In the very beginning of my time here at NPN I was devastated because my images weren't sharp enough at some huge size, and I learned then that all they need is to be sharp at actual pixel size, which I think is 100%. But now I want to make sure I'm right about that. I have lots of images that look awesome at even 50% of their size, but not at 100%. I'm deleting them all. I should be deleting them, right? Or is there some point in keeping them? Thanks.
Posted by Luke Ormand on January 28, 2011 at 4:07 PM
what exactly are you deleting? Never ever ever ever ever delete a .RAW image unless you want nothing to do with it ever again.
When you look at an image at 100% it will almost never look good because you are viewing it so close you can see the pixels. This is compounded when you are using a lesser quality monitor. I have a 27" iMac which has double HD resolution so my images look crisp at larger sizes but if I'm right in front of the monitor and look at something at 100% it will still not look good...if I step back a few feet it will look great.
Can you give us some more information? like what monitor you use, if it's calibrated, if you are looking at the images as .jpegs or .RAW etc.? Thanks.
Oh, wow. I think back to all that I have deleted, and double wow.
I thought everything had to be sharp at 100%, so I look at all the JPGs at 100% and delete the ones AND their CR files that aren't reasonably or completely sharp at 100%. Before the Rebel I only had JPGs, and I deleted all the JPGs that weren't reasonably or completely sharp at 100%.
and big sigh.
My external monitor (I use a laptop) is not calibrated. It is a Samsung SyncMaster. It's not widescreen, because for my regular work I need a taller monitor rather than a widescreen monitor. The resolution is set at 1600x900.
So, at what size (percentage) do I delete them if they are not sharp at that size?
Unless an image is clearly out of focus, you should save it. After converting from RAW, making adjustments and resizing, sharpen the image to taste. If you cannot get the image to sharpen, it likely is either out of focus, or has suffered from camera and/or subject movement.
In PhotoShop I zoom into/or out of an image so that when I take a ruler and hold it up to the ruler on the screen, they are equal (or close to it); i.e., actual size rather than actual pixels. I then perform my sharpening. Next, I view actual pixels (100% view) to check for halos, or dust spots I may have missed.
--P
Posted by Lon Overacker on January 28, 2011 at 6:07 PM
Barbara, yes it's disheartening to look at an image file at 100% and seeing an image that is not close to what you would expect in terms of detail and sharpness. I have quite large, 10,000px+ images many over 500MB in size and when I look at these images at 100%, it looks disappointing.
However, trying thinking in terms of output - printer or screen and how these will be viewed. If you've sized an image for a print say 16x20 at 300dpi for a printer output, you can view at 100% as Preston was eluding to, for determining the correct amount of final sharpening. Perhaps there is some amount of experience to know how much to apply for a sharp looking print, but regardless you should be thinking about the output. What I mean is once you have your 16x20 print and you're holding it in your hands how does the sharpness and detail look at arm's length? Remember if you hang a 16x20 print on the wall with matting and a nice frame, you will probably be at least a couple of feet away from the print. Now, if you stuck your nose against the print with a magnifying glass, you actually might see a similar softness or lack of detail that you are seeing on your screen. A sharp image or a sharp print is a relative term. It has to do with final output size and just as importantly viewing distance. For example, a 60foot billboard looks pretty good and sharp when you're driving on the freeway a tenth of a mile away. It's all relative.
So, inspecting an image at 100% does not mean it's not sharp or wasn't focused properly. What you might want to do is experiment. Take a series of images where you manually focus from tack sharp to out of focus and compare the results at 100% so you can start to see and understand the difference.
To a large extent it also depends on how the image will be used. Large prints? Yeah, peep those pixels till your eyes cross. Internet use? If they're sharp at 600-1000 pixels on the long axis, you're golden.
But culling? Omigod. Nuh-uh. I keep them.
Posted by Brian Gatlin on January 28, 2011 at 7:28 PM
Barbara, memory is cheap these days, so there's an easy case to be made for keeping things. And of course, the required level of sharpness depends totally on how you want to use the photo. If you want to show your pics on the computer, if it looks good at 1200 pix, then that is good enough. If you want to print huge prints, it's another story. On the other hand, for sheer file management, it may be best to cull plenty of shots, not just based on sharpness.
Since you're asking about 100%, it might be helpful to post some 100% crops that show what we as individuals feel is acceptable or not. I'll try to find a couple examples when I have time in the next day or two.
Barbara, Unless something is truly horrible, keep it. You can get an inexpensive 2 TB external hard drive for less than $100 to keep nothing but pictures on, and not use up your laptop "space." It will hold many thousands of images.
We never know what processing miracles are coming to salvage the "almost..." shots,, and we certainly all improve in our processing as time goes on. Many shots you think are horrid now may be quite good later on, with improved skills. You'll be surprised.
Sandy
Posted by Steve Canuel on January 28, 2011 at 8:46 PM
Barbara,
Lots of good advice above. I read somewhere that if an image viewed at 50% looks good, it should print out to a decent size. I've used that guide since I started this hobby and its proven true for prints for me at sizes of 8 x 10 and 11x 14. Another thing to think about is paintings. If you look closely at paintings, they're just globs of different colored paint. Step back a bit and they become masterpieces. When you purge your images from a shoot, make a folder called possible trash or something similar. Move the images you're not sure about there and go back some time later (hours, days, weeks) and look at them. If you still think they stink, then get rid of them. Sometimes the "bad" photos become good candidates for photo art projects. In the end, if you like a photo, keep it regardless of how "flawed" you may think it is.
Posted by Richard Ditch on January 28, 2011 at 10:04 PM
Essentially, every raw image needs some sharpening, and the best practice is to apply sharpening only once as the final step in your work flow.
It takes experience to learn how to judge the raw files for sharpness when you are reviewing the images from a session. But you should be able to learn this by looking at lots of you own images and processing many of them to final form.
I always review my raw files after transferring them to my Mac: the first quick cut is make in Nikon ViewNX 2 that lets me see each image filing the screen and also lets me easily view it at 100%. I delete the obviously out of focus stuff, bad exposures, and bad compositions at this stage.
Second time around I look more critically at every image at 100% using Preview - an application that comes pre-installed on every Mac. This is a much slower step than viewing at 100% in ViewNX but shows more detail.
The attached file is a screen grab of a raw file created this morning, viewed at 100% in Preview. Taken with my D200, 300/2.8 lens and 2x converter when the Peach-faced Lovebird landed about 12 feet in front of me. ISO 400. Absolutely no processing here - no exposure adjustment, no saturation boost, no sharpening of any kind. This is a file that I now will look good when I'm finished running it through raw conversion and post processing in photoshop CS3.
I'd suggest that you look at some other raw files from other people on your computer. GO to dpreview and look through some of the recent camera reviews there. I was able to download a sample image taken with a Nikon D7000 with the 18-105 lens at ISO 100, use Abode DNG converter to pump out a .dng version that I could open in PS CS3 with ACR 4.6. Doing this with whatever raw files from whatever camera you are interested in will give you some idea of what to expect from your own raw output.
Thanks Barbara for asking the question and thanks everybody else for the informative answers. This has been a useful discussion thread for me.
Posted by Jim Bullard on January 29, 2011 at 8:56 AM
Sharpening in digital is nothing more that increasing the contrast between adjacent pixels. Don't confuse it with focus.
I have to take exception to the notion that you should sharpen only once just before printing. I know that is the conventional wisdom but while it is true that you need to do a final sharpening for output (either print or screen) you need to sharpen for source (the fact that digital capture is inherently soft), for detail (to enhance details), for effect (to direct the viewer's attention) and lastly for output. It's too complicated to explain in a forum post (whole books have been written about sharpening) but I sharpen in Lightroom and (at the least) I do a two pass sharpen on a separate layer when I pass the image to Photoshop. I also resharpen any time I resize the image.
"...you need to sharpen for source (the fact that digital capture is inherently soft), for detail..."
The same concept may also apply to scans from film. Sometimes I will apply a bit of sharpening at the scan stage; sometimes not. It depends upon how I want the image to look, or if the image needs it.
As Jim and others have said, you will need to sharpen at the specific size and use.
As Jim and others have said, you will need to sharpen at the specific size and use.
On top of that, there can be compelling reasons not to sharpen at all. The graphic artists used by most of our clients insist on no sharpening whatsoever. They know which of the 9,345,499 possible sharpening methods works best for their repro processes, and they want to do it themselves. For that reason we do all our sharpening on a separate layer, preserving the original. If we decide we need to retain a sharpened version for any reason, we do a Save As and label it carefully, then dump the layer from the source file.
On top of that, there can be compelling reasons not to sharpen at all. The graphic artists used by most of our clients insist on no sharpening whatsoever. They know which of the 9,345,499 possible sharpening methods works best for their repro processes, and they want to do it themselves. For that reason we do all our sharpening on a separate layer, preserving the original. If we decide we need to retain a sharpened version for any reason, we do a Save As and label it carefully, then dump the layer from the source file.
Very true, and a very good point to make, Hank. I do all my sharpening on a layer so that I can dump it, if needed.
--P
Posted by Brad Hill on January 29, 2011 at 7:29 PM
Hey Barbara,
There's a critical element to this that's getting lost a little in all the information you're getting that can help simplify things - ask yourself this question: Are you going to use every pixel in the image in your chosen method of displaying it? If the answer to this question is "yes", then you're right in throwing out non-sharp images at 100%. BUT (and this is a HUGE but), these days VERY FEW images are being used in a way where all pixels (and sharpness at 100%) are needed. Let me clarify:
1. Web use (or emailing images): If you have a 12 MP pixel camera, it will have somewhere around 4000 pixels on the long axis. Try to find a computer monitor ANYWHERE that can display 4000 pixels - they don't exist (or are EXCEPTIONALLY rare). The BIGGEST images used on this website are 1200 pixels. Which isn't too far off 25% of the pixels found on that 12 MP pixel camera. Thus, if you're using a full-frame image on this website it has to be downsampled (reduced in resolution) a LOT to get it down to 1200 pixels. AND the act of downsampling an image increases its apparent sharpness. So...if the image can look sharp on your computer monitor at about 25% magnification, then it can look OK on this website. SO, for any form of online display use, you needn't begin with an image that is tack sharp at 100%.
2. Print Use: Most inkjet printers will produce good results at a resolution of about 150 dpi or higher. If we take that same 12 MP camera that has "about" 4000 pixels on the long axis, then that means if you use ALL pixels you could produce a print about 27" wide (4000 divided by 150 = 26.67"). To have that print tack sharp at 27", the original file will have to be quite sharp when viewed at 100%. BUT, how often are you printing at this size? If you're printing at HALF that size, well, you can downsample your image by 50% (which will be equivalent to viewing it at 50% on your monitor), and the act of downsampling will increase the apparent sharpness (i.e., it needn't look tack sharp at 100%). MOST amateurs aren't printing images 27" wide on a regular basis, so normally they don't need images that are tack sharp at 100%.
If you have much higher aspirations than inkjet prints - think National Geographic or Nature's Best Magazine - these images are printed using much higher resolutions (often up to 450 dpi). To have your image printed relatively large (approaching full page) at these resolutions, you're likely to be using ALL or most of the pixels in your image. Which means it better be darned sharp when viewed at 100%.
3. Another issue: CROPPING: If you're going to crop your images a LOT (crop, NOT downsample them), then you very well might have to start with tack sharp images (at 100%). Say you're going to show an image on this website and only want to display the "regular" size version (720 pixels on long axis). If you're using that same 12 MP camera with 4000 pixels on the long axis, you could theoretically crop off 3280 pixels (4000 - 720 = 3280 pixels) and STILL have a tack sharp shot shown on this website, BUT ONLY IF THAT IMAGE WAS SHARP AT 100%. Conversely, if you're not cropping this much (and most folks don't crop that much) you're likely to be downsampling (reducing the resolution of) your image for most uses and, again, this means your image might not need to be tack sharp when viewed at 100%.
At the end of the day, only understanding pixels, the output resolution needed for various display options, and how you're likely to use your own images will help you really determine how sharp they must be (at 100% or any other magnification) to meet your needs.
By the way - never assess image sharpness on your computer screen at magnifications of 33% or 66%. These magnifications don't involve 2:1 pixel reductions and always look a little "odd" on a computer monitor.
Cheers...
Brad
PS: If my images aren't tack sharp at 100%, I chuck them out. If I didn't it would totally jam up my hard-drive (with countless tens of thousands of images which would have very limited use and I'd almost surely never use anyway). My only exception to this rule is if the subject is SO unique (think Sasquatch in desert, or cougar/grizzly bear hybrid) that it will have value regardless of technical quality...
Posted by Scott Fricke on January 29, 2011 at 8:35 PM
In addition to all the other things mentioned above, do not sharpen or do other permanent adjustments to the original file, whether it is a RAW or JPEG, always make a copy and rename something along the lines of "XXXXXXX_Master Edit". this way you can mess it up all you want and still have the original to go back to for another try.
Posted by Lon Overacker on January 29, 2011 at 9:51 PM
PS: If my images aren't tack sharp at 100%, I chuck them out.
This is really the $64,000 question isn't it? What does it mean, or what does it look like, 100% view and being "tack sharp"
Let's take Richard's example above - and thanks Richard for posting the example. An educated guess tells me that with good processing techniques this image will look "tack sharp" at 1200px 72dpi on screen and tack sharp in an 11x14 print (assuming the original pixels can go this size.)
But looking what he posted of the 100% view? Is this "tack" sharp? And how does one determine if it is? Maybe from experience someone might be able to say this is, or isn't. But what about those that are asking the question? What is "tack" sharp? of a raw, unprocessed file?
I would say Richard's image is "tack sharp", but clearly it does appear so in a screen shot.
Lon
Posted by Paul DeCesare on January 30, 2011 at 1:01 AM
Barbara, I have some images that look like dogmeat at 100%, but at 480x720 on the web thse images look great. Don't delete any more images using the 100% View method! You may be killing some good usable web images!
Here's an unscientific suggestion: Set your image to 25%, if it looks good, then the smaller web version will also. If it looks good at 50%, then an 8"x12" print will also. Sometimes I feel like 50% on my monitor matches the quality of my print which is why I stated that. Anyway, many have piped up with good suggestions, so you've come to the right place for advice!
Hank, you're exagerrating, there are only 7,845,692 known sharpening techniques amongst the graphic experts.
I would rather have a good fuzzy shot than a bad sharp shot. Stop deleting shots based on sharpness. That's my advice. Determine the value of a shot with your printer, not by pixel peeping. Remember that sharpness has to be recovered following the a/a filter and cameras don't all handle this the same way. So we all do not get to see the same thing at 100% anyway. Chill. You are making some of us nervous.
I have thought that if a shot looked good small but not larger that I would be "cheating" to pretend that it was a good shot. Obviously (what a surprise! ) I haven't understood at all before now.
Thank you again.
Posted by Brad Hill on January 30, 2011 at 9:34 AM
Barbara,
RE:
I have thought that if a shot looked good small but not larger that I would be "cheating" to pretend that it was a good shot.
I doubt too many would view it as cheating, but if an image only looks good small it definitely limits the options you'd have in using the shot. A good shot at a small size is, well...a good shot at a small size!
Cheers...
Brad
Posted by Ed Cordes on January 30, 2011 at 9:43 AM
Barbara, one topic not addressed here is the possibility if image recovery so you may start again. if your deleted files are merely in your recycle bin just restore them. If they are deleted out of the recycle bin it may be possible to run an image recovery program to recover them. Do an Internet search there are many good recovery programs out there.
Regarding the question of sharpness you have received great advice. I use PK Sharpener which is a highly regarded 3 phase sharpening regime. The "capture sharpening" gives just enough to compensate for the problems of digital capture. The "creative sharpening" is where you selectively or globally sharpen to taste on an otherwise fully processed image. finally, the "output sharpening" is used after the image is re sized for whatever output the image will be used.
My point is that many images that look borderline are rendered excellent with careful processing.
Good luck with your image recovery.
Posted by Richard Ditch on January 30, 2011 at 11:49 AM
I have thought that if a shot looked good small but not larger that I would be "cheating" to pretend that it was a good shot.
Perhaps I gave you that view, and if I did I will not retract it.
I think some of the advice given here is a recipe for mediocrity, and it is a bad strategy to rely on using small versions only because the image is not sharp enough to present at larger size. Do yourself a favor and set high standards early for your own work - you will never regret it later.
However, I do advise not deleting raw image files until you are sure you can read sharpness. You will develop that skill in time, and then you can purge your files to weed out the marginal shots.
Posted by Barbara Graham on January 30, 2011 at 12:09 PM
Thank you, Ed and Richard.
Regarding the other options about using small versions, I do think that if an image is going to be posted here on NPN it should be sharp enough to post a larger version. There are times when compression quality will be so low in a complex image that a larger version can't be posted, but most times people don't say why they don't post a larger version (except the Alaskans and any others who are using dial-up). If it's not posted because it's not sharp enough at even a 1200-pixel size, that seems to me to mean the image is not that great. But it's pretty frustrating to people trying to critique an image to see only the smaller version, and if it's a stunning image I really, really want to see the larger version.
Posted by Richard Wong on January 31, 2011 at 12:03 AM
Sharpness depends on what you plan on using it for. I have had some images that aren't "tack-sharp" that sell for publication because of the type of photo it was. Now I wouldn't make a 20 x 30 print from a landscape that wasn't quite in focus so every photo and use has a minimum threshold of "acceptable sharpness". Over time that is something you will probably figure out.
Also sharpness is subjective. We're all just talking hypothetical here since I really have no idea what your idea of sharp images means because I haven't seen any examples.
A very interesting thread, indeed. I learned a lot from your posts.
I can't say anything better on the matter of sharpness.
About the editing process, as a personal taste, I only delete after the download the pictures that are technically wrong (say out of focus, unwanted blur by non steady camera or moving subject, subject view obstructed by something else, unwanted "mutilation" of the subject, way too much over/under-exposition). And, with the exception of few sure keepers, I let the so-so images rest a little bit after being keyworded.
This Limbo phase may be even months long.
I found that looking with a dispassionate eye my work often helps the judgement and this process needs a "detachment" mood that takes time to happen.
Actually, I am a heartless picture deleter: only about 2-3 % survive!
Discussion
28 repliesWhen you look at an image at 100% it will almost never look good because you are viewing it so close you can see the pixels. This is compounded when you are using a lesser quality monitor. I have a 27" iMac which has double HD resolution so my images look crisp at larger sizes but if I'm right in front of the monitor and look at something at 100% it will still not look good...if I step back a few feet it will look great.
Can you give us some more information? like what monitor you use, if it's calibrated, if you are looking at the images as .jpegs or .RAW etc.? Thanks.
I thought everything had to be sharp at 100%, so I look at all the JPGs at 100% and delete the ones AND their CR files that aren't reasonably or completely sharp at 100%. Before the Rebel I only had JPGs, and I deleted all the JPGs that weren't reasonably or completely sharp at 100%.
My external monitor (I use a laptop) is not calibrated. It is a Samsung SyncMaster. It's not widescreen, because for my regular work I need a taller monitor rather than a widescreen monitor. The resolution is set at 1600x900.
So, at what size (percentage) do I delete them if they are not sharp at that size?
P.S. Thank you very much, Luke, in spite of my
Unless an image is clearly out of focus, you should save it. After converting from RAW, making adjustments and resizing, sharpen the image to taste. If you cannot get the image to sharpen, it likely is either out of focus, or has suffered from camera and/or subject movement.
In PhotoShop I zoom into/or out of an image so that when I take a ruler and hold it up to the ruler on the screen, they are equal (or close to it); i.e., actual size rather than actual pixels. I then perform my sharpening. Next, I view actual pixels (100% view) to check for halos, or dust spots I may have missed.
--P
Barbara, yes it's disheartening to look at an image file at 100% and seeing an image that is not close to what you would expect in terms of detail and sharpness. I have quite large, 10,000px+ images many over 500MB in size and when I look at these images at 100%, it looks disappointing.
However, trying thinking in terms of output - printer or screen and how these will be viewed. If you've sized an image for a print say 16x20 at 300dpi for a printer output, you can view at 100% as Preston was eluding to, for determining the correct amount of final sharpening. Perhaps there is some amount of experience to know how much to apply for a sharp looking print, but regardless you should be thinking about the output. What I mean is once you have your 16x20 print and you're holding it in your hands how does the sharpness and detail look at arm's length? Remember if you hang a 16x20 print on the wall with matting and a nice frame, you will probably be at least a couple of feet away from the print. Now, if you stuck your nose against the print with a magnifying glass, you actually might see a similar softness or lack of detail that you are seeing on your screen. A sharp image or a sharp print is a relative term. It has to do with final output size and just as importantly viewing distance. For example, a 60foot billboard looks pretty good and sharp when you're driving on the freeway a tenth of a mile away. It's all relative.
So, inspecting an image at 100% does not mean it's not sharp or wasn't focused properly. What you might want to do is experiment. Take a series of images where you manually focus from tack sharp to out of focus and compare the results at 100% so you can start to see and understand the difference.
Hope that helps.
Lon
But culling? Omigod. Nuh-uh. I keep them.
Since you're asking about 100%, it might be helpful to post some 100% crops that show what we as individuals feel is acceptable or not. I'll try to find a couple examples when I have time in the next day or two.
We never know what processing miracles are coming to salvage the "almost..." shots,, and we certainly all improve in our processing as time goes on. Many shots you think are horrid now may be quite good later on, with improved skills. You'll be surprised.
Sandy
Lots of good advice above. I read somewhere that if an image viewed at 50% looks good, it should print out to a decent size. I've used that guide since I started this hobby and its proven true for prints for me at sizes of 8 x 10 and 11x 14. Another thing to think about is paintings. If you look closely at paintings, they're just globs of different colored paint. Step back a bit and they become masterpieces. When you purge your images from a shoot, make a folder called possible trash or something similar. Move the images you're not sure about there and go back some time later (hours, days, weeks) and look at them. If you still think they stink, then get rid of them. Sometimes the "bad" photos become good candidates for photo art projects. In the end, if you like a photo, keep it regardless of how "flawed" you may think it is.
It takes experience to learn how to judge the raw files for sharpness when you are reviewing the images from a session. But you should be able to learn this by looking at lots of you own images and processing many of them to final form.
I always review my raw files after transferring them to my Mac: the first quick cut is make in Nikon ViewNX 2 that lets me see each image filing the screen and also lets me easily view it at 100%. I delete the obviously out of focus stuff, bad exposures, and bad compositions at this stage.
Second time around I look more critically at every image at 100% using Preview - an application that comes pre-installed on every Mac. This is a much slower step than viewing at 100% in ViewNX but shows more detail.
The attached file is a screen grab of a raw file created this morning, viewed at 100% in Preview. Taken with my D200, 300/2.8 lens and 2x converter when the Peach-faced Lovebird landed about 12 feet in front of me. ISO 400. Absolutely no processing here - no exposure adjustment, no saturation boost, no sharpening of any kind. This is a file that I now will look good when I'm finished running it through raw conversion and post processing in photoshop CS3.
I'd suggest that you look at some other raw files from other people on your computer. GO to dpreview and look through some of the recent camera reviews there. I was able to download a sample image taken with a Nikon D7000 with the 18-105 lens at ISO 100, use Abode DNG converter to pump out a .dng version that I could open in PS CS3 with ACR 4.6. Doing this with whatever raw files from whatever camera you are interested in will give you some idea of what to expect from your own raw output.
I have to take exception to the notion that you should sharpen only once just before printing. I know that is the conventional wisdom but while it is true that you need to do a final sharpening for output (either print or screen) you need to sharpen for source (the fact that digital capture is inherently soft), for detail (to enhance details), for effect (to direct the viewer's attention) and lastly for output. It's too complicated to explain in a forum post (whole books have been written about sharpening) but I sharpen in Lightroom and (at the least) I do a two pass sharpen on a separate layer when I pass the image to Photoshop. I also resharpen any time I resize the image.
The same concept may also apply to scans from film. Sometimes I will apply a bit of sharpening at the scan stage; sometimes not. It depends upon how I want the image to look, or if the image needs it.
As Jim and others have said, you will need to sharpen at the specific size and use.
--P
On top of that, there can be compelling reasons not to sharpen at all. The graphic artists used by most of our clients insist on no sharpening whatsoever. They know which of the 9,345,499 possible sharpening methods works best for their repro processes, and they want to do it themselves. For that reason we do all our sharpening on a separate layer, preserving the original. If we decide we need to retain a sharpened version for any reason, we do a Save As and label it carefully, then dump the layer from the source file.
Very true, and a very good point to make, Hank. I do all my sharpening on a layer so that I can dump it, if needed.
--P
There's a critical element to this that's getting lost a little in all the information you're getting that can help simplify things - ask yourself this question: Are you going to use every pixel in the image in your chosen method of displaying it? If the answer to this question is "yes", then you're right in throwing out non-sharp images at 100%. BUT (and this is a HUGE but), these days VERY FEW images are being used in a way where all pixels (and sharpness at 100%) are needed. Let me clarify:
1. Web use (or emailing images): If you have a 12 MP pixel camera, it will have somewhere around 4000 pixels on the long axis. Try to find a computer monitor ANYWHERE that can display 4000 pixels - they don't exist (or are EXCEPTIONALLY rare). The BIGGEST images used on this website are 1200 pixels. Which isn't too far off 25% of the pixels found on that 12 MP pixel camera. Thus, if you're using a full-frame image on this website it has to be downsampled (reduced in resolution) a LOT to get it down to 1200 pixels. AND the act of downsampling an image increases its apparent sharpness. So...if the image can look sharp on your computer monitor at about 25% magnification, then it can look OK on this website. SO, for any form of online display use, you needn't begin with an image that is tack sharp at 100%.
2. Print Use: Most inkjet printers will produce good results at a resolution of about 150 dpi or higher. If we take that same 12 MP camera that has "about" 4000 pixels on the long axis, then that means if you use ALL pixels you could produce a print about 27" wide (4000 divided by 150 = 26.67"). To have that print tack sharp at 27", the original file will have to be quite sharp when viewed at 100%. BUT, how often are you printing at this size? If you're printing at HALF that size, well, you can downsample your image by 50% (which will be equivalent to viewing it at 50% on your monitor), and the act of downsampling will increase the apparent sharpness (i.e., it needn't look tack sharp at 100%). MOST amateurs aren't printing images 27" wide on a regular basis, so normally they don't need images that are tack sharp at 100%.
If you have much higher aspirations than inkjet prints - think National Geographic or Nature's Best Magazine - these images are printed using much higher resolutions (often up to 450 dpi). To have your image printed relatively large (approaching full page) at these resolutions, you're likely to be using ALL or most of the pixels in your image. Which means it better be darned sharp when viewed at 100%.
3. Another issue: CROPPING: If you're going to crop your images a LOT (crop, NOT downsample them), then you very well might have to start with tack sharp images (at 100%). Say you're going to show an image on this website and only want to display the "regular" size version (720 pixels on long axis). If you're using that same 12 MP camera with 4000 pixels on the long axis, you could theoretically crop off 3280 pixels (4000 - 720 = 3280 pixels) and STILL have a tack sharp shot shown on this website, BUT ONLY IF THAT IMAGE WAS SHARP AT 100%. Conversely, if you're not cropping this much (and most folks don't crop that much) you're likely to be downsampling (reducing the resolution of) your image for most uses and, again, this means your image might not need to be tack sharp when viewed at 100%.
At the end of the day, only understanding pixels, the output resolution needed for various display options, and how you're likely to use your own images will help you really determine how sharp they must be (at 100% or any other magnification) to meet your needs.
By the way - never assess image sharpness on your computer screen at magnifications of 33% or 66%. These magnifications don't involve 2:1 pixel reductions and always look a little "odd" on a computer monitor.
Cheers...
Brad
PS: If my images aren't tack sharp at 100%, I chuck them out. If I didn't it would totally jam up my hard-drive (with countless tens of thousands of images which would have very limited use and I'd almost surely never use anyway). My only exception to this rule is if the subject is SO unique (think Sasquatch in desert, or cougar/grizzly bear hybrid) that it will have value regardless of technical quality...
This is really the $64,000 question isn't it? What does it mean, or what does it look like, 100% view and being "tack sharp"
Let's take Richard's example above - and thanks Richard for posting the example. An educated guess tells me that with good processing techniques this image will look "tack sharp" at 1200px 72dpi on screen and tack sharp in an 11x14 print (assuming the original pixels can go this size.)
But looking what he posted of the 100% view? Is this "tack" sharp? And how does one determine if it is? Maybe from experience someone might be able to say this is, or isn't. But what about those that are asking the question? What is "tack" sharp? of a raw, unprocessed file?
I would say Richard's image is "tack sharp", but clearly it does appear so in a screen shot.
Lon
Here's an unscientific suggestion: Set your image to 25%, if it looks good, then the smaller web version will also. If it looks good at 50%, then an 8"x12" print will also. Sometimes I feel like 50% on my monitor matches the quality of my print which is why I stated that. Anyway, many have piped up with good suggestions, so you've come to the right place for advice!
Hank, you're exagerrating, there are only 7,845,692 known sharpening techniques amongst the graphic experts.
Nah, cuzz I just invented two more.
I have thought that if a shot looked good small but not larger that I would be "cheating" to pretend that it was a good shot. Obviously (what a surprise!
Thank you again.
RE:
I doubt too many would view it as cheating, but if an image only looks good small it definitely limits the options you'd have in using the shot. A good shot at a small size is, well...a good shot at a small size!
Cheers...
Brad
Regarding the question of sharpness you have received great advice. I use PK Sharpener which is a highly regarded 3 phase sharpening regime. The "capture sharpening" gives just enough to compensate for the problems of digital capture. The "creative sharpening" is where you selectively or globally sharpen to taste on an otherwise fully processed image. finally, the "output sharpening" is used after the image is re sized for whatever output the image will be used.
My point is that many images that look borderline are rendered excellent with careful processing.
Good luck with your image recovery.
Perhaps I gave you that view, and if I did I will not retract it.
I think some of the advice given here is a recipe for mediocrity, and it is a bad strategy to rely on using small versions only because the image is not sharp enough to present at larger size. Do yourself a favor and set high standards early for your own work - you will never regret it later.
However, I do advise not deleting raw image files until you are sure you can read sharpness. You will develop that skill in time, and then you can purge your files to weed out the marginal shots.
Regarding the other options about using small versions, I do think that if an image is going to be posted here on NPN it should be sharp enough to post a larger version. There are times when compression quality will be so low in a complex image that a larger version can't be posted, but most times people don't say why they don't post a larger version (except the Alaskans and any others who are using dial-up). If it's not posted because it's not sharp enough at even a 1200-pixel size, that seems to me to mean the image is not that great. But it's pretty frustrating to people trying to critique an image to see only the smaller version, and if it's a stunning image I really, really want to see the larger version.
Also sharpness is subjective. We're all just talking hypothetical here since I really have no idea what your idea of sharp images means because I haven't seen any examples.
I can't say anything better on the matter of sharpness.
About the editing process, as a personal taste, I only delete after the download the pictures that are technically wrong (say out of focus, unwanted blur by non steady camera or moving subject, subject view obstructed by something else, unwanted "mutilation" of the subject, way too much over/under-exposition). And, with the exception of few sure keepers, I let the so-so images rest a little bit after being keyworded.
This Limbo phase may be even months long.
I found that looking with a dispassionate eye my work often helps the judgement and this process needs a "detachment" mood that takes time to happen.
Actually, I am a heartless picture deleter: only about 2-3 % survive!