Sandman <
m...@sandman.net> wrote:
>
fl...@apaflo.com (Floyd L. Davidson) wrote:
>> Histograms are almost certainly the *most* useful tool
>> brought to the camera by digital technology! It is
>> *vastly* better than any judgment you can make about
>> exposure by looking at the image displayed on the
>> camera's LCD display (rest assured you cannot get even
>> close!).
>
>I'll take your word for it! :)
Don't take my word for it, do some research. An LCD
display will appear different in different ambient
light. And even in the same light the human eye cannot
distinguish absolute brightness levels with any degree
of accuracy. Your eyes will probably be able to tell
you that an image is massively over exposed, or
horrendously lacking in contrast... but you cannot
determine that half a stop more light will produce a
better image, never mind 1/3 of a stop.
With a light meter half a stop is a relatively large
distinction. If your light meter uses an histogram
it is possible to determine exposure within 1/10 of
a stop, though 1/3 is a reasonable level of accuracy
and 1/2 is good enough for most people.
[snipped a lot]
>> But the right edge of the graph should extend all the
>> way to the right edge *if and only if* there are pure
>> white areas in the image. And it should extend all the
>> way to the left if there are black pixels.
>>
>> And that means you can tell if an image is over or under
>> exposed, and with some care it can be determined within
>> about 1/10 of an fstop.
>
>And you're saying that an image that is over/under-exposed by the
>histogram could possibly not be obvious by just looking at the
>picture?
Absolutely.
>I'm guessing that it's for finding hot and cold spots in the
>image rather than a more general "this image is slightly overexposed"
>kind of declaration?
The normal graphical display of an histogram will not
help to find hot and cold spots.
On the other hand, a "highlight display", which is
actually just a different kind of histogram, works
extremely well for that purpose. The area that is
clipping will blink white, and the area that is blocking
to black will blink black.
>Say that the hump in the histogram is 40% to the
>right, and you retake it to move it to only 30% to the right -
>wouldn't such a difference be discernable by just looking at the
>picture and conclude that it's slightly under/over-exposed?
If you can look at the two images side by side you will
be able to see that one is relatively brighter than the
other. If you look at either of them separately 5
minutes later, you probably won't be able to see the
difference, and certainly will not be able to determine
that if it is 30% from the right side of the histogram
then it is probably about 1-2/3 to 2 fstops under
exposed.
>To determine whether you have cold/hot-spots, I could totally
>understand the use of a histogram, but I just can't remember seeing an
>image in post product and it having a problematic amount of cold/hot.
>Sure, it has happened, but not to the degree where I have thought that
>I should have had a way to determine and adjust it when I shoot it.
An example would be taking a picture of a person's face
on a cloudy day, with the sky as the background. There
is little reason to worry about whether the sky is
clipping or not, as there is no detail there anyway.
But you wouldn't want the white parts of the person's
clothing to clip, nor earrings, nor teeth, nor
reflections off the forehead or cheeks.
By using the blinking highlight display that can easily
be differentiated. Or, if one wants to avoid even the
cloudy sky being into clipping, the regular historam graph
would show that very clearly.
Judging that by eye while looking at the image on a very
small LCD screen while shooting will get results scattered
over at least 2 and maybe 3 or 4 fstops.
>Plus, doesn't a lot of DSLR's offer cold/hotspot flashing for images?
>I'm sure I've seen that.
That is another form of histogram.
>> If you make an exposure and the right edge of the graph
>> is about at the middle line along the horizontal axis
>> you know that the exposure could be increased by
>> approximately 2-1/2 stop before highlights would be
>> clipped. And if the graph shows a vertical line on the
>> far left edge, it is very clear that highlights are
>> clipping and exposure must be reduced if clipping is to
>> be avoided.
>
>SO, basically, you'd use it to see, at a glance, what your limits are
>for the current subject? How much of this could easily be dine in post
>production?
Exposure should normally be nailed at the time the
shutter is released.
But it is also absolutely true that histograms are an
essential tool while processing/editing! Adjustments to
brightness and contrast, to saturation, and for
sharpening, should all be done with a very careful eye
on the histogram.
>I.e. if the image looks ok in the LCD after I shot it, and I have
>hot/cold flashing to determine whether something is several
>under/overexposed - sure I could tweak the image within the limits in
>post production? I.e. what do I gain by reshooting it with a new
>timing in-camera at that point?
If it is flashing to indicate clipping by "several
under/overexposure", it will probably make very little
difference what you try while post processing. Clipping
means the data is *gone*.
>Be sure to realize that I'm not actually *questioning* your usage of
>the histogram, I'm not using it myself, so I can't make any other
>comments than my guesses above :)
Yes, it is quite clear that you are entertaining the
concept as a mental exercise, and haven't actually
worked with it! That is of course a good thing to do.
As I an many others who have a lot of experience with
this kind of technical aspect will tell you... you *do*
want to take it a step farther and physically work with
it for at least a while. You very likely will find using
histograms to be very very useful... but it is also
perfectly acceptable to decide they aren't for you. Different
people approach the art of photography in different ways, and
I would like to inspire you to do what comes natural for you,
rather than inhibit you to what works for me! You need to
explore it well enough to understand exactly what it is and what
it is not, and then make a choice that fits you.
Just don't pass it off as unuseful because you aren't
doing it now and don't understand where it would help.
Learn it well, then decide.
>> Do a web search on it, and you can spend hours reading
>> and learning!
>
>I just might do that :)
Keep you busy for the rest of your life...