Paul – generally that is what I do – the only other thing I find that may need alteration after resizing is the scaling/bevel size on any layer styles that I have applied to the images as they sometimes need tweaking.
Susan S.
Well, Susan,
I guess yours is the only advice I will be receiving. I was hoping to find out I was doing something wroing, but I guess not. Thanks for your help.
Paul B.
I was hoping to find out I was doing something wrong, but I guess not
Does that mean you don’t like what you are getting, Paul? My guess would be that your problems are resolution-related, in that case.
Photoshop is not the best software for image resizing. I usually make changes and apply filters on the original image using photoshop and then move it to macromedia to resize it.
Barbara,
No, there is nothing that is making me think I am doing something wrong. But it is satisfying to find out that by making some simple change, pictures will look better. Since I don’t yet have the reasons for why something should be done before or after downsizing, I lack the confidence that the practice that I had developed was optimal. I was hoping to be steered clear of another myth, like I was with the number 72. I tentatively inferred from the fact that no one else had said anything after Susan, that others agreed with her. But I wasn’t sure.
I suppose there is only one uncertainty left. I have adopted the attitude that one should prefer to start with the largest digital image available, and reduce from that. The alternative, presumably, would be to pick that image size that would require the least amount of reduction. For most purposes I have no choice, but for some of my photos, there is a range of scans from quite small to quite large. The fact that there is more ‘information’ in the largest size, has made me want to start there–although I have some doubts about whether that makes sense. I do tend to crop a lot, but still the reduction is usually pretty great. Perhaps a better way is to look at the maximum photo first, and if it seems that after cropping there will be a significant reduction necessary, open up the smallest image possible and use that. I have no idea how ‘intelligent’ the reducing process is using Photoshop Elements. I gather Matt is saying that Macromedia products (e.g., Fireworks, which I have) do a better job. It sounds also as though if your goal is to put images on the Web or make small prints–and you do your own scanning–one is best off scanning to the size that will require minimal reduction.
But this would raise another question: you don’t have to do any reduction of an image in order to print it at a smaller size. You just tell the computer and printer what size you want the picture and it will take care of the matter (constrained by the number of pixels per inch you want to use for printing). Do pictures look better if you do reduce the image (if you can) before printing? Or is the process of making the output smaller effectively the same in either case?
–Paul
Paul, you’re definitely thinking about this one, aren’t you?!
It’s hard for me to come up with a reason not to start with the most pixels possible, regardless of how many of those pixels might ultimately discarded. I have a friend who insists on taking pictures at 640×480 pixels and brags about how many images he gets on a memory card. He was doing it primarily for web use, but….he had an image that he wanted to print and was disappointed with the results. I shoot everything at my camera’s max image size and minimum compression; if I have an image that’s only going to be used for the web I can convert it using Save for Web and ditch the original (although I never do).
The ‘intelligence’ of downsampling hasn’t had much play on the forum; we’re usually fixated on the concerns of upsampling. I’ve assumed that downsampling is done pretty much the same by Photoshop and other apps – maybe not. Would be something worth knowing, although the only time I downsize is for web use, and that’s hard for me to get excited about.
With respect to printing, I never toss away pixels for printing – I just change the linear dimensions in Image<Resize<Image Size with resample unchecked and let the resolution float upward. I read something a few months ago that implied that high resolution did NOT cause higher ink consumption. Don’t know what to believe in that regard, but I’m sticking with the approach of delivering max resolution to the printer – until otherwise convinced!
Chuck
Paul,
Scan for your intended output (which seems to be for Web and small prints).
Why not do your cropping in the scanner software, you say you crop a lot. After the preview scan, there should be a dotted line box which you can drag diagonally to encompass the part of the image you wish to keep. Draw your box on the scan preview and have it do another preview scan. Now only the desired area of the image is showing in the scanner software, ready for the final scan.
Here’s a "for instance"…
If I scan a 4×6 at 240ppi (my Epson printer likes to print to 240ppi) I will create an image of 960 pixels by 1440 pixels (4×240=960 and 6×240=1440). This will print out at 4×6 using 240res. If I want that image to nearly fill a screen on a web page, top to bottom in this case, I would go to Image Resize, and with Resample/Constrain check marked, type in 600 for the height and width will fill itself in at 400. If I wanted two images to fit on a web page, stacked up, I would Resize by filling in 300 for the height and the width would fill in with 200. Go to Save for Web, choose the quality of the jpg you are converting to. You don’t need to change the dimensions of the image, it’s been done.
If I want to use only part of the image, use the crop tool in the scanning software. If I drew a box which included about half my image, say 2×3 in this case, but I would like to be able to print this portion out as a 4×6 at 240 res, I need to scan at 480ppi. Double the scanning res. because I will be making a 4×6 print out of my selected 2×3 area. To print…Image Resize = UNCHECK resample, change the res. from 480 to 240 and don’t change anything else. Now it will show as a 4×6 and print at that size. This 2×3 scanned at 480 also is 960 pixels by 1440 pixels. To modify it for a web page, follow steps from above…in Image Resize, CHECK MARK resample/constrain and type in 600 or 300 for height (per our example). Go to Save for Web etc.
In proceeding this way, you have allowed for printing the image at a modest size, however, if I thought I might like to print out at 3 times the actual size, I would scan at 720, or at 960 for 4 times. Decide what part of the image to scan, how large you might ever want to print it and use those figures. You said small print, so that’s how I set this up. In doing so, you do not have unneccessarily large file sizes produced from your scan.
Hopefully this has answered some of your questions.
Nancy
edit—whoops, I was typing as you were replying Chuck
According to what I have read you want as much of the original image as possible to do your edits with. For example, to edit the white balance you would want to edit an uncropped image rather than one that may have had much of it’s highlights cropped even though it may be cropped later. The same would hold true for color or contrast I would think. I have always cropped after making adjustments so I can not tell you how the edited image would be affected otherwise.
—
Have A Nice Day,
jwh 🙂
My Pictures
http://www.pbase.com/myeyesview
Nancy, excellent rundown! Only one minor quibble: a 2×3 image isn’t 1/2 the size of a 4×6, it’s 1/4…
🙂
Chuck
Now you just leave Nancy and I alone, Chuck! We have our own way of calculating this stuff, and to us it’s half as big! 🙂 However I do know what you mean.
Beth, you’re right. If it’s half the length and half the width, it’s only logical that it’s half as big. What was I thinking?! Sorry, Nancy….
🙂
Chuck,
Not 1/2 in volume, the linear dimensions are halved 🙂
Paul,
To clarify one point…my Epson printer has a stated resolution of 720/1440. I SEND it an image with a res. of 240 (720/240=3=an even number) and tell the printer to print it at 720dpi. Sometimes I use 1440, but really I hardly see a difference…which may be because I don’t have a top of the line photo printer and I almost always print on matte paper.
Nancy
end of this story
Oh, I have to throw in!
"Not 1/2 in volume, the linear dimensions are halved"
I’ve seen some loud photos, but I didn’t know a photo could actually have volume!
Eric,
:)—glad I provided a laugh for you today…
It just seems an easy way to explain to someone why you would double the scanning resolution when wanting to again produce a print that is 4×6 after you have cropped it down to 2×3. I am getting defensive it seems…I retract "half the size".
Nancy
May I show my ignorance and ask what the Unsharp Mask is, Paul?
Thanks,
Lorace
Lorace, you definitely have some homework to do on this one! Unsharp Mask is the premier sharpening tool in Elements and as such one of the most used tools to take the ‘softness’ out of digital camera images. Lots of good tutorials out there on its use…
Chuck
Chuck said the only time I downsize is for web use, and that’s hard for me to get excited about.
Now that it has been made explicit, I guess that is pretty much the right attitude. There is never a need to down sample for the sake of printing.
Once again, I think we have evidence that with respect to digital photography no one ever needs to hear the number 72.
Nancy says Scan for your intended output (which seems to be for Web and small prints).
I dont have a scanner, but according to Chuck, there is no reason not to scan at maximum size for the purposes of printing (other than to save hard disk space). And what is the loss of scanning at maximum size for the Web? Unlike in the case of printing, you will have to downsize, but at least you get to see what detail is available in case you want to crop a section. Sometimes I use both a complete image plus a detail as a second picture. Doesnt it make more sense to make these sorts of decisions on the computer using Photoshop Elements, rather than on a scanner. Often you dont know what youve got until you fiddle with the image. I have images photos where the background looks completely black, but which reveal all sorts of stuff after using levels. This mainly be relevant to scanning from negatives.
What is the benefit of cropping on a scanner, where you cant really see the image very well. I dont have a scanner, so I may misunderstand the process.
Jwh says I have always cropped after making adjustments.
That would be different from the way I do it. Do others agree about this? The only thing that has stopped me from cropping right away is the need to make some adjustments to see what I am working with. It may turn out that the really dark background can be made to reveal stuff that is worth keeping. Otherwise, it seemed to me wrong to modify the image with stuff that is going to be dropped anyway. Cropping allows you to focus on what will matters. But are there reasons of quality, to make color and tone adjustments on the uncropped image?
Nancy says I almost always print on matte paper. Why?
Lorace asks
what the Unsharp Mask is
? It is the filter (under sharpen) that should almost always be used to sharpen (rather than unsharpen) your image. I start at the settings of 85%, .05, 0, and make modifications according to what looks good to me. (Is that optimal?)
–Paul
Paul, I get very nervous when I read the words ‘according to Chuck’….I’m no authority on anything, especially on scanning; I have a cheap scanner that I’ve used maybe five times. There are techniques I’ve read about for double scanning very contrasty pictures and using Elements to ‘sandwich’ the two shots and bring out details in both the highlights and shadows. And scanners can be cranked up to very high resolutions that make absolutely enormous files; are they really any better than scans at lower resolution? But if I ever start giving opinions about scanning again, someone should remind me to get back in my place!
Good post, by the way….may comment on some of your points later after others have chimed in.
🙂
Chuck
Paul – I also use Epson’s heavy weight matte paper – because it’s half the price of the glossy, gets nice even gradations of tone on my printer and under glass looks the same. I don’t have a high end photo printer and the high quality Epson photo papers don’t give a particularly superior result – if I had a better printer I might experiment a bit more.
When I scan I do most of the adjustments in Elements – the scanner software I have isn’t too flash. If I were using something like vuescan which I think can do basic adjustments using more bits, then it might be worth while doing more in the scanner. I always scan at the highest optical resolution of my scanner – the thing is so slow I don’t want to have to repeat the process at a higher resolution later.
I use unsharp mask heavily (probably too heavily!) – I am a bit lazy with a lot of my images and tend to sharpen for print and then just resize for web purposes which really needs a lot less sharpening. Different settings (lower percentages, much higher radius) can also be used to reduce haze and increase contrast in landscape shots.
I guess the problem with re-jigging the set up in Elements is that one person’s weird filter is another person’s bread and butter. I use an entirely different set of tools when I am processing photos to when I am processing calligraphy or original art work or creating digital "art".
Interesting thread..
Susan S.
Paul wrote:
Jwh says "I have always cropped after making adjustments."
That would be different from the way I do it. Do others agree about this? The only thing that has stopped me from cropping right away is the need to make some adjustments to see what I am working with. It may turn out that the really dark background can be made to reveal stuff that is worth keeping. Otherwise, it seemed to me wrong to modify the image with stuff that is going to be dropped anyway. Cropping allows you to focus on what will matters. But are there reasons of quality, to make color and tone adjustments on the uncropped image?
James replies:
The reason I do it this way is so that I have all available data to make corrections with. For example, one tool I use often in another program is a Black/White/Gray point adjustment. If the optimum white (black or gray) points fall outside my intended crop area I am forced to choose another point within the crop that may not give me the results I am after for my final cropped image.. By having the entire image I can make my correction using the proper points to get the image I want and then make my crop. I do not think there is a right or wrong way of doing this it’s just my preference.
—
Have A Nice Day,
jwh 🙂
My Pictures
http://www.pbase.com/myeyesview
Paul,
Hmm…I can see my image quite well on my scanner preview. I do a rough crop at this point. Cropping at this stage is not, for example trying to crop exactly around a face or some tiny detail. What is the point of scanning an entire photo of a darling house with an ugly gas station next to it when all I want is the house? I would just discard the gas station portion by cropping in Elements anyway. It is excess baggage to me. Rather like the "Blue Plate Special", I want the roast beef and it comes with spinach and a baked potato. In order to get what I want, I have to take the whole thing and then just discard the portion I don’t desire. Maybe it is about having vision of where you want to go when creating an image.
I said, scan for your intent…if the person ONLY wants to place images on web pages (need an image 800×600 pixels or less) and maybe make some small prints, a very modest size image file is what he requires. He doesn’t need to purchase a 1 ton truck to transport a box of Kleenex! Again, it is a matter of knowing what you want to do with your digital file. If one might like to print this file out at poster size, well of course you would scan at a high res.
It works better for me to crop and then do my tonal adjustments (usually). When using only part of an image it no longer has a relationship with the missing parts. Perhaps this part has limited tonal variation, compared to the whole. This part would be more interesting if it had more range and I can make that happen.
Matte paper is about half the price of some of the glossys. I get good results using the Epson Heavyweight matte on my Epson printer. As Susan says, under glass it looks the same. All my prints are under glass.
Nancy
I get good results using the Epson Heavyweight matte on my Epson printer.
Nancy,
Epson Heavyweight Matte is my favorite paper! Brent Bertram, a longtime contributor to this forum recommended it a long time ago. He said it gave him the best results. When he was having printer problems, that was his "standard." If the print didn’t look good on that paper, he knew he had a problem.
Re workflow order: I prefer to crop first and then make adjustments. That way, I am adjusting levels, brightness, sharpness, etc. based on only the pixels that will be in the final image. That makes the most sense to me. I do sharpening last, after all other adjustments.
Bert
Thanks for the various comments. I was under the impression that in regular film printing matte paper produced an image that was less sharp. I figured this must be true for digital printing too, and that this was why matte paper was cheaper. Based on what several of you have said, I will try matte–Epson Heavy Duty Matte, in particular.
I think I have absorbed the collective wisdom on the other points–at least as much as is necessary until I actually own a scanner. I guess everyone agrees that with the exception of using the unsharp mask, all changes should be made before reducing the size of an image–something that one only needs to do all for the sake of the Web.
As to whether to crop on the scanner, it can be profitably done when there is someting one is sure one will not use.
There is one person who finds a benefit to adjusting tone, etc. on an uncropped image. Most crop, then make tone and color adjustments.
There is some variation among users as to what the frequently-used tools are. I would guess, nonetheless, that there are some tools that most people use most of the time, and some tools that most people almost never use. So, if nothing else, a way to customize the shortcuts bar would be a improvement of Photoshop Elements. Thanks again.
–Paul
Paul,
It gets down to whatever works best for each individual. Each person’s needs, requirements and sense of ‘pleasing’ are different. …so many variables, even the matte paper issue. It is not out of the realm of possibility that printer brand/model may influence results on different papers. These differing results are often expressed as YMMV, your mileage may vary.
I definitely agree with Nancy. With the exception of some basics (like don’t print at 72ppi and expect a masterpiece and don’t do extensive editing and resaves in JPEG) so much of this is a matter of both personal preference and variations in equipment. I do have some Epson Matte paper, but I don’t have an Epson printer, so I’m not real pleased with the results, although I’ve found a couple of other uses for it. And sometimes I crop before I scan and sometimes I don’t; it depends on what I’m scanning and what I think I might want to do with the image.
It seems there are times people would like us to point at something and say "it has to be done just like this", but that’s rarely true, especially when using Elements. The best we can do is collect ideas and see which ones "fit" and which don’t. I’m a "marquee parker" and other people prefer parking on the little hand. Does it matter? Heck no. And I probably like more vivid color in my prints than somebody else, but I’m the one who looks at them, so why should anybody else care? 🙂
Beth,
Very wise words. Unless you’re a professional, you’re doing this for yourself, primarily, so that’s the one person you have to please. If you’re doing it for friends or family, they’re getting your services free, most likely, so they will be happy with whatever you do.
Moreover, there are usually MANY ways to what you need to do. In my case, the way I do things is not necessarily the quickest or most efficient, but it’s a way that I KNOW will work, and I’m used to doing it that way. Each PS user has to find those familiar paths through the maze and perfect their technique. If it takes a few more keystrokes, so what?
Bert
OK, but this is starting to sound ecumenical-relativistic-Canadian. Would you at least be willing to concede that the number 72 is irredeemably evil?
–Paul B.
Paul- What is ecumenical-relativistic-Canadian? Canadians that believe certain numbers are evil as they relate to dpi?
6 x 6 x 2 = 72
From the Bible, you know about those satanic sixes, don’t you? 🙂
Bert
Yes, but what if they got it wrong…perception wise that is… and it’s really 999 that’s evil? Now that throw a monkey wrench into the mix.
Yeah, you’re right….they could have just been looking at it upside down. Never thought of that. 🙂
Bert
