[tuco]

...
What did you meter? 1 sec @ f/2.8?
At f/16 it'd be 30". Make it so.

If you selected 9 multiple shots the camera will shoot nine 30-seconds (270" or 4 minutes and a half) shots and average out the exposure as well, giving you basically a 9X ND filter.

Try it.

A 9X ND filter is ≈ 3 stops if you're talking filter factor instead of density. f16@30 seconds with a 3 stop exposure time increase = 120 seconds. Thanks for the tip!

[Huck Finn]

A 9X ND filter is ≈ 3 stops if you're talking filter factor instead of density. f16@30 seconds with a 3 stop exposure time increase = 120 seconds. Thanks for the tip!

Ok, X*2^9, then.
Actually I've seen some manufacturer using the "nX" expression to mean an "n stops" factor so I just followed suit....

[tuco]

Ok, X*2^9, then.
Actually I've seen some manufacturer using the "nX" expression to mean an "n stops" factor so I just followed suit....

I don't know why filter manufactures just don't be consistent, mark and advertise filters by stops. Some give it in density and some in filter factor. When given by filter factor, for example, how many people can solve for x (stops) in the following equation when they want to know how many stops it is. Gee whiz.

400 = 2^x

If anyone wants to know it is:

ln(400) = ln(2^x) = x * ln(2)

Solving for x:

x = ln(400)/ln(2) ≅ 8.6 stops

Where ln = the natural log.

So my Hoya NDX400 filter is advertised as 9 stops but really it's a little short.

[falconeye]

An ND filter cannot (currently) be implemented by camera firmware.

The workaround discussed here is called stacking. The Pentax firmware supports up to 9x stacking. But unlike HDR, it doesn't do alignment between images.

Stacking can be more useful though when done in post processing using software like PhotoAcute. It can replace a tripod for waterfall shots etc.

However, to divide 5s into 1/50s requires 250 images and isn't everybody's darling


Future CMOS sensors which incorporate a one word memory register with each pixel will be able to support ISO below ISO 0.01.

[philbaum]

Great trick! (the multiexposure) I'm going to use it for waterfall shots next summer! Makes it a lot easier to compose/focus without an ND on.

Huck finn is right on. I've been using multiple exposures instead of ND filters for a year, works great for water effects in broad daylight.. its one of those features that K5 has but not D7000.

[tuco]

Huck finn is right on. I've been using multiple exposures instead of ND filters for a year, works great for water effects in broad daylight.. its one of those features that K5 has but not D7000.

On a sunny day you'd be looking at f16 @1/125 exposure (ruffly). With only 9 shots at 1/125 you are still an order of magnitude short from even a 1 second "long" exposure. I'd say you still need a ND filter for bright conditions if I understand this method correctly.

[Ishpuini]

That's not what I was suggesting. I'm not saying to divide a 10 seconds exposure in 5 2 seconds ones. I'm saying that if you tick the "Auto EV adjust" check box the camera will average out the images, therefore obtaining the same result of an exposure as long as the sum of all the exposures made.

The result will still be multiple exposures interrupted by short periods of non exposure. Of course it depends on what you want to achieve with your 30s exposure. Imagine a cityscape at night where you want to capture the trails of car headlights. Whether the exposures are averaged out of added doesn't really make a difference as the combined light trails will not result in continuous lines either way. For this you really need to use one or more ND filters. If your total exposure does not include one continuous movement or if the interruptions won't be visible in the end result there is no problem with your approach of course.

Another difference is impact of mirror movement and alignment between frames. I found that a very sturdy tripod is required, more than when using a single long exposure with ND. The mirror going up and down between every exposure increases the impact of mirror shake compared to working with an ND, and if you eliminate this by using 2s MLU timer that increases the lag between exposures even more. Mirror movement - even if the impact on individual frames is eliminated with (2s) MLU - may also slightly move the position of the camera on the tripod (if not mounted properly) resulting in significant loss of sharpness because of misalignment between the frames (esp between the first and the last). This may sound massively exaggerated, but I have actually experienced this with the camera seemingly properly mounted on the tripod. If the camera is mounted level the risk is small, but if not you would be surprised how easily gravity can pull ever so slightly on a camera and change its position a fraction, and vibration resulting from mirror movement helps this process along. A misalignment of very few pixels is enough to result in significant loss of sharpness.

BTW this will also have a good effect on any digital noise you may have in dark areas, since the noise is randomly scattered and it will be smoothed over by the multiple exposure.

For the noise that is random this is true. However, some noise is not random, e.g. the noise that is eliminated by DFS. Note BTW, that if you leave Slow shutter speed NR turned on, the interruptions between exposures will always be longer than the exposures themselves once you use 1 second or more.

Actually this works better than a physical ND filter, since this kind of filters (at least the darker ones) often induce color dominants in the picture. To have a truly neutral ND filter you have by a quality, expensive filter.

True, filter quality is paramount. But that goes for every type of filter, actually for every optical element you add to the optical chain between your subject and the sensor. Good quality (expensive) filters are the only way not to compromise your camera's IQ imho, but that's a personal opinion... Good quality ND filters exist and should be preferred imho...

hth, Wim

---

Last edited by Ishpuini; 12-23-2010 at 01:42 AM.

[Huck Finn]

On a sunny day you'd be looking at f16 @1/125 exposure (ruffly). With only 9 shots at 1/125 you are still an order of magnitude short from even a 1 second "long" exposure. I'd say you still need a ND filter for bright conditions if I understand this method correctly.

Aw, right!
It's 9 times the original exposure, not the equivalent of a 9 stop ND filter.
I beg your pardon for the incorrect calculations, you were right, it's a 3 stop ND effect...

[philbaum]

On a sunny day you'd be looking at f16 @1/125 exposure (ruffly). With only 9 shots at 1/125 you are still an order of magnitude short from even a 1 second "long" exposure. I'd say you still need a ND filter for bright conditions if I understand this method correctly.

Tuco,
the last time i did this i merged 4 exposures at 1/6 sec, iso 100, f16. Its not equivalent to the creamy smooth water surface that one finds with a long exposure. Its more like a filmy see-thru effect, so you're right, its not the same thing. More than 4 exposures would have helped.

Looked at some of your BW exp. out on the Oregon beaches, outstanding images!

[006]

the trick with multishot is good, but it's not the same that neutral filter.

for example with neurtal digital filter U coud use flash lamps with constant brightnes with aperture level of Your choice and without changing position of the lamps.



I thought You can always lower the sensivity of any sensor without losing quality - just the software way.. simply by dividing level of output signal coming from each pixel... that could even make quality higher.. due to statistics.. but maybe it's realy impossible, but why ?

sorry 4 my english