PDA

View Full Version : To CROP or not to CROP, that is the question


tomlaronge
April 21st, 2007, 04:59 PM
I have been in the habit of editing, cropping and saving an image as a PSD file that I want to eventually print. In otherwards, if I eventually think that I may want to print an image as a 4" X6" print, a 5" X 7" print, an 8" X 10" print, an 8" X 12" print and an 11" X 14" print, I have been editing, dropping and saving 5 separate and appropriately labeled PSD files.
1) Is this the best way to preserve the information or should I edit the image without cropping and save this as a PSD and crop only at the time of printing?
2) Is there yet a better way to accomplish saving an edited image in a non-lossy format for eventual printing in several different formats?
Thanks to all for educating me.
Best regards,
Tom

Juergen D
April 21st, 2007, 05:28 PM
Tom,
Save the files non-cropped. You only have one file, which you can crop any way you want when it comes to printing. You do not even have to save the cropped version if you do your own printing.

Juergen

TonyW
April 21st, 2007, 06:18 PM
I agree with Juergen although I usually only save a PSD file if I've done a lot of editing. Most of the time I just keep the original as is from the camera (RAW or JPG) as "the negative". I've found that as I get better I can do any basic adjustments a) better and b) quicker so in most cases the original works for me. I've just been going back to some 5 year old PSD files which must have been done with PSE1 or PSLE and realised I really didn't know what I was doing then plus my editing software has got a lot better - fortunately I still had the original JPEGS :)

tomlaronge
April 21st, 2007, 07:45 PM
juergen and Tony,
Much gratitude for your assistance.
Please tell me one other thing related to this. Now I edit a JPEG or NEF image file, insert 300 dpi as the desired printer resolution, insert the image ratio or the exact height and width, crop the and print and then I have been saving the result as a PSD for each different print size. If I do this say for a 4" X 6" proof image, how do I recrop the PSD to print an 8" X 10" image? What am I missing as the aspect raio of the 4' X 6" PSD is very different than that of an 8" X 10" image as you know. Perhaps in recropping, I need to include some of the original image cropped prior to making the PSD. I am quite confused on these points.
Thanks again,
Tom

Juergen D
April 22nd, 2007, 01:13 AM
Tom,
This is another reason to save the uncropped image file. You will be cropping away parts of the image when going to certain sizes. Those parts cannot be restored if you want to go to another aspect ratio later.
In this example, if you cropped to 10 x 8 first, you would loose a substantial portion of the image, which you would want to keep for a possible 6 x 4 or 7 x 5.

http://www.pixentral.com/hosted/1s0IXwUMlnS69utnxgK5dMPQouZXrL1_thumb.jpg (http://www.pixentral.com/show.php?picture=1s0IXwUMlnS69utnxgK5dMPQouZXrL1)

Juergen

TonyW
April 22nd, 2007, 08:46 AM
It's also better not to put a number (like 300) in the resolution box when you crop. If you do that you're telling Elements to add (or subtract) pixels by interpolation. It's a bit of a myth that 300 ppi is the right resolution for printing. It's better I think to just keep the original pixels intact and avoid resampling (interpolation) if at all possible.

Tony

tomlaronge
April 22nd, 2007, 11:21 AM
Juergen,
What you wrote makes all the sense in the world and I shall do this from now forward. This is exactly why I have been using all that extra memory to store multiple files that I have first individually edited and cropped.
Thank you for helping me and hopefully many others.
Regards,
Tom

****************************************

Tony,
As I started to learn to use PSE and then to print, one of the sources I read stated that you get the best results with a Canon printer at about 300 dpi, with an Epson printer at about 240 dpi and that 72 dpi ought to be used for a web published photo.
1) What is wrong with what I read?
2) If you should best not insert a number in the resolution blank, what is the purpose of this blank, when should a number be inserted herein and how to you control the reolution being sent to the printer?
3) Do you know where there is a good, readable description of resolution, captured, edited and fed to printers that I may review to get my own thinking straight. What I quoted above is from one of the commonly recommended reading sources for learning and for using PSE. If this is incorrect, I sure would like to learn.
Tony, one great big thanks to you for your help. I trust this may help a whole lot of us PSE users.
Have a great day and best regards,
Tom

TonyW
April 22nd, 2007, 12:15 PM
Tom: There's nothing really wrong with those guidelines except that it's not dots per inch - it's ppi (pixels per inch) and there's a big difference between the two (printers use dpi and the numbers are much higher) and the 72 ppi guideline for web publishing is wrong. The only thing that counts for web publishing is the number of pixels wide and high. Monitor resolutions are all different and I doubt there are any left that are 72 ppi (mine's 90 I think but it doesn't really matter).

My argument for leaving the resolution box blank is that the printer takes pixels and converts them to many dots of different colours. So the printer software is doing a lot of conversion to get what it wants and I'm happy to leave it to do it's thing rather than try and second guess what it prefers (and I have tested my Epson and I couldn't see any difference).

Why do they have the box? Maybe there are some printers that like to be told a specific resolution but I've never owned one and I can't remember ever putting a number in that box.

As for the good readable description of resolution I'm sure there is one but I haven't really seen it. I've seen several that look good but actually contain errors which just add to the confusion. Of course maybe I haven't got it right either and I'm always willing to be convinced that I'm wrong and they're right. ;)

Tony

Juergen D
April 22nd, 2007, 01:34 PM
Tom,

Let me throw in my 2 cents... Tony is perfectly right recommending to leave the box blank. This way no resampling (up or down) will take place. Since you are storing the uncropped image, which you may crop at a later time, you should also not alter the resolution until maybe later.

Usually your image will have plenty of pixels to accommodate a good printing resolution when the time comes. Let's assume that you have a 5 megapixel point and shoot camera and your regular image, as it comes off the camera, has 2592 x 1944 pixels. If you crop this image to 10 x 8 you would be left with 2430 x 1944 at a resolution of 243 PPI. Perfect to print. Now take the same image and crop to 6 x 4 and you'll have 2592 x 1728 pixels at 432 PPI. Also perfect to print. So, no reason to degrade the quality of the original image by resampling.

There is a time when some resampling may be necessary, for instance when cropping in very far into a small portion of the original. Then, however, the resampling should be done very carefully (possibly in multiple steps) and not with the Crop Tool, but via >Image >Resize >Image Size. This is another story altogether.

Juergen

kroberts
April 22nd, 2007, 01:57 PM
Juergen,
Thank you for helping me and hopefully many others.
Regards,
Tom
****************************************
Tony,

Tony, one great big thanks to you for your help. I trust this may help a whole lot of us PSE users.
Have a great day and best regards,
Tom


Tom's right, this discussion has helped me! I just noticed the resolution box today (very observant, huh?:o ) and wondered if I should be putting a number in when I crop a photo.

Tom, Thanks for asking the question! Jeurgen and Tony, Thanks for your responses! There is so much to learn and so many to learn from on this forum!:)

tomlaronge
April 22nd, 2007, 04:31 PM
Tony and Juergen,
The information you provided id invaluable. Juergen, I followed your calculations to a tee, I just never did the calculation myself previously. And, I just filled in all the blanks based upon what I read in the more and more apparent not so magic books.
1) So, where did you two learn this stuff correctly. What you wrote is so simple and straightforsard, but it is certainly, to my observation what is in the great reading sources recommended throughout this forum and elsewhere?
2) Using Juergens' direct mathematics, it becomes easy to see that a 16" X 20" print from his 5megapixel camera would be markedly deficient in pixels for sound printing in accordance with the standard guidelines in the basic texts discussed in this forum. I have seen some excellent prints of this size and larger made from similar file sizes. How does this work/result? What are the reasonable limitations on the effective size of good prints made from a camera having a half 35mm CCD or equal, stable electronics, good chip cooling and a file size between say 5 megapixel and 10 megapixel, i.e. the range in which most of us probably work these days?
3) Now I have to add my observation. This is that specific workflow ought to be laid out, taught and the reasons for such explained to us beginners. This would greatly help us become encouraged as we achieve reasonably satisfactory results that we can fine tune as we learn rather than go broke buying ink and paper due to having to throw out those hideous initial results, and of course, magnums of aspirin, Tylenol or the like to attempt to quench the pains of confusion from the $25 to $35 wonder texts.
In conclusion and awaiting your experienced replies, THIS FORUM IS AWESOME!!!!! I for one deepky appreciate the privilege of being a participant.
Thank you all once again,
Tom

Juergen D
April 22nd, 2007, 05:08 PM
Using Juergens' direct mathematics, it becomes easy to see that a 16" X 20" print from his 5megapixel camera would be markedly deficient in pixels for sound printing in accordance with the standard guidelines in the basic texts discussed in this forum.
Not necessarily... There is this rule, which states, the larger the print size the lower the resolution may be. This is due to the observer's perception, looking at a large image from a farther distance than at a small one. The 20 x 16 would still have the previously mentioned 2430 x 1944 pixels, distributed over twice the width and height. This brings the resolution to exactly half, 121.5 PPI, which may yield a quite acceptable print (at that size!). Of course, depending on the subject and the quality of the photo, it may be more or less acceptable...

Juergen

Rusty
April 22nd, 2007, 05:16 PM
Juergen, you are exactly right.

I went to a gallery Friday. They were selling matted and framed photographs at (at least for me) unbelievable prices. Most were very large ... no 8x10s in this bunch :)

At 8 or 10 inches the grain (most were film) was very noticeable. At normal viewing distance they looked just fine.

Rusty

Juergen D
April 22nd, 2007, 05:21 PM
I have seen some excellent prints of this size and larger made from similar file sizes. How does this work/result?
If you have a fairly low resolution like 121.5 PPI and want to increase it, you can best do so in small increments. You go to >Image >Resize >Image Size. In the dialog box check Resample Image and increase the resolution by about 10%. Do this step several times until you have reached a targeted higher resolution, in this case maybe 180. This procedure will soften the image to some degree since new pixels are added. But the printed output may look more pleasing.

Juergen

TonyW
April 22nd, 2007, 06:00 PM
Tom: Like Juergen says it's not quite as simple as one resolution fits all (although many books and tutorials try and make it that simple which just adds to the confusion). The bigger the print the less resolution you need because you look at it from further away. A good analogy is a wide screen TV. Those in fact have rather low ppi's and the bigger they get the lower the ppi but they get away with it because you sit further away to watch it (or that's the theory anyway :) ). So you might need 300 ppi for a 4x6 but by the time you get up to say 24x36 you can get away with a lot less - typically about 110-120 ppi.

Here's a table that I have - it came from various sources but pretty much matches the recommendations in Scott Kelby's "The Digital Photography Book"

953

Tony

tomlaronge
April 22nd, 2007, 06:19 PM
O.K. thank you all. I neglected to consider the viewing distance. I do know from video work that proper viewing distance is about 3 to 4 times the image diagonal.
Based upon everything Juergen and Tony have shared, it seems that PSE would very much benefit if one could revert or undo the crop function at any time regardless of where in the editing process the file is cropped. Frankly, I prefer to look at the image in the editor, select the image size I wish to print, crop and edit the cropped image. After printing, I sure would benefit if I could undo the early cropping step so that I could then store the entire edited image file for later recall and printing. All one would then need to do is to bring the stred and edited file into the editor and crop whatever size is desired at the time and then print. WOW, I should think this would greatly improve the utility of PSE for printing.
You all are great and very helpful.
Tom

TonyW
April 22nd, 2007, 06:29 PM
Tom: You're right - that would be a good feature. The only way to do it now is to crop last but being able to delete an earlier step in the Undo History would be a very desirable feature.

Tony

Chuck S.
April 22nd, 2007, 06:35 PM
Tom, I used to keep several crops of some images, thinking that I would go back and print more of a certain size....but it hardly ever happened. Now my workflow is generally something like this:

Open image.
Perform all edits except sharpening.
Save as PSD.
Crop to print size #1
Sharpen
Print
If edit issues show up in print, Undo History to step 2 and try again.
After acceptable print is achieved, Undo History to last edit before cropping.
Save as PSD (uncropped and unsharpened)Of course, I could also make use of the Organizer version sets by saving after acceptable print was achieved. Then I'd have an original, an uncropped and unsharpened edit, and a cropped and sharpened edit of one size. Each succeeding crop could be saved as part of the version set. And eventually, it wil be time to buy yet another external hard drive for all this extra stuff....:)

Juergen and Tony: great responses! Thanks for helping all of us refocus on this topic.

tomlaronge
April 22nd, 2007, 06:36 PM
Tony,
I just submitted the idea to the WISH LIST of this forum. I cannot thank you and Juergen enough for all of your help here to say nothing about how much time and memory in my computer I will be able to save doing as you two have iinstructed.
If you do not mind a mostly bald top, my hat is off to you and to Juergen. Please have a great week.
Tom

JackF
April 23rd, 2007, 08:35 AM
Tom: There's nothing really wrong with those guidelines except that it's not dots per inch - it's ppi (pixels per inch) and there's a big difference between the two (printers use dpi and the numbers are much higher) and the 72 ppi guideline for web publishing is wrong. The only thing that counts for web publishing is the number of pixels wide and high. Monitor resolutions are all different and I doubt there are any left that are 72 ppi (mine's 90 I think but it doesn't really matter).

I respectfully disagree, 72 ppi is perfectly fine for web publishing. There is no reason to go higher. While it's true that different monitors have different ppi's that's really a non issue. You cannot tell the difference in a 72ppi image vs a 90, 96 etc. image so it makes perfect sense to use the smaller 72ppi for a "smaller" file size.

JackF
April 23rd, 2007, 08:46 AM
I respectfully disagree, 72 ppi is perfectly fine for web publishing. There is no reason to go higher. While it's true that different monitors have different ppi's that's really a non issue. You cannot tell the difference in a 72ppi image vs a 90, 96 etc. image so it makes perfect sense to use the smaller 72ppi for a "smaller" file size.

I need to opologize to you Tony. Sorry :D

You are correct Tony, I've never tested that when resizing an image. Pixels wide x pixels high is all that matters for the web, resolution has no effect on file size. I need to look into this a little further so "I" will understand it better.:eek:

I do always use the Bicubic Sharper interpolation when downsizing, don't know if that really makes a difference...never tested it but it's what "they" say to do when downsizing. LOL

TonyW
April 23rd, 2007, 09:15 AM
Jack: It's very common to find that 72 ppi recommendation for sizing for the web and it would be perhaps true if you cropped to inches rather than use Save for Web or Image Resize and put in pixel dimensions. But if you did crop to inches and use 72 ppi the only people who will see it the same size in inches are people running an 800x600 display on an 11" wide monitor and there can't be too many of those around these days.

I use Bicubic Sharper too for downsizing and I do think it makes a difference - particularly noticeable since I got a bigger MP camera that shoots sharper images in the base case. One Photoshop feature missing in Elements is that in Photoshop Save for Web lets you select Bicubic Sharper when downsizing - I don't know what Elements uses but I think it's probably just Bicubic.

Tony

Chuck S.
April 23rd, 2007, 09:20 AM
Tony, I wondered about that (whether Bicubic Sharper made a difference when downsizing for web). I'm going to have to experiment with some fairly sharp 10 MP images I took recently; I just did a simple Elements Save for Web to size them for my gallery. Maybe there's more that can be accomplished there. Thanks for the tip!

tomlaronge
April 23rd, 2007, 02:21 PM
Chuck,
Thanks for your succinct and appropriate comments. Your workflow summarizes exactly what I read as the suggestions put forth by Juergen and Tony. Frankly, I have been doing the workflow you described last and getting tired of the huge demand upon memory and also on its obvious inflexibility.
One comment that I truly believe to be a major weakness in PSE is the fact that yyou cannot crop first and then undo the crop after printing and before converting to a PSD and saving. Just as beauty is in the eyes of the beholder, it is often both easier and better to edit a cropped selection in preference to an uncropped selection just as one would burn and dodge during printing of a cropped enlargement in the darkroom. Working on exactly what you desire as a finished product has many benefits.
Chuck thanks again for your comments and for summarizing these into a workflow. Have a wonderful week.
Regards,
Tom

Chuck S.
April 23rd, 2007, 02:36 PM
Tom, I agree with you regarding the visualization advantages inherent in cropping first. Now if we can find a way to preserve all those edits between cropping and printing so they can be applied to other crops of the same image.

In full Photoshop, you may be able to group all those edits after the initial into an Action, which then could be saved and re-used after each subsequent crop. There would be some limitations to that approach, but it could help streamline the process. It wouldn't be foolproof (e.g., if you did any editing that involved a mask it might be problematic) but it would work for many types of edits that you might perform.

tomlaronge
April 23rd, 2007, 06:43 PM
Chuck,
Thanks for the comment. I am afraid this is way too cumbersome unless your business is selling memory devices. We really need a new generation of Elements/Photoshop sequenced editing steps are individually addresable nd modifiable, at least to the extent of reversion and redoing. I think such software would revolutionalize the photographer's acceptance of the digital darkroom.
Best regards,
Tom