A month or so back I asked if there was a noticeable reciprocity

[flamin lib]

departure for digital sensors. Apparently there is. Reciprocity is defined as the media responding to light intensity in the expected way, ie if light intensity decreases by 1/2 the exposure should increase by 2 to compensate.

It appears that most digital sensors lose sensitivity quickly as the light intensity falls. In ordinary situations it can be exhibited as excessive contrast like the WJS post here: http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=280x41043

In the film days Kodachrome was known for its high contrast. In particularly contrasty light such as high altitude clear sky conditions the shadows simply disappear. That seems to be the way digital behaves if the light is in the least bit contrasty. What this means is that the sensor is not as sensitive at low light levels (shadows) as film normally is.

In practicality what this means is that in low light conditions exposures will be longer than one expects.

The contrast in pictures like WJS can be minimized to a point in post-processing but if carried too far the image will exhibit noise and artifacts (funny looking clusters of pixels).

This is more than you wanted to know but it affects my shooting a lot as the best wildlife shooting is early morning and late evening near sunset/rise. ISO 100 in full daylight behaves more like ISO 25 late in the evening.

[WannaJumpMyScooter]

1. O crap... now I see what you all see

I had been looking at this pic all day on a laptop, and on that screen, it looked fine. Now I am home and I see what you mean.

No, I think there is more detail in that shot, and I am going to find out.

Here it is with minimal post...

[mth44sc]

3. Looks like you been henbit

aren't the wild flowers down this way just somethin'!

[DS1]

2. larger sensors help

each pixel takes in more light

[flamin lib]

4. I'm not sure about that. My college instructor has Canon (15x22 mm sensor) 8 mp

and his pictures show the same contrast. He corrects it in-camera because he doesn't like to post process and tends to shoot it right in-camera.

I also have a Nikon Coolpix with a sensor about 3x4mm 8 mp and it isn't visually any more contrasty than my Oly with a 13x17mm 5 mp sensor.

No, I think it has to do with energy density not total energy absorbed.

Keep in mind that this whole thing, OP included, is anecdotal with no science behind it.

edit to add mp count.

[DS1]

5. I was talking more about the low light performance/ noise with larger sensors

http://www.clarkvision.com/imagedetail/does.pixel.size.matter2/

[flamin lib]

7. Very interesting article. I'd have to do a lot of research to compare

sensor to sensor to see the actual relevance of the chart tho. The Canon S70 uses a CCD sensor and the D1 a CMOS--like comparing Kodak 400 negative film to Ectachrome 64. I'll see how much time I have to weed out the different sensor types 'cause I'm a curious person who wants to know.

[F.Gordon]

6. What Teach' is talking about......

In the olden days when people shoved these smelly rolls of paperish stuff in their camera there were certain types of this smelly paperish stuff (film) that suffered from a problem called Reciprocity Failure.

What does this fancy phrase mean? Well... basically... it means that at some point when your exposure time is increased the film would go all wacky-like and would not react to low light in the manner that was expected using any number of different kinds of exposure "formulas". For example... if your initial exposure was f/4 at 1/250 second.. in theory... at f/8 your exposure should be about 1/125 second.

The obvious result of this fancy Reciprocity Failure problem is that colors get all gooey and melt together.

WJS's photo is being used as an example of how DIGITAL sensors (the new "film") have this same problem. Well, I say Nay Nay. I've never seen proof of this in Digital.

And since I'm the guy who.... "fucks up so you don't have to"
:dunce:

I snap quite often during the WORST possible time of day cause' I'm not a morning person and I hate "waiting" for the light at sunset. This is the time of day when the light is EXTREMELY contrasty (extreme differences between bright light and low light). I will lose detail in areas where very bright highlights mix with the shadows but I've never seen Reciprocity Failure.



I don't know what went on with the WJS photo. I thought it was a processing thing he just used. IF WJS would be willing to post the EXIF maybe someone can figure out this "mystery".

:hi:

[flamin lib]

8. Thanks for weighing in F. Gordon.

Color film was particularly susceptible to RD at both ends of the exposure speed range. Longer than 30 seconds and shorter than 1/10,000th (automatic flash). Fun Fact: a beam of light leaving an automatic flash at it's closest setting is about 19 miles long. Some film would record bleached blond hair as green using the very short flash exposures causing some discomfiture with wedding candids. B&W film just showed a lack of density at long exposures--over 30 seconds.

In the days of Ansel's zone system and density curves it was easy to see the effects of RD, but with digital sensors there isn't an equivalent of a density curve. There is a lot of discussion of "dynamic range" which as I read it is equivalent to contrast or the range of brightness a sensor can record. It appears the consensus is that digital sensors have less dynamic range than film which is equivalent to a very steep density curve. So extrapolating from this, digital sensors are less sensitive to lower light levels, hence the shadows drop out and need help in post processing.

For me it's a gut feeling. I spent many hours in a canoe on the swamps of Caddo Lake shooting K64 through a 300mm 4.5 at 1/30 to 1/60 in deep shade. Under the same conditions I'm getting 1/8 at ISO 100. In full daylight I get exposures I'd expect, i.e. f16 @ 1/125. At lower light levels the ISO 100 behaves more like ISO 25, hence my belief that digital suffers RD at a much shorter exposure. While I've not seen any color shifts like those in color film, I am experienceing a much longer exposure than I would expect with film in subdued light.


K64


K64

It's all moot because I'll shoot what I have and not go back to film with all its time, labor and expense. It's just that I'm a bit of a technician and can't let arcane academic stuff go.

In the dictionary next to anal retentive you will find my portrait . . .