Exposure compensation

Hi
Can someone explain to me about the need to reduce the aperture with a
digital camera to give more 'punch'. Also how does this relate to Exposure
compensation and Exposure value (if at all)
Cheers and thanks to you all

 

I see a complex possibility of answers because the question is vague, but no
worries. When I see a question regarding aperture and an adjective like
punch, sharpness, crisp, and so-forth I can only think of resolution.

So, for best resolution, you want a _modest_ aperture, typically a stop or
two down from maximum (the smallest number, widest aperture). That is a
general statement. Stopping down all the way increases depth-of-field, but
has less sharpness due to diffraction.

You also need to be aware of the shutter speed being used. Faster is always
better... up to say, 1/250 where faster than that doesn't help most
situations for sharpness. (The human body moves, has a heartbeat, and we
aren't aware of how much we move even when holding still. Maintaining a
steady grip, smooth shutter release is always important.

If you means something else, then it might be about color saturation, or
fidelity. Let us know what track you are on. (Another typical error is
where people don't do proper white-balance, but that's another subject.)

 

For exposure compensation, try shadows and highlights under Adjustments.

Digital sensors have limited range, but worse is that they blow highlights
easily. The remedy to use the above is a work-around that many appreciate.

 

Most of the reading I have done concludes that digital sensors have a
*larger* dynamic range than most film.

 

Keep reading.

 

Good advice, but it won't change the fact ... ;-)

Bart

 

Bart, my experience is limited to Olympus - the 8mb model before the EVOLT
(can't remember which it was, but we have four of them) and Canon XL-1
digitalvideo and in both cases, blown highlights or alternatively severely
compromised shadows are a big problem. I've no similar issues with MF or LF.

 

What exactly you refer to with "punch" and "to _reduce_ the aperture"?

Timo Autiokari

 

One surely can find plenty of such conclusions from the Web but they are
either biased conclusions or written without better knowledge.

Film, both color-reversal and negative have about 10 f/stops dynamic
range, they can capture that much of scene range. Can be seen from the
film specifications. The thing is that it is not very easy to get all
that image data from the film (a very high quality scanner is needed)
and then the data has to be color-managed properly. Color-reversal as
projected on the white-screen (or on a light-table) only shows some 5 to
6 f/stops in such way that is proper for the vision. Similarly a print
that is made from negative only shows some 5 to 6 f/stops properly. But
this does not change the fact that the dynamic range of the film is
about 10 f/stops.

I've just measured the dynamic range of Canon 1D MK2 and Canon D60
dSLRs. They both have about the same dynamic range, about 8 and 2/3
f/stops. Sadly, all digital camera manufacturers conceal this very
important specification. On the Web there are some reports of dynamic
range measurements of digital cameras but the results are not
consistent, they vary a lot. Partly this is because of some of the
measurement/evaluation techniques are prone to fail considerably, partly
because there is no single good standardized method available, and in
some cases personal preferences seem to affect a lot. I have found only
one conclusion that claims over 10 f/stops range for a digital camera,
all the others conclude it it be around 9/stops.

Timo Autiokari

 

Correct me if I'm wrong, but it seems to me that you are referring to
the in camera JPEG results, which is somewhat like judging film
capability by looking at a one hour drugstore print. Lots of
information got lost.

Most decent current digicams can produce a 10 - 11 stops of dynamic
range (and 1 stop of various types of noise) from its 12 bit internal
processing, but you'll need to use Raw camera data.

Film is very much limited by its grain structure which raises the
noisefloor and limits the signal to noise ratio to something like 6 -
10 stops depending on whether you use slide or negative film and use
low ISO.

This has all been, and no doubt will be, discussed and demonstrated on
more appropriate newsgroups than this one.

Bart

 

This fault, that in-camera finalized JPEG shows lesser dynamic range
than what can be had from the RAW, is particularly a fault in Canon
cameras, all Canon models have this fault.

Do you suggest that this fault is similarly found from all the other
digital cameras also?

Or do you you suggest that there would be some kind of limitation with
the JPEG image format that would limit the dynamic range that it is able
to show to some lesser quantity compared to what can be had from the RAW?

While it is possible in theory (like on the Excel sheet), in the real
life they do not. They do approach, with the aid of severe signal
processing that produce quality issues like plastic apperance, loss of
sharpness etc, about 9 stops.

Timo Autiokari

 

9 Stops, or 2^8, or 256 levels.

 

I have not personally tested the dynamic range of JPEG output of all
camera's but there seem to be a lot of unquantified user remarks
pointing in that direction. Poor exposure on Auto may also play a role
in that.

No, not necessarily so because with a proper tone curve all 12-bit raw
data can be mapped in the 8-b/ch JPEG. It can be more of an issue to
linearize the tone curve in order to reconstruct the linear
relationship between luminance and signal level and the derived noise
level which sets the floor for Dynamic range measurements. JPEGs are a
more troublesome medium for DR quantification.

That would depend on the (absence of) noise reduction and average
noise floor level, but it is not difficult to exceed those 9 stops if
a bit more noise is tolerated, as I've demonstrated (in a 'worst' case
scenario of tungsten lighting and lots of added lens flare, and a raw
converter that clips the noise below 127 out of 4096) in a more
appropriate newsgroup:
<http://www.xs4all.nl/~bvdwolf/main/foto/Imatest/1dsm2/1436_Step_2.png>
..

Bart