20
\$\begingroup\$

I have read that longer exposure times lead to increase/improved saturation and that landscape photogs use ND filters to allow for longer exposures to get better saturation.

Is this true and does this technique apply to both film and digital. If true for digital, why? Surely, the sensels receive the same number of photons for a particular EV regardless of the time.

Also, is the quality of increased saturation much different than what you would get from just pumping up saturation in PS?

\$\endgroup\$
2
  • 2
    \$\begingroup\$ I was always under the impression that, for film, slight under exposure increased colour saturation. \$\endgroup\$
    – labnut
    Jan 20, 2011 at 15:49
  • 1
    \$\begingroup\$ @labnut: Depends on whether you're talking about negatives or chromes. Negatives like slight over-exposure; chromes like slight under-exposure. Velvia kind of broke most of us chrome shooters of that habit -- it really got garish shooting at 64 instead of 50. VPS -- the wedding photogrpher's weapon of choice -- was rated at 160, but most folks shot it at 125 or 100. \$\endgroup\$
    – user2719
    Jan 20, 2011 at 17:20

5 Answers 5

9
\$\begingroup\$

What you may have heard is that you should slightly over-expose your digital shots with modern DSLRs when wanting better saturation. Similar to film, where you'd under expose so that you don't "lose data", if you over expose your digital shot, without clipping the highlights, then you'll have a better saturation without noise when you adjust it in post-processing.

The reasoning behind this thought is that it reduces the noise is shadows since digital sensors are better at picking up information in the highlights than they are in the shadows. Because of the heat and electrical interference within the sensor the lowlights and shadows will have random data picked up due to the sensors own errors. This increases with the ISO setting.

If you under-expose a digital shot by 1/3 or 2/3 stop, like you would in film, then you increase your chance of having to deal with noise in the lowlights and shadows during post-processing. This is a result of having to boost values, and therefore the errors already present from the noise, in the dark areas of the composition.

For this reason, many digital professionals will tell you to over-expose by 1/3 or 2/3 stops so that you can then lower the stops in post production, since the sensor picks up more information in the highlights than it can in the lowlights and shadows. This only really works with the newer cameras with 14-bit RAWs, otherwise you ran the risk of clipping the highlights and losing just as much information as you would if you under-exposed the shot. This works because the highlights and actual like are strong enough, and the values are high enough, that it overrides and electrical and heat interference on the sensor. But again, the results are going to vary based on the ISO and exposure length used.

\$\endgroup\$
3
  • \$\begingroup\$ 32-bit raw? Which cameras are those? Current Canons for example are 14-bit at most, I think. \$\endgroup\$
    – Conor Boyd
    Jan 20, 2011 at 19:50
  • \$\begingroup\$ I meant 42-bit, since it 14 per primary, but I can understand that that will make things more confusing, even without my typo. \$\endgroup\$ Jan 20, 2011 at 20:00
  • \$\begingroup\$ Overexposing greatly helps with noise, but it wont improve saturation, in fact it's the opposite of what you should be doing as it can lead to one colour channel clipping (even if you have a 14bit ADC) which is very bad for colour reproduction! \$\endgroup\$
    – Matt Grum
    Jan 21, 2011 at 22:41
14
\$\begingroup\$

Firstly I can't think of any reason a longer (sunlight) exposure would affect saturation nor have I heard of this idea before.

There's a couple of things that might be causing confusion here:

  • Longer exposures can cause individual pixels to be saturated, in the sense that they can't store any additional charge. This is a totally different use of the word however.

  • Under artificial light a longer exposure can result in better colours as the spectrum emitted by incandescent and fluorescent bulbs varies over a 50Hz or 100Hz cycle and so you want to capture several full cycles. You wouldn't normally need an ND filter indoors though!

  • Landscape photographers make regular use of ND filters to get long exposures to blur water or other moving objects.

If you want really saturated colours you need to underexpose slightly. This is because with the RGB model as colours get more saturated one channel might hit the maximum possible value (and start clipping) before the others even though the overall tone is jot that bright. Also the colour gamut (the range of possible colours) is at its highest in the midtones (middle brightnesses) as you get brighter the number of colours, and their saturation decreases until you hit peak brightness, where there's only one colour (white) which has zero saturation!

So there are cases when you might use an ND to prevent overexpsoure (and thus desaturated colours) but it's got nothing to do with exposure time (shutter speed).

\$\endgroup\$
9
  • \$\begingroup\$ I'll try and find the original source. Useful info though, about underexposing a bit so not to clip the dominant channel. \$\endgroup\$
    – Ken
    Jan 20, 2011 at 13:01
  • \$\begingroup\$ In fact, if anything, you'd want a shorter exposure time to underexpose your shot (assuming you're not using a ND filter like Matt says)/ \$\endgroup\$ Jan 20, 2011 at 15:43
  • \$\begingroup\$ Isn't this a tip for shooting slide film too? Maybe the same mechanism is in effect. \$\endgroup\$
    – gerikson
    Jan 20, 2011 at 15:45
  • 1
    \$\begingroup\$ There might have been a reliance on whatever the electronic equivalent of reciprocity failure is on a particular sensor involved here as well (as I recall, it was all Kodak back in the Before Time), leading to a cargo cult usage over time. \$\endgroup\$
    – user2719
    Jan 21, 2011 at 22:29
  • \$\begingroup\$ @stan you might be right but unless the issue is the shutter using a longer exposure + ND filter shouldn't make any difference as it's the same amount of light coming into the camera \$\endgroup\$
    – Matt Grum
    Jan 21, 2011 at 22:39
4
\$\begingroup\$

I don't know anything about this, I use ND filters for Waterfall shots and in the sun to get my aperture to f/8, f/11 where the lens is more sharp.

I didn't notice any saturation improvement. Maybe this rumor started from photographers that use cheap filters that will alter the color a little bit. I know this because in the past those ND filters where my only option :)

\$\endgroup\$
2
\$\begingroup\$

I always understood it to be that it was down to accuracy in timings of the mechanical shutter: If you consider that, in most modern cameras, a shutter consists of two moving curtains, there will be a time (albeit really quite short, but not perfectly the same every time) for how long it takes to release the shutter, and have each curtain start to move. For the sake of argument, let's say the random error introduced by this mechanical error is +/- 10 microseconds; If you then consider that at a really fast shutter speed of 1/8000 sec (so 1/8 ms or 125 microseconds) then you random error can account for being out by 8%. If you introduce a neutral density filter that takes the exposure down by 4 stops to give an exposure of 1/500 sec, then that 10 microseconds error is all of a sudden only 0.5% of the exposure time. A more accurate exposure time, is less likely to differ to the metered exposure and be over (or under) exposed

\$\endgroup\$
4
  • \$\begingroup\$ This becomes very important when shooting timelapses as the random variations cause "shutter flicker" however I wouldn't have thought it to be a problem shooting stills as you can reshoot if you get exposure problems, and seeing as the cameras metering won't be as accurate as 8% you'll end up taking several photos anyway! \$\endgroup\$
    – Matt Grum
    Jan 20, 2011 at 14:51
  • 1
    \$\begingroup\$ @Matt more significant in the days of film when you didn't get a histogram straight away, and some habits (and advice) die hard. I would also hope that timing reliability is improving over time. Of course the 10 microseconds was made up on the spot (like 85.4% of all statistics) and not based on empirical data (as I don't have access to it) \$\endgroup\$ Jan 20, 2011 at 18:50
  • \$\begingroup\$ A nanosecond is 1x10^-9 seconds. You're looking for microseconds, 1x10^-6. 1/8 ms is 125 microseconds. \$\endgroup\$
    – Evan Krall
    Jan 20, 2011 at 20:08
  • \$\begingroup\$ @Evan oops. My bad -- fixed :) \$\endgroup\$ Jan 20, 2011 at 20:20
2
\$\begingroup\$

With regards to digital cameras, the only explanation of this thesis I could guess is when photographing in low light conditions. There you typically have two options: to increase exposure time or to increase the camera sensitivity (say, to ISO 1600).

Although higher sensitivity often allows getting a sharp image when shooting handheld, the colors and dynamic range of the photo unavoidably suffer. A long exposure photo taken at ISO 200 in the same conditions typically turns out much nicer -- with reacher colors, smoother tones and less noise. It does require a tripod, however. :)

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.