Resizing Photos

[caleb]

Ok, I can think of three reasons to resize photos:
-Printing?
-Web
-Editing*

Printing has a question mark next to it because I have been using full res files (even for 4x6).This doesn't seem right to me because I've read that sharpening should be done for final output size AND I have a feeling using the techniques you guys will tell me will downsize the files for printing a lot better than the photo lab.

Web. This one's obvious. Two concerns are file size and, as above, downsizing that may degrade image quality. Again... I work on full size images, check everything at 100% which is usually 7-10x larger what I output the file as.

*Editing. I starred this one because today is the first time I ran into ever wanting to resize for editing. I was making the attached image. Since it's 6 or 7 merged files, my width was something obscene like 12,000 px. Although I didn't notice too much processor lag, I can see that it's at least plausible for me to want to resize so that I'm not working with such massive files.

What I currently do: All post processing techniques at full size. Sharpen until I see halos at 100% (sometimes I sharpen on multiple layers if I have like skin, trees, and maybe a cell phone or something that can take a lot more sharpening). File>Save for web/other devices... I use this the same for outputting to web and when getting together files for printing. I use maximum jpg output (unless file size is an issue) with sRGB color profile.

What I should be doing: [see posts below]

[Photo-John]

Caleb-
I apologize for not getting to your question sooner. I've had it open on my desktop since you posted it, if that makes you feel any better

Most of the time I am resizing for the Web and ocassionally, when I have a client, for print. I never resize to edit. I a can see how you might want to with a stitched image. But even then I work with the full-size file and only resize when I want to share it.

For the Web, I resize in steps. I first resize to 2500 pixels on the long dimension. The I resize to 1500 and use Unsharp Mask. I then resize again to 1000 pixels and then once more to my final Web site - usually 600 to 800 pixels. After the final resize I use Unsharp Mask again. Unsharp Mask settings vary depending on the image. But my standard sharpening settings for Web are Amount: 100, Radius: 0.3, Threshold: 3. I find that works pretty well for most Web images.

For printing, I don't usually resize more than once. Some people get really anal about print reasolution. And I'm sure it makes a difference. But I haven't found it makes enough of a difference for me to me too particular. Generally I set my output dpi to 250 or 300, at the final print size. To do that I open up the "Image Size" dialogue box in Photoshop (tip: right-clik on the blue bar at the top of your image), uncheck the "Resample Image" box and then change the dimensions to my final print size. Then I recheck the "Resample Image" box and change the resolution to 250 or 300. If I have to upsize, I may do it in steps. But if I have to downsize, I don't worry about it. After I've resized I sharpen. In theory you should check your sharpening at output size. But this is usually some weird ratio that Photoshop won't render well. So I usually toggle around between print size, 50% and 100%. I've found that if it looks good at 50%, it usually looks good in print.

Ok, that's all I've got for now. Let me know what I didn't answer or glossed over too much. And if anyone disagrees with me, by all means, tell us why. I can take it

[caleb]

Hey, no problem and thanks for answering. I've never thought twice about it until you posted something in another thread.

1) Why you sharpen at 1500 and again at 600-800?
2) I'm very curious about color profiles and which ones printers can handle. When outputting photos that I will be printing, I worry about the printer not being able to reproduce some minute shifts in color. Is there any standard (from what I understand, the standard for the web is sRGB?) for printers/jpg file format? Should I save as a different file format?

I just don't fully understand why resizing multiple times would produce a better image, although I will be trying this for a week or so to compare. Thanks for sharing your USM settings, by the way. I've always left the radius in the 1.0-1.4 range, regardless of image size. Although in hindsight, that doesn't make too much sense

[drg]

A couple of other notes to consider.

One suggestion for down sizing/sampling is to try using the Bilinear model as opposed to Bicubic resizing algorithm. Whether you use steps of sizing or not with some type of images it may result in a sharper image with out added problems associated with tweaking in USM.

Another idea to not induce a color issue from sharpening is to convert to the Lab model and just sharpen the Lightness layer via USM. Then you are getting even closer to the way USM worked mechanically.

Both of these apply to Adobe Photoshop only. Other products have similar or identically labeled features but I am thinking in PS-CSx or PSEx specifically. Lightroom has a rather different sharpness and resizing approach and now they've indicated they are going to go further with that approach in the latest Beta version.