LEE and Cokin system is very attractive. The only problem is that I can't find reviews on the quality of filters. At B&H and Adorama there are some reviews on Cokin filters and it seems consistent that people say the ND and GND change the color of the photo, a magenta cast. My first question: could that be easily resolved in digital with white balance?

I'm tempted to forget about buying several screw-in filters (except for the CPL) and instead adopting Cokin or LEE, but without at least some samples of how image looks like with filters made for these systems I can't tell if would be a good investment. By that I mean I would get not only GND filter, but also ND, special effects, color, warm etc.

I wish there were review sites like photozone or dpreview with filter reviews. Any suggestions?

Best,
leo

[arith]

I don't know if this is any good but it has a GG2 Kood filter on it; I can't tell any difference between Cokin and Kood, but certainly Kood stepping rings are as good as any and a lot cheaper.

GG2 is a soft 2 stop ND gradient. So about 1 stop top to bottom in this and the only ambition is to get a little more detail in the wheels.

[Hansm]

Hi Leo,

Investing in LEE or Cokin adaptors is never a waste. You can use the Cokin, LEE or Singh-Ray filters, but also you can use every other type of glass you want.
This gives you lots of freedom in experimenting with unusual combinations.
About the colour cast also LEE filter give cast if you put two or more after each other.
When only using one filter I didn't see any cast.

[Colin Southern]

So, I followed the suggestion on rectangular filters from Colin Southern. The concept of LEE and Cokin system is very attractive. The only problem is that I can't find reviews on the quality of filters. At B&H and Adorama there are some reviews on Cokin filters and it seems consistent that people say the ND and GND change the color of the photo, a magenta cast. My first question: could that be easily resolved in digital with white balance?

Hi Leo,

In short, best quality is Singh-Ray - medium quality is Lee - and "budget" quality is Cokin. Personally I use ONLY Singh-Ray GND filters, and people tell me my stuff seems to come out OK

In terms of correcting colour-casts ... it CAN be an issue with any resin filter (which the sliding rectangular ones are). Generally speaking it's not an issue with around 3-stops, but it can start to be an issue with 5 stops or more, and it can be very difficult / impossible to zero out in post-processing primarily because it's an infra-red component that upsetting the apple cart. In these situations I use a Singh-Ray Vari-ND for the basic attenuation (2 to 8 stops) and then use a 3-Stop resin GND filter for dynamic range compression. If you do that - and shoot RAW - then so long as you expose correctly for the highlights then you shouldn't have an issue with shadow detail.

Unfortunately, when it comes to filters there are only two kind; good, and cheap!

My advice would be that if it's a degree of overall attenuation that you're after (for smoothing water / blurring clouds etc) then beg - borrow - sell the girlfriend - do whatever it takes, and get a Vari-ND. You then get 2 to 8 stops of attenuation (in one filter) - continuously variable - and NO colour shifts. And you'll have it for the rest of your life, unless you damage it. I shoot with a Canon 1Ds3 camera ... and the Vari-ND is probably higher quality! It is one classy bit of kit (and they now have a version with a CP built in if that's important to you) (personally I never use a CP).

In terms of a GND filter - made sure you gear up for the 100mm wide variety (and Lee holders will be higher quality than Cokin) (I'd recommend the Lee foundation kit) (can supply links if you need them), and then just get the best quality you can afford. Again, if you go for the expensive ones - and look after them - then you'll have them for life. For sure they're not cheap, but neither are cameras and lenses, and what's the point of visibly degrading your images after forking out big $$$ for camera and lenses (and tripods and ball heads and remote timers / releases) because you bought a budget quality filter? (THat's like buying a Ferrari and putting pram wheels on it to save money).

Hope this helps

[arith]

I have not noticed colour cast and I've got both Cokin and Kood, but I'm not into wasting money they are different. If anything I think Kood is better, it defies logic but when I use Cokin I think I have got a lot of dust on the sensor. It is weird because it is all gone when I change to Kood.

[leo]

Thanks everyone for posting.

Colin, I think I will not go into GND filters right now, since to make it worthwhile it's costly. Apparently it is possible to reproduce the similar effect of a GND by taking several photos with different exposure and then creating a gradient blend with some software. I will first try this procedure; in situations that something is moving I may try the continuos shooting with a large exposure bracketing. Then if I'm not happy with the results of digital GND I will consider buying some of those filter systems.

I got a B+W ND 1.8 filter for long exposures. I did not get anything lower than that because I figured that if I want a bigger aperture I can just set the shutter speed high. I have a Nikon CPL so if I need more than 1.8 ND I may try combining both.

Anyway, just a thought: Wouldn't be a trivial thing for a CMOS sensor to have a gradient of ISO along some specified direction? It seems natural that one could set up a portion of the photodiodes to be at, say, ISO 100, and another portion to be, say, at ISO 800, with some blending in the middle. This would allow to create a low quality GND; better than nothing it seems, specially for small prints and if the camera has a very good noise control. And one could circumvent a bit the noise issue by averaging pics.

Best,
Leo

[Dave Humphries]

Anyway, just a thought: Wouldn't be a trivial thing for a CMOS sensor to have a gradient of ISO along some specified direction? It seems natural that one could set up a portion of the photodiodes to be at, say, ISO 100, and another portion to be, say, at ISO 800, with some blending in the middle. This would allow to create a low quality GND; better than nothing it seems, specially for small prints and if the camera has a very good noise control. And one could circumvent a bit the noise issue by averaging pics.

Hi Leo,

That idea came up a few weeks ago and I can possibly see it coming in say, 2 years.

Here's how I envisage it being useful; imagine a touch screen LCD about the size of an iPhone, which as we know, has a nifty multi-touch screen allowing one to easily position (up/down/left/right) and twist, the gradient to suit the image content as shown on the screen view Live View. Of course, since an iPhone sized screen is too big to fit on the back of most DSLRs, it would have to be a wirelessly linked Live-View screen (Dream on Dave) which would also function as a remote release and flash controller. (Dream on Colin)

Back to reality - one might argue that the Nikon's D-Lighting, and whatever the Canon equivalent is called, are already a manifestations of active picture content aware dynamic range control, because they are, in localised areas, varying the effective sensor iso. The trouble is, they can leave halos around bright or dark things if they are surrounded by a large area of uniform mid-tone. This bugs me so much I leave mine switched off, as this is one control that I believe (but I may be wrong) affects the RAW data, so cannot be easily reversed in PP

Cheers,

[Colin Southern]

Colin, I think I will not go into GND filters right now, since to make it worthwhile it's costly. Apparently it is possible to reproduce the similar effect of a GND by taking several photos with different exposure and then creating a gradient blend with some software. I will first try this procedure; in situations that something is moving I may try the continuos shooting with a large exposure bracketing. Then if I'm not happy with the results of digital GND I will consider buying some of those filter systems.

Hi Leo,

There are big debates over "digital GND" -v- "in-camera GND" - I use either; so as far as I'm concerned it's just another tool in the toolbox, with each having potential advantages and disadvantages depending on the situation. With regards to movement, it all depends on where the movement occurs; if it DOESN'T occur in your transition zone then it usually doesn't create any issues. If you're going to combine images digitally then using a gradient blend often doesn't produce an optimal result; in my opinion, a better way is to stack the two layers - attach a mask to the top one - and then paint the mask with an opaque brush (perhaps set to around 10%, but varying as required) so you can blend the areas as required, which is what I did with this image (the bottom portion is 3 shots taken with a GND filter that are effectively averaged for the mist effect, whilst the top is effectively a different exposure, and I've hand-painted the transition zone for a seamless blend).

I got a B+W ND 1.8 filter for long exposures. I did not get anything lower than that because I figured that if I want a bigger aperture I can just set the shutter speed high. I have a Nikon CPL so if I need more than 1.8 ND I may try combining both.

It's a good theory anyway

Anyway, just a thought: Wouldn't be a trivial thing for a CMOS sensor to have a gradient of ISO along some specified direction? It seems natural that one could set up a portion of the photodiodes to be at, say, ISO 100, and another portion to be, say, at ISO 800, with some blending in the middle. This would allow to create a low quality GND; better than nothing it seems, specially for small prints and if the camera has a very good noise control. And one could circumvent a bit the noise issue by averaging pics.

The problem is that the photosites are geared for a base ISO of 100 for Canon and (I believe) 200 for Nikon; there's no way to make the photosites less sensitive than that by varying A/D gain, so you can only go the other way (by raising the A/D gain or "ISO") and - unfortunately - with landscape photography you generally have far bigger problems with ISO not being low enough (I'd brefer to have ISO 100, 50, 25, 12, 6, 3 than 800, 1600, 3200, 6400 etc).

[Nass]

Hello Leo,

It depends what you're stepping up from and to, and what stepups you're using. I think, for example, if you tried to mount for argument's sake an 88 mm on a Sigma 10-20 (itself 77mm) and you were using the 10-20@10mm, then its field of view is so wide that stacking the 2 or 3 stepups required would cause vignetting, yes. On a lens of a different mm, say 28mm, from personal experience it's not a problem at all. I have an ND110 in 77m and I use about 6? stepups to be able to attach it to any 49mm threaded lenses I want.