Do you recompose in PP?

I often crop a bit. Even with a 1D3, the AF point frequently is not exactly where I want my subject to be in the final composition. So I frame a bit loose.

BrandonSi wrote in post #12044017
Cropping an image doesn't enlarge it. Two different things.

Cropping an image and reducing your chosen print size accordingly doesn't result in any degradation or increase in blur/etc..

But again, this is all for print. Getting back to the point, simply cropping an image is non-destructive to original IQ (if done correctly).

Simply cropping the image doesn't change the remaining part at all - I agree to that.

It's the real world use of the cropped file (with the subject recorded smaller than it could have been) that brings out the difference.

Forget pixels for the moment, think about the physical size of the subject of interest projected by the lens onto the sensor (or film). The smaller the subject size on the sensor, the more it has to be enlarged to reach the same required print or viewing size on the monitor. For that reason, when the subject is viewed at the same display size, the version coming from a cropped file will be less sharp.

The source reason being that the blur at the edges is the same absolute width, regardless of subject size. By enlarging the small subject image more, you get wider blur for the same final subject size than from the frame-filling subject. In other words, less sharpness.

Bottom line: Whatever the print/view physical size you are using when viewing your cropped image, if the subject filled the frame it would look sharper.

That said, some cropping may be unavoidable at times, but you always pay the sharpness price. The more you crop, the higher the price.


EDIT:: During the evening walk with my labrador, I came up with a simpler explanation.

Again, forget pixels. The problem sources from the image projected by the lens onto whatever screen.

Imagine me taking a head shot of a model with a film camera. In the first picture, the model's head fills the frame. Then I take a second shot from twice the distance and the head fills only half the frame in each direction.

Now I inspect the developed film with a magnifier and estimate that in the first picture the blur width amounts to 1/10 the width of an eyelash. In the second picture, the blur width is the same (being the same lens, at the same aperture), but the eyelash is half thinner, therefore here the blur width amounts to 2/10 the width of the eyelash. So, relative to the image features, the blur in the second picture has doubled. This fact remains, regardless of the way the picture is viewed or displayed. Every feature from picture 2 is less sharp then from picture 1. In picture 2, a small detail will simply disappear in the lens blur.

I hope the fresh air helped.

scubthebub wrote in post #12040519
I typically frame so I don't have to crop, but I have a real hard time keeping a straight horizon for some reason so i tend to loose a little in corners when I straighten. Otherwise I rarely crop, but there is nothing wrong with doing so (to a certain point anyways).

I have the same problem with my 5DII. For the life of me I am perplexed with trying to capture a straight horizon. I do not have the same issue with the 7D.

I will sometimes shoot with the intent of cropping later, it depends on what is surrounding the subject when I shoot. That is a benefit of having a large sensor with a lot of megapixels...

agedbriar wrote in post #12044996
Simply cropping the image doesn't change the remaining part at all - I agree to that.

It's the real world use of the cropped file (with the subject recorded smaller than it could have been) that brings out the difference.

Forget pixels for the moment, think about the physical size of the subject of interest projected by the lens onto the sensor (or film). The smaller the subject size on the sensor, the more it has to be enlarged to reach the same required print or viewing size on the monitor. For that reason, when the subject is viewed at the same display size, the version coming from a cropped file will be less sharp.

The source reason being that the blur at the edges is the same absolute width, regardless of subject size. By enlarging the small subject image more, you get wider blur for the same final subject size than from the frame-filling subject. In other words, less sharpness.

Bottom line: Whatever the print/view physical size you are using when viewing your cropped image, if the subject filled the frame it would look sharper.

That said, some cropping may be unavoidable at times, but you always pay the sharpness price. The more you crop, the higher the price.

EDIT:: During the evening walk with my labrador, I came up with a simpler explanation.

Again, forget pixels. The problem sources from the image projected by the lens on whatever screen.

Imagine me taking a head shot of a model with a film camera. In the first picture, the model's head fills the frame. Then I take a second picture from twice the distance and the head fills only half the frame in each direction.

Now I inspect the developed film with a magnifier and estimate that in the first picture the blur width amounts to 1/10 the width of an eyelash. In the second picture, the blur width is the same, but the eyelash is half thinner, therefore here the blur width amounts to 2/10 the width of the eyelash. So, relative to the image features, the blur in the second picture has doubled. This fact remains, regardles of the way the picture is viewed or displayed. Every feature is less sharp then from picture 1.

I hope the fresh air helped.

I would think that there is no doubt that the more you crop the less latitude you have in big prints.

I don't think for example that people do themselves a favor by shooting totally loose in their composition just thinking "I'll crop it later". But sometimes that can't be helped.

And with me, the shots I get of little critters in the wild so often need to be cropped -- I know that maybe an 8x10 print would come out decently but I don't delude myself into thinking that a big print of it would hang in a gallery.

Also, there are times when I shoot say a person in landscape orientation just because it seems to work at the time but in processing I see it more appropriate in portrait orientation and I crop accordingly. It's hard to say how large I could print a shot like that...it would depend on the image.

tonylong wrote in post #12045501
I don't think for example that people do themselves a favor by shooting totally loose in their composition just thinking "I'll crop it later". But sometimes that can't be helped.

I agree on both points.

It's all about being aware that you will pay a price for that convenience and that a higher MP camera will not reduce that price to zero.