High ISO or DOF?

[Fabs Forns]

This is a full frame image, taken with the 500/4 and the 7D, on a very high ISO and quite wide aperture. CAn you tell if the apparent lack of detail is due to the high ISO or is it a matter of DOF?
Please vote and will see what the consensus is :)

Model: Canon EOS 7D
Lens (mm): 500
ISO: 1600
Aperture: 5
Shutter: 1/640

[Alfred Forns]

... great post Fabs !!! ... have been told not to vote till later :)

[Lance Peters]

Hi fabs
can't tell from my phone.
Thought you said it was high iso :)
1600 is my starting point -- ohh I forgot I have a d3s ;)
will have a vote when I can see on big screen

p.s wife is doing well. Couple more days in hospital ;)

[Tom Graham]

My answer (D) -non of the above-
The lack of detail is due primarily to lack of contrast. You need edge contrast to have detail. Here is PS screen grab of the levels histogram. The BIF shows very little contrast.
.
http://www.mediafire.com/imgbnc.php/795d3ea04703cc4b693e5db799a3f0976g.jpg.
.
Was that meant to be a "trick" question? :)
Tom

[Desmond Chan]

My answer (D) -non of the above-
The lack of detail is due primarily to lack of contrast. You need edge contrast to have detail. Here is PS screen grab of the levels histogram. The BIF shows very little contrast.

It seems to me you're saying the bird is under-exposed. Is that right?



Was that meant to be a "trick" question? :) TomYou never know. The moderators have been throwing test after test at us lately :D


I think it's a DOF problem.

[Fabs Forns]

Tom, the histogram shows the color of the bird to be on the dark side and a lot smaller than the sky. There are no mid tones in the image.
How would you suggest to get contrast on the bird?

[arash_hazeghi]

Fabs,I think it's both, part of the near wing that is oof is out of dof and has some motion blur too. the head and beak are in focus and within DOF but they somewhat lack crispness with a bit grainy look, that is coming from high ISO noise. BTW, can't really tell noise in a small size image :) I posted ISO 12,800 Ibis shot a while ago, looked OK at 1024 pix!!!!

Lol, I agree with Lance 1600 is not considered "very high ISO" these days:D

BTW, very good image given the circumstances and camera used!!! Exposure is perfect and nothing is underexposed.
And cool new signature pic!

[Tom Graham]

How to get contrast on bird? Too late now!!!! :)
But since I had it in PS anyway, below is orig and then hit with brightness and contrast - NO sharpening. Does it look any sharper simply from contrast increase?
A secondary problem could be the small jpg size at 100KB. The pixels on BIF look a bit gross. Anyway here is that contrast before/after. Please, please, if I'm waaay of base on this, tell me, this is the "eager to learn" forum after all!!!
.http://http://www.mediafire.com/imgbnc.php/3f848e50bd4b1ea758d5c439eaf421996g.jpghttp://www.mediafire.com/imgbnc.php/3f848e50bd4b1ea758d5c439eaf421996g.jpg
Tom

[Fabs Forns]

Tom, great that you are "eager to learn" :)
Arash nailed the problems. You version is not realistic on a cloudy day, where glossys don'ts show the iridescence for lack of sunshine.

[Tom Graham]

Agree my re-work is not true life, was trying to show how contrast affects apparent sharpness. And do believe your original is like you said underexposed.
What Arash said is all true and the end result are pixels edges that do not have good, if any, contrast with its neighbor. Call me the "edge fanatic" if you like :)
Tom

[Fabs Forns]

Tom, very difficult to reconcile your understanding and mine. I'm a photographer who does PS as a necessity, not a computer savvy who takes pictures. Slight but powerful difference.

[Michael Gerald-Yamasaki]

Fabs,

Greetings. DOF Calculator (Alfred's suggestion) says less than 7 inches at 50 ft. And still only about 15 inches at 75 ft distance. I'm unfamiliar with 500mm at a 1.6 crop regarding framing on an uncertain sized bird ;) but I'm taking a wild guess that you were closer than 100 feet? (as such dof would be a problem, methinks).

Arash has it right...

Tom,


Please, please, if I'm waaay of base on this, tell me, this is the "eager to learn" forum after all!!!

Er, uh, well, histograms do not show contrast or lack of contrast. For instance there is plenty of contrast between the near wing and the far wing. How would you see this in the histogram?

Neither can one contrast an edge from a blur without creating noise (you can't generate signal from noise). Besides, again, the near wing is in good contrast to the far wing, yet the edge still seems oof.

Best to let the optical systems handle focus ;) ,,,

Cheers,

-Michael-

[Fabs Forns]

You guys analyze too much :)
Photography is an art, not a science. To me, anyway.
Measuring to much is detrimental to art, IMO

[Tom Graham]

Thanks Michael, so is there a way in Photoshop to "map" or "evaluate" the contrast of an image? That is, instead of saying "I see it, why can't you", (been there, done that) which may very well be due to monitors, to take a measurement. Turn it into a number. Like if something measures RGB 0,0,0 it -is- black. Is there a way?
Here's my "reasoning" for using PS levels to show contrast. Below I took a small image of a Kingfisher, and did a "levels" look at it. Then I lowered the contrast of the image and took another "levels" look at it. Here are the two images with "level" charts -
.
http://www.mediafire.com/imgbnc.php/72a254ea1089b8308059b2dd91c8b3106g.jpg
.
Clearly seen is that the histogram has been compressed within its possible range of 0-255. To me this demonstrates the change (lack of) of contrast. Is there a better way?

Perhaps I am talking apples-and-oranges here? I agree that for detail/sharpness we are looking not so much at the overall range as shown on the histogram. But at the contrast or luminance change at the pixel-to-pixel level. I haven't tried it but perhaps you could crop a small area of an image where detail is expected and do a histogram "levels' only on that crop?
Eager to learn :) Tom

ps - Sorry Fabs, it's a guy thing :) . You know how we like to take apart something even if it isn't broken!!!

[Fabs Forns]

Well, for me, a pic works or doesn't work and I don't need to measure anything on it.
I go by visual and nothing farther from my mind than techs when I look at one.
It must be a girl thing :)

[Julie Brown]

I have to agree with Fabs on this. Either it works or it doesn't. My initial thought (in my limited experience) was that it was underexposed, but that was not the question. Shouldn't the 7D be able to handle ISO 1600? I am as interested in PP as much as most people, but I just want to know what to do in the field to make better images. :)

[Michael Gerald-Yamasaki]

You guys analyze too much :)
Photography is an art, not a science. To me, anyway.
Measuring to much is detrimental to art, IMO

Fabs,

Greetings. Methinks, science is an art (albeit with a bit more restrictive "rule of thirds" type rules), but in any event it's just my style (or you could think of it as my affliction if you'd like ;) )... BTW, I think just looking at an image is the best way to evaluate contrast (not to mention deciding whether "it works" or not)...

(starting to stray off topic... last comments on this follow)

Tom,

Greetings. Contrast is a spatial metric, that is, a range of levels is only meaningful if associated with the area which those levels appear. The histograms you show are for the whole image so tell us nothing about contrast, say, local to the wing. Therein lies the rub, one can select areas here and there and measure the range of levels across them, but the only contrast information one gets is specific to that area... measuring the wing doesn't tell you anything about contrast between feathers, etc. One can say, these whites have a range of between 220 and 250 which shows reasonable contrast (or conversely only between 250 and 254, so there isn't much contrast - or detail - that can be seen), as "the whites" is shorthand for the "area that is mostly white" giving the spatial bounds of the range of levels.

Probing images for levels information is in my experience easiest with curves (click-hold in the image shows the level in the curves chart with a dot on the curve and numerically displayed as input (below the chart).

IMO, the best way to evaluate contrast is by just looking at an image... the mind is very good at integrating all manner of spatial information from global contrast to pixel level micro-contrast ("sharpness") at a glance ;) .

One further aside... The last images you put up show how setting black point and white point is a contrast adjustment (where compressing them decreases contrast). Levels and Curves adjustments are essentially image-wide contrast adjustments (where excepting for setting black/white point increasing contrast in one area necessarily decreases contrast in another. :eek: ).

It's this last bit that supports the advice of getting it right with the capture rather than trying to fix it later with pp.

Cheers,

-Michael-

[Tom Graham]

Many thanks again Michael. I'm going to carefully review what you say and try it. While I was "composing" the below, your reply was posted. But I'll go ahead and put it up since I think it is more about the the heart of the technical thing. follows-

Allow me one last "appeal" at this. The main reason I'm interested in such (attempt at) technical analysis is to improve communication between us. Sort of like someone saying - The new Canikon lens cost a lot of money. That could mean $300 or $3,000 to them, or to me.

I think the biggest variable or hindrance to our communication about the images here is our monitors. Not to get into monitor technology and calibration stuff, but if you haven't thought much about it, you should. Another recent example at a (not BPN) forum. I pointed out that a large portion of the image looked black. Another commentator said it did not look black to him. And I'm sure it did not look black to him - on his monitor. So, I took the image into PS and PS measured the "black". It was RGB 15, 7, 7. Now we can have a discussion about how black it is, not about what it looks like to me/you.

And let me hasten to add that the technicals, what monitor, how black is it, is contrast low, was ISO low, I agree has little to do with whether we like the image!!!

Tom

[Desmond Chan]

I have to agree with Fabs on this. Either it works or it doesn't. My initial thought (in my limited experience) was that it was underexposed, but that was not the question.

But we were invited to answer some specific questions to begin with, not about if the image works for you or not. And an under-exposed image could give the impression that details are not there or they are insufficient. What Michael and Tom are doing, IMO, is simply providing evidence and reasoning to support their arguments/position. Not a bad thing to do, IMO again.

[Tom Graham]

Well said indeed, Desmond!!! :)
Tom