Luminous Landscape Forum

Equipment & Techniques => Landscape & Nature Photography => Topic started by: howard smith on November 01, 2004, 03:57:40 PM

Title: using graduated nd filters correctly
Post by: howard smith on November 01, 2004, 03:57:40 PM
A split ND will accomplish the same thing, it is just much harder to place to edge.
Title: using graduated nd filters correctly
Post by: howard smith on November 01, 2004, 06:18:37 PM
Generally, a reflection should be a little darker than the thing being reflected.  A reflection does not intensify the light.  At best, they are the same.  Try a spot meter to see jst how difference they are.

I still think the idea should be to expose for the lower tone you want and add ND to get the brighter hightlights the tone you want with that exposure.

To do this, you will need to know the exposure latitude of the film of camera you are using.
Title: using graduated nd filters correctly
Post by: howard smith on November 03, 2004, 12:37:47 PM
Being a user of film, I prefer to get as much on the film as possible rather than patch together pieces.

I usually use my spot meter (1 degree) and meter a place in the foreground that I want to look right.  That is usually a shadow, so I stop down about 2 stops.  I then pick a place in the upper portion and do the same metering.  I then use ND to bring it the right number of stops above the foreground spot.

I have always had a lot of trouble placing the filter, so I don't use them much if I can help it, and bracket the placement.

If you that and bracket the exposure, you can run up a lot of frames.  I prefer to learn how to meter and trust that.  Maybe someday I will get the hang of split ND.
Title: using graduated nd filters correctly
Post by: jamesdak on November 19, 2004, 07:54:05 PM
On the Maxxum 7 I am able to use a custom function of the AE lock button to get the metering value of all 14 segments to show on the back LCD.  Makes it real easy to place a split ND for an even exposure across the frame.
Title: using graduated nd filters correctly
Post by: Ray on November 20, 2004, 08:17:15 PM
Photomatix? Never heard of it!  :)

I'm a bit skeptical of many of these programs which essentially take a Photoshop process and turn it into some sort of 'action', whether or not it be named as such.

The parameters are usually set within a range that accommodates, perhaps, most situations but not the hard cases.

They can be great time savers, perhaps, but don't encourage Photoshop skills.

I'm also spreading myself a bit thin. Instead of writing this, I really should be out mixing some concrete for my driveway  :D .
Title: using graduated nd filters correctly
Post by: Jonathan Wienke on December 04, 2004, 07:43:55 PM
Quote
Quote
If you're using an aperture if f/16, you're only using a small portion of the center of the lens and a graduated ND filter isn't going to accomplish much.


That's a joke, right?  Just so we're clear.
It's not a joke. Stopping down reduces the effectiveness of ND grad filters unless they have a very sharp transition from the ND portion to the non-ND portion. As you stop down, the area of the front lens element actually used for the exposure becomes smaller and smaller, and unless the gradient transition of the filter lies completely within this area, the effect of the gradient will be diminished. The ND grad filter will behave more like a standard ND filter. See this page (http://visual-vacations.com/Photography/aperture_and_lens_usage.htm) for a visual explanation.
Title: using graduated nd filters correctly
Post by: Jonathan Wienke on December 05, 2004, 07:59:05 PM
I've posted a follow-up to this discussion in a new thread here (http://www.luminous-landscape.com/cgi-bin/forum/ikonboard.cgi?act=ST;f=3;t=1089).
Title: using graduated nd filters correctly
Post by: jayz on November 01, 2004, 01:23:38 PM
i finally got some grad ND filters recently, but am finding my exposures still are not exposed correctly.
could someone tell me what i am doing wrong with my expoure technique.
last week i took some sunrise pics of a mountain reflected in a lake.
exposed for the water with mountain reflection (foreground), and the snowcovered mountain (background) seperately.
the foreground expousre was metered using spot metering at f16 and 1/4 sec.
the background sunlight snow covered peak spot meter reading was f16 and 1/30.  i added two stops since the sunlight snow was brighter then medium (f16, 1/125).

i exposed this scene at f16, 1/4 sec, and stacked a 3x and 2x graduated ND filters, with the clear part on the reflection and the grey nd part on the sunlight peaks.  I used the tree line as the point of graduation.

my exposure was still off.  The water was dark, and the sunlight peaks were washed out.  what went wrong????
Title: using graduated nd filters correctly
Post by: Sam NI on November 01, 2004, 04:04:47 PM
Quote
If you're using an aperture if f/16, you're only using a small portion of the center of the lens and a graduated ND filter isn't going to accomplish much.
???

-------------------------------------------------------------

Well, Jayz, meter for a mid-tone anywhere you can, slide the grey grad into position and, with slide film, bracket your shots for insurance. If you nail your mid-tone right, you won't have to worry about the snow or reflections.

But, if you're using a DSLR, the chances are you'll have more problems with contrast and will have to get down and dirty with contrast masking. As a keen amateur, here's the way I see it, FWIW:

Quote
Expose separately for the sky and the land.

You can bracket for the landscape – -0.5, meter reading, +0.5 – but try starting off by simply exposing for the mid-tones in the shot and then push the exposure to the point where the histogram shows more information to the right of the display, without the abrupt peaks that indicate blown highlights. The histogram is really the key here, in my experience. Doing this will maximise your exposure for the ground area and give you more to work with later on your computer.

But, if there's a lot of water or bright reflections in the shot, I've found it will really take a couple of land exposures to get it really right and good digital blending (from RAW) to make use of all the data. I'm not a big fan of RAW - as an amateur with a modest workflow, it's a bit of a fiddle in my opinion! - and rarely use it, but it will certainly help to rescue a couple of stops of highlight detail that you may not think are there.

Depending on how bright the sky is, you could try -2.0 and work up from there, although I've found even with a cloudy, bright sky the highlights can still blow somewhere! Try RAW if this is a persistent problem and do a search on RAW-editing methods, and levels and curves too. The Luminous Landscape has very good tutorials on this, dealing with sky problems and highlights elsewhere in the shot. Blending is very effective, and impressive. It's not that difficult really, but perhaps a little time consuming for some. These together will will give you the best control over difficult highlights.

So, if it's digital landscapes, shoot RAW, expose for the sky and the land separately, bracket if it helps you, get the sky the way you want it without blown data, and always keep a close eye on the histogram.

 From SLRs to DSLRs (http://www.theimageplane.net/tip2_075.htm)
Title: using graduated nd filters correctly
Post by: Jonathan Ratzlaff on November 01, 2004, 08:20:46 PM
Generally I use grad filters at small apertures, (no, Jonathan W you are not correct, the centre of the lens is not used for small apertures)  stop the lens down to f 11 or 16 to assist in placing the grads.  A reflection is usually about 3 stops darker than the object causing the reflection.  
For most images I use spot metering, however when I use graduated filters I usematrix metering.  The graduated filter in the bright portion evens out the exposure to a great extent.  For further control, I bracket downward from the given exposure, 2/3 stop for each bracket.  My experience with my equipment is that these images tend toward overexposure and that bracketing downward is more useful than bracketing over and under.  
I usually find that the 2/3 under exposure gives me the best result.
I think that in this case the next time you try this, take a number of images with different combinations, take notes and compare the results on a light table.  This way you will be more confident with your results.
Title: using graduated nd filters correctly
Post by: Bruce Percy on November 03, 2004, 11:26:32 AM
Quote
If you are going to blend, you don't need split ND.  Bracket enough you don't need an exposure meter.  Split ND is good for getting a good exposure on a single frame.  A good meter is useful for getting a good exposure the first time.
Hi,

I tried blending, and found it did not give the same result as using an ND grad. In my own experience, I found using an ND grad produced a more natural result, because sometimes the blend in photoshop would look really false.

I tend to expose for the ground, and use a Sekonic 608 light meter (spot at 4degrees) to measure the contrast between sky and ground. If there is 3 stops of a difference, I use a 3 stop grad. Surprisingly, I found the hard grads much better than soft grads, and that's with using a rangefinder camera too!

In terms of placement, I tend to look at the scene, decide if the grad needs to be 1/4, 1/3 or 1/2 over the lens, and then place the filter on the holder to suit. This is done without looking through the lens (I'm using a rangefinder). But I maybe take several identical exposures with different placings (usually they look similar on the light table), if i'm not sure.

Perhaps you are metering bright areas of the ground? Try to meter something in the ground area that looks fairly 'average'.... experience will help as time goes by.

Hope this helps.
Title: using graduated nd filters correctly
Post by: christoforos on November 05, 2004, 11:20:42 PM
Was the ISO setting on the meter the same as the film you were using???
Title: using graduated nd filters correctly
Post by: didger on November 20, 2004, 08:33:51 PM
Quote
I'm a bit skeptical of many of these programs which essentially take a Photoshop process and turn it into some sort of 'action',
This is a stand alone program, not a PS plug or "action".  It's clearly quite sophisticated and it can blend any number of images, not just two.  I've had field situations where I felt that two bracketed shots might not be enough, so it would be good to have a choice of blending more than two.  The sampe pics and the few user feedback reports are pretty impressive.
Check it out (http://www.hdrsoft.com/index.html)
So, just how important is that driveway project?
Instead of writing this I should be continuing to work on my overhead camper project for my truck so that I can get out for some more shooting!!  
Title: using graduated nd filters correctly
Post by: didger on December 04, 2004, 08:11:08 PM
Quote
It's not a joke.
Hey, I also thought it was a joke.  Who's going to straighten us out here if Jonathan ever leaves?
Quote
I don't think that separate software for blending is needed just good knowledge of PS.
Not even good knowledge of PS is required.  The layer mask method described in Michael's blending tutorial (http://www.luminous-landscape.com/tutorials/digital-blending.shtml) is so dead easy that anyone can do it with no real PS expertise at all.  As if that weren't enough, here's an action (http://www.didgeridoings.com/BlendingAction/) to do it for you.  Just take save the action to your Photoshop actions folder and then load it into your actions pallette and run it.  I have a about a zillion bracketed blending pairs and I've found this method quite bombproof and it costs nothing, not even any time to speak of.  
Title: using graduated nd filters correctly
Post by: Ray on December 05, 2004, 02:28:47 AM
There's something else going on here (I mean apart from the extreme modesty of some members of the forum  :D ).

I also have a slight vignetting problem with my Sigma 15-30. For those not familiar with this lens, it has a fixed lens hood and no means of attaching a protective UV filter or polariser except by slipping a short tube (threaded at one end) over the lens hood. For full frame 35mm it's a pretty useless arrangement, but for the cropped frame of the 20D it's almost perfect.

However, at 15mm the rim of the filter causes very pronounced vignetting. But it's only in the extreme corners and covers a very small area of the frame. It can usually be cloned out quite easily, but it's still a nuisance.

After reading Jonathan's theory that at f16 one is using only a small portion of the centre of the lens, I couldn't help wondering why I hadn't noticed that the vignetting problem with the Sigma is no longer a problem when the lens is stopped down. So I've just taken a few shots of a cloudy sky at various apertures ranging from f3.5 to f22.

The vignetting is very noticeable in all shots but its character is slightly different at different apertures. But no big deal. There's no aperture which removes the vignetting and oddly enough, it's actually worse at f22 in the sense there's a slight imbalance with the vignetting in one corner being more opaque and solid, almost as though the rim of the filter was more in focus at f22.

Explanations, anyone?
Title: using graduated nd filters correctly
Post by: sieracki on January 09, 2005, 10:51:14 PM
I'll use grad filters only under certain conditions. First is that the SBR exceeds the dynamic range of my sensor (I'm shooting digital) and second, there are no objects (such as trees) that extend up into the highlight region. For example, I might encounter a scene such as Moraine Lake in Banff NP in Canada. Quite often, clouds cover the peaks surrounding the lake. These clouds are often very bright and the lake is quite dark since it might be in shade early or late in the evening. I could just shoot the lake, ignoring the clouds and peaks, or I could just shoot the peaks and clouds and ignore the lake. If I want both areas with detail then I think about using a grad filter.

How do I know if I need a grad? Well, most digital cameras have about a five stop (or more) dynamic range. You need to test this yourself and it isn't completely objective (it's up to you to decide just how dark shadows can get before they are too dark). For the sake of argument, say we have five stops to play with. This means that I can see detail in a print or on a monitor in areas that are five stops apart in brightness. Let's suppose I meter the clouds and they read EV 16. Most spot meters allow readings in EV's or exposure values. Then I meter the lake and I see that it is reading EV 9. That's a 7 stop difference. Either my highlights (the clouds) or the shadows (the lake), or both, will lack detail, depending how I determine exposure. In this case, if I place a two step grad filter, I'm hoping to bring down the highlights two stops to be 5 stops, equivalent to my dynamic range. But the clouds may not be stopped down two stops since the filter is only 2 stops at the top edge. In this case, I might reach for a 3 stop grad, try that, and then go back to the 2 stop for insurance. The placement of the graduated portion of the filter matters too. Using more strength than is needed can be quite counterproductive. You also have to remember that your foreground exposure will increase if your highlight exposure is increased. Something bright in the foreground might be overexposed as a consequence (I had a stick in the water in one of my shots that got overexposed and it looks like there is a bolt of lightning in the water...).

So I base the whole thing on knowing my dynamic range and what the SBR of the scene is. Then I try to position the camera so that I don't have trees or anything vertical protruding into the highlight region since trees will be dark on top and light on the bottom.

Placing the grad properly involves using your depth of field preview button if you have one. Stopping down the lens makes the grad line more obvious in the frame. Jiggling the filter a bit helps you see where the line is.

Also look for really neutral grad filters. I've been using Hitech and Singh-Ray's. If you can afford them, go for the Singh-Ray filters. Never buy the round grads. Only buy the square or rectangular systems. The round ones cannot be adjusted as far as where the grad line is falling.
Title: using graduated nd filters correctly
Post by: howard smith on November 01, 2004, 01:43:41 PM
I would suggest checking the exposure of the upper half and lower half.  Expose the lower half to look the way you want.  Add as much ND to the upper half to make it look like you want using the same f/stop and shutter speed as the bottom.
Title: using graduated nd filters correctly
Post by: howard smith on November 01, 2004, 05:04:22 PM
If you are going to blend, you don't need split ND.  Bracket enough you don't need an exposure meter.  Split ND is good for getting a good exposure on a single frame.  A good meter is useful for getting a good exposure the first time.
Title: using graduated nd filters correctly
Post by: slacker on November 02, 2004, 12:10:27 PM
Hi Jayz,

The method you're using appears to be fundamentally sound; this is almost exactly how I meter and shoot these scenes and have had pretty consistent success with exposures.  Like you, I shoot film and use grads on at least 50% of my landscape shots.  I use soft grads, but that will only mean a difference in the transition zones and not in the fg and bg where you seem to have not achieved expected results.  Reading the explanation of your process, I can suggest a couple of things to look for and try that may help find the problem or improve your chances:

- you indicate you spot metered the fg; you have to be careful, then, that you are choosing an appropriate spot.  I use spot metering also but I will point it at several spots in the fg before choosing my base exposure.  The reflected snow peak will still be quite bright, relative to darker places in the water; you indicate your result was still underexposed in the fg

- you also indicate though, that the bg (snow) was still way overexposed.  I admit this part seems very unlikely given your process (setting exposure based on fg spot meter and adding 6x ND to the top of the frame).  

First thing I would check is that you definitely set the exposure you think you did - i.e. you didn't let the camera choose the exposure (AP, SP, P etc.) with the filter in place.  I definitely am not questioning your ability or knowledge - but hey, we all occasionally forget to flip a switch.  Make sure you're on manual exposure mode.

Second theory:  you're suffering from bad placement of the filter combined with spot metering from a dark spot (darker than mid gray) in the fg.  I don't know if you already do this, but when placing a grad filter, it is a good idea to use your depth of field preview button to look through the lens while sliding the filter slowly downward in its holder so you can see the effect in the frame.  Johnathan's comment about the small aperture has been pointed out as mistaken (or a joke  but one thing that is certain is that you cannot tell accurately where the transition will occur in the frame when exposed at F/16 if you look through the lens at F/2.0 - or worse - don't look through the lens at all.  I usually have to slide the filter out and back down a couple of times before I'm satisfied it is well placed.

If you spot metered off a dark place in the fg and then placed the 6X (total) ND too low, you could potentially still overexpose the peaks while with the filter too far down it would obviously negate your proper metering of the fg and render it very dark.

Hope something in here is of help.

Cheers,
Bruce
Title: using graduated nd filters correctly
Post by: Tim Gray on November 03, 2004, 04:25:52 PM
Quote
You have captured remarkable and powerful images (http://www.thelightandtheland.com/pages/Portfolios/iceland/iceland_15_b.html).
At the risk of beating a dead horse...

This image is truly first class - but there are circumstances where the contrast in a shot like this is too much to handle.  No neutral density filter would help this kind of composition.
Title: using graduated nd filters correctly
Post by: Jonathan Wienke on November 20, 2004, 06:04:44 AM
I've never used a split ND, and have no interest in getting one, for the simple reason that most of the high-contrast subjects I shoot are irregularly shaped, and a filter with a straight-line graduation would be useless. For me it's usually sunsets with mountains in the background rather than slot canyons, but the same concept applies.
Title: using graduated nd filters correctly
Post by: Ray on November 20, 2004, 07:41:02 PM
Quote
I've never used a split ND, and have no interest in getting one....
Same here. Autobracketing and digital blending work fine for me, although I sometimes have a bit of difficulty getting rid of unnatural halos around dark objects like trees against a brighter sky. I assume this is what Bruce was referring to and is why he prefers to use split ND filters.

Perhaps we need an additional tutorial on how to remove unnatural artifacts of digital blending  ???  Adjusting the levels of the individual blended layers can largely overcome the halo problem, but there may well be other adjustments in the process that can help.

Excellent photos, Bruce. Strong compositions with a tendency towards minimalism which I like.
Title: using graduated nd filters correctly
Post by: Ray on November 20, 2004, 09:56:27 PM
Didger,
Thanks for the link. The capacity to blend 3 images is appealing. I'll give it a try.

In the meantime, I'll have to finish the concreting because the only exit from my property until the concrete's cured will be by foot or bicycle  :) .
Title: using graduated nd filters correctly
Post by: Jonathan Wienke on December 04, 2004, 08:25:57 PM
Quote
Quote
It's not a joke.
Hey, I also thought it was a joke.  Who's going to straighten us out here if Jonathan ever leaves?
Well, I don't know that I'm comfortable with the wearing the mantle of The Ultimate Arbiter Of All Truth And WhizDumb, but I do attempt to fact-check before posting.

I was wrong once, in that uber-long thread about the required exposure difference between 4x5 and 8x10 format film...
 
Title: using graduated nd filters correctly
Post by: Brock on December 28, 2004, 05:14:43 PM
November issue of Outdoor Photography or maybe December...had a good article on using split ND filters....and trying to get the square ones like for Cokin so you can move, twist and adjust the location of the split as you need it instead of being stuck with a solid ND or split one on a fixed filter...
Title: using graduated nd filters correctly
Post by: Jonathan Wienke on November 01, 2004, 03:24:31 PM
If you're using an aperture if f/16, you're only using a small portion of the center of the lens and a graduated ND filter isn't going to accomplish much.
Title: using graduated nd filters correctly
Post by: jayz on November 01, 2004, 05:55:17 PM
i actually am using splid nd filter (hard grad?).
i think what i'm having trouble detrmining is what to assign meter value to assign to the reflection.  I assigned it as my midtone, maybe is should be brighter then midtone.
then i added in my grad/splid nd filter for the number of stops over for the sunlit snow.  that was 3 stops over foreground, and added another two stops since i figured the sunlit snow was two stops over grey.
i am not looking to combine over under expousres, shooting film with a nikon f100.
Title: using graduated nd filters correctly
Post by: Peter McLennan on November 01, 2004, 10:13:30 PM
Quote
If you're using an aperture if f/16, you're only using a small portion of the center of the lens and a graduated ND filter isn't going to accomplish much.


That's a joke, right?  Just so we're clear.
Title: using graduated nd filters correctly
Post by: Sam NI on November 03, 2004, 03:07:33 PM
Bruce,

I reckon you should know how to go about this because I've just spent quite some time visiting your site. I think I'll close mine down!  :D  

You have captured remarkable and powerful images (http://www.thelightandtheland.com/pages/Portfolios/iceland/iceland_15_b.html).

I just wanted to let you know how much I appreciated your work.

Good luck for the future,

Sam
Title: using graduated nd filters correctly
Post by: didger on November 19, 2004, 08:38:33 PM
Quote
No neutral density filter would help this kind of composition.
I got my introduction to latitude problems in slot canyons and there the contrast is generally extreme and virtually never fixable with ND filters, so I figure it's best just to do whatever it takes to learn go get good results with bracketing and blending.  There's just too many situations where nature does not oblige with a simple vertical division of light intensity.
Title: using graduated nd filters correctly
Post by: didger on November 20, 2004, 07:55:08 PM
Quote
Perhaps we need an additional tutorial on how to remove unnatural artifacts of digital blending
Well, we have a definite "hard core" group of committed digital blenders, but no certainty about how to get the best results.  I downloaded the demo version of Photomatix, but I don't really know when I'll have a chance to mess with it.  I can't spread myself any thinner with things I'm trying to do for the time being.  There's lots and lots of sample images and details about the program.  It looks vastly more sophisticated than the Miranda blending plug and the demo images look very very good, so how about somebody else doing some serious impartial testing?  I've done my share with lens and tripod research and testing!!   :D
Title: using graduated nd filters correctly
Post by: christoforos on November 22, 2004, 03:23:00 PM
Blending two exposures with digital is really easy, but don't forget that a lot of folks still shoot film (especially medium and larger formats). Sure you can take two exposures with film scan them and blend them in PS, but the cost adds up (film processing and scanning) and the blending of the scans is a little more difficult.
I don't think that separate software for blending is needed just good knowledge of PS.
Title: using graduated nd filters correctly
Post by: didger on December 04, 2004, 08:36:20 PM
Quote
I was wrong once
Well, don't feel bad, I was wrong about something one time too, but it was so long ago I forgot what it was.  Maybe by now it's been proven that I was right after all; probably so; almost certainly.   :D
Title: using graduated nd filters correctly
Post by: [email protected] on December 06, 2004, 09:25:38 AM
Does someone need to "straighten out" John Shaw??? Here's what he says:

"By the way, the more you stop down for your final exposure, the greater the filter's effect. This is particularly true with wide-angle lenses, which are the lenses you'll use most with these filters. If you work with your lens wide open, the filter will have little effect, since it lies far outside the depth of field of the wide-open aperture. Stick with apertures around f/16 for the best results."

See http://www.singh-ray.com/shawarticle.html (http://www.singh-ray.com/shawarticle.html)