I am somewhat of a newbie and I have been seeing a lot of D7000 comparisons using a 100% crop image or at least referring to it. Could someone please explain technically what that is and why so many people use it to show quality of the D7000 images.
When posting at Nikonians, it's not really possible to get an idea of the fine detail recorded in a file by re-sizing the whole image to fit within our 900 pixels / 150K limits. In this context, a "100% crop" is a section of up to 900x900 pixels cropped from the image, which can then be posted and viewed full-size.
As to why so many use 100% to show quality, personally I think forum-dwellers are over-reliant on 100% and it is a little overdone. I don't usually look at my own photos at quite that magnification (it works out to several feet wide and feet tall). So I don't think looking at this level in isolation fully helps. The folks who show the whole image plus a 100% crop of a section of the image are the most helpful. To look at someone else's images at just that level of magnification, not too sure how helpful that is.
Advice I have received here is looking at one's own images at 50% (try the 50%/100% slider in View NX2 free Nikon software) is a more realistic way to evaluate an image.
And there is no substitute to actually downloading a full-size image from Nikon or one of the test sites and processing it in your own post-processing software.
We joke about pixel-peeping and measurebating on the one hand and then jump to 100% on the other hand
While I'm on my soapbox I'll add, whenever a new camera comes out everyone demands to see images straight out of camera with no sharpening or post-processing. Once again, is that how I ultimately view and display my own images - never!
The reason people show 100% crop is because that is where the critical differences between cameras are. Looking at a photo that has been resized for the forums is like looking at a 4x6 print. Pretty much any DSLR can do a good job of making a 4x6 print. It is when you start targeting large prints that the differences really become apparent. Since we can't display a large print on the Web, we crop out part of the digital file to show what one small piece of the large print would look like.
I have always wondered what is the actual full size of the image at 100% crop? Does anyone have an idea? I guess it would be really large much larger that what we would usually print 16 x 20, 20 x 24 etc.
>For example, a D700 produces images of 4,288 x 2,848 pixels >(when in ), at at 300 pixels/inch resolution
And to view this image at 100% on a monitor would require a monitor with a screen with 4,288 x 2,848 pixels. That's roughly six times the size of today's monitors. In other words, you'd need two rows of three monitors stacked on top of each other. That's around 5 feet wide and 3 feet tall of screen area. It is unusual to be staring at an image that large from only 2 feet away (an example distance of eyes to monitor) - it demands a circle of confusion of about 1/3 the diameter, or put another way, reduces acceptable depth of field to about 1/3 what one would calculate on a usual DOF calculator.
So for sure, viewing at 100% is unusually demanding. It is like requiring a car speedometer with two decimal digits of accuracy.
Actually it would be more proper to state that the D700 has a sample rate density of about 3000 SPI giving its full resolution images a pixel dimension of 4256x2832. I say about 3000 for the vertical sample rate is slightly less than the horizontal rate. I seem to recall reading somewhere that Nikon’s senor sensel’s are actually rectangular and not square, but don’t recall where I read that.
The PPI tag in the EXIF data just signifies to a printer driver or RIP what physical medium size you wish to print to.
Peter’s comment is the general argument in regards to printing larger. It goes that when you print larger the normal viewing distance is greater so you do not need the level of detail at that distance as you would viewing closer. Unfortunately you cannot always control how people will view your large print, and a number of people that do gallery exhibitions will tell you that people will walk around viewing prints from afar taking in the whole image at once. But if they find an image that peaks their interest, they will then step in and view it at about 20 inches to admire the fine detail. So if you want to insure high quality sharp gallery prints you need about an actual 300 PPI of image resolution.
So while I cede that Peter’s argument is a valid one, I do not fully ascribe or agree with it. I have seen people make comments such as “there seems to be motion blur of 2 to 3 pixels”. To me the only way of seeing that is to view the image on screen at 100% or greater. When doing so you’re not judging the general sharpness of the whole image but trying to assess the area of critical focus or the focus plane, of which in an image there is really only one. If you’re view an image at 50% view then one screen pixel represents 4 image pixels and I would think at that point certain areas of critical focus could be altered appearing sharp when in fact they are not. And as to whether any anomalies discovered at 100% view will be apparent when making a print would be up to the viewer and certainly a moot point. Even viewing an image on screen as print size is only an indicator. Actually when applying sharpening for inkjet prints it is said that you need to make the screen image look over-sharpened to achieve the proper amount of print sharpening. Thus signifying to me that the screen view is only and indicator and not to be relied upon for judging printed output.
Of course I am only an amateur hack and could be misunderstanding all of this. It’s not like it would be the first time that happened!!
>While I'm on my soapbox I'll add, whenever a new camera comes >out everyone demands to see images straight out of camera with >no sharpening or post-processing. Once again, is that how I >ultimately view and display my own images - never! > >Best regards, SteveK
A long time ago, I bought a new stereo tape deck. When I took it home and plugged it in to my system, it didn't sound anything like what it sounded at the store. But in the bottom of the box, there was a small casette tape marked something like "Demo." It sounded great! I came to find out that the demo tapes used in stores were "juiced" a little... I don't know, magnetic sterroids or something I guess. Anyway, my lesson was learned and I always took one of my own tapes/CDs with me when I got new stuff.
So, I'd rather see a portion of a photo that has not been altered / retouched / post-processed... whatever. I'd like to see the camera's ability, not someone's Photoshop skills. So I can see why people want to see the pre-processed images when evaluating a camera. No, we don't usually end up with them that way but it's what we all start with. : )
#10. "RE: 100% crop" In response to Reply # 1 Mon 08-Nov-10 07:05 PM by HayMower
>When posting at Nikonians, it's not really possible to get an >idea of the fine detail recorded in a file by re-sizing the >whole image to fit within our 900 pixels / 150K limits. In >this context, a "100% crop" is a section of up to >900x900 pixels cropped from the image, which can then be >posted and viewed full-size.
Thanks, Brian. Just to clarify in my mind - a 100% crop would be any cropped image that is not resized. Is that correct?
And for it to be posted at Nikonians, it must be 900 pixels or less and 150K or less?
There are valid reasons for looking at the image at 100% zoom. It's far easier to see how sharpening is being applied, for example to see sharpening halos, when zoomed in like that. If you can't see them at 100%, you sure won't see them in a print of any size. Likewise with noise reduction - you want to see where you're going to determine if your NR parameters are appropriate. Similarly if you're looking for things like motion artifacts or merge/alignment issues with panorama, HDR or image stacking software, 100% is the way to do that.
> inkjet prints ... oversharpened
True, because there is very little bleeding from pixel to pixel on the screen, especially LCD screens. But on real paper, particularly some surfaces, adjacent ink droplets merge together and actually soften the perceived sharpness of the image. Accordingly sometimes it is appropriate to sharpen a file to visibly distracting levels knowing that the printing process will counteract that.
Your monitor is about 75 dpi, so when you're viewing at 100%, a 4200 x 2800 image works out to be about 50 x 37 inches - about the width of a generous front door on the house, but a bit shorter.
_____ Brian... a bicoastal Nikonian and Team Member
My gallery is online. Comments and critique welcomed any time!
#13. "RE: 100% crop" In response to Reply # 0 Mon 08-Nov-10 11:22 PM by ZoneV
On the web, all or nearly all images are displayed at 100% magnification. In order to see an image online without resizing, it is necessary to crop it so it fits the monitor. Viewing an image without resizing gives a better approximation of the image's native quality. However, when making small prints or viewing whole images (resized) on a monitor, some of the resolution evident in the 100% view is discarded (same principle as viewing it from a further distance, except that this is the digital analog to viewing distance). So something that appears yucky at 100% view may look fine at 8x10 inches (uncropped; resized downward and hopefully sharpened for outupt).
As for people who don't belive in ever resizing images, they are either stubborn, lazy, or don't understand that all images have to be resampled prior to printing. Whether they do it in software or not, the output device will make certain it's at the correct size prior to printing.
And as far as pixels per inch are concerned, it's really nothing more than a conversion factor (from pixels to the inches on a print). There really is no such thing as a "300 ppi file" for example. And people who think they are protecting their images by decreasing the ppi from 300 to 72 for web use might be surprised to find out that they are not really changing the data in the file (unless they resample it downward too). ppi is a conversion factor and nothing more.
An undeniable paradox: To think that there is any such thing as an absolute rule is at worst naïve, and at best, shortsighted. There is no such thing as an always-true, all context- or situation-salient, absolute rule that always holds true…including this one!
Pete: you make good points and I agree. I was not commenting at all on printed images. Just pointing out the extraordinary demand on photography equipment when viewing something at 100% on a screen. I wonder how 100% viewing on a screen compares to using a loop on a slide? I would guess the loop is not as high magnification. Peter