Hello all, I am at a point where I am starting to sell prints of my captures. Most of my stuff are taken with a 5D or a 1DS2. I don't know crap about processing for print so for now, I will just get some software (like Alien Skin Blowup 2) to do it for me.
I am interested to hear how you do your prints (in details). Specifically, I am curious to know what you do from the PSD file with all your edit layers all the way to the final image right before you print. The largest print I would like to make for now is 16x24.
What am interested to know what is your step-by-step procedure to get it from the PSD to the final output.
Step 0: Make all the editing desired, Raw to PDS or TIF in 16 bit Adobe98. Save this base copy.
Step 1: I flatten all layers. I'm making a copy for print, don't need all the adjustment layers.
Step 2: Crop for composition. Check the PPI of the final crop. If it's greater than 150, I'm OK. If it's less than 150 I'm in
trouble. Have to decide whether to uprez or simply abort. Usually I abort. My threshold of 150 PPI comes from years of experience
and testing. Others will argue strongly for a higher minimum PPI. You should do your own test to decide who to believe.
Step 3: Apply Curves and Saturation adjustments that I've developed for the specific printers I use. In spite of a well calibrated
system, the final print never matches the monitor exactly. My curves/sat adjustments were developed after much trial-and-error testing on specific printers. For example, my home printer is an Espon 2200. For that printer I boost sat about 10% and leave curves alone. I use Mpix for lab prints. I don't make any curve/sat adjustments for Mpix. I print Blurb books. Every image going into a Blurb book gest a small s-curve adjustment.
Step 4: Convert to 8 bit and convert to sRGB if necessary. Images going to Mpix and Blurb must be in sRGB. Images going to my Epson printer are left in Adobe98.
Step 5: Perform "Output" sharpening, if any, but I rarely do any. I find my "Creative" sharpening in the base image is usually
good enough.
Step 6: Soft proof using the profile for the specific printer I'm targeting and look closely at the gamut warning. If there are
out-of-gamut areas, try to fix them as well as possible using masked curves and sat adjustments. But don't worry about getting them all out. Gamut warnings can be very misleading.
Step 7: Save as jpeg level 10 for Mpix and Blurb, save as TIF for Epson. Note that my PPI might be much greater than the mythical 300. I just leave it as is. I find that the various print drivers do a very good job of adjusting resolution and sharpening. As good or better than I can do using Photoshop to set a 300 PPI and re-sharpen.
Step 8: Print and review. 90% are acceptable as is. 9% require re-work and re-print. 1% look bad and just can't be fixed. Most common re-work is to add more "pop" with s-curve contrast adjustments or add more color with sat adjustments. On portraits, the most common re-work is to adjust skin tones. It seems to me that prints have the most trouble with reds and yellows not matching the monitor image. Most of my "bad" portrait prints are too red/pink.
Step 9: Show to people. 99.9% think the print looks great. But they are un-trained and unable to spot defects. 0.1% are trained and experienced photograpers and photoshoppers. They can pick nits that are invisible to even me.
Disclaimer: 99% of my prints are 11x14 or smaller. Most people seem to focus a lot of attention on resolution (insisting on 300 PPI) and output sharpening. My workflow pays little attention to these areas. But if I was printing 20" prints and larger they might be more critical issues.
Redcrown's workflow is very good, I will argue that with my Canon iPF5000 prints look better if I resample them to 300ppi at size. I do a lot of 16x20s and 16x24s and they look very nice. As Redcrown said, there are always folks who are better than you. Right now the folks whose opinion matters are artists, I know a couple of painters and they are hard on an image, you better have that composition right and have an interesting image. Photographers knit picking over minor tech details is boring to me right now.
redcrown your information is very useful to me so thank you for your time.
xfbo: i prefer to stay in ps, i did look at qimage and i think it might not work for my workflow, i want to be able to do all adjustments in ps.
tomm101: can you provide more details on how you resample to 300. is it the image size dialog in ps? if so, what exactly do you do there. so for example i have an image that is 10x16 native at 240ppi. what exactly do you do to make it 16x24, 300 ppi? thanks.
Your 10x16 @240ppi image would be 2400 x 3840 pixels. A image 16x24 @300ppi would be 4800 x 7200 pixels. That's twice as big, requiring a 100% or 2X magnification. That's a lot! Any resampling method you use will have to "invent" twice as many new pixels.
There are custom programs out there that do the job, and of course claim superior results. The most famous is "Genuine Fractals" and sells for $160. The Photoshop method is to use the Image/Image Size command. Enter your new dimensions and then choose "Bicubic Smoother".
You may find some old info that claims you should do the Photoshop Image Size in several small steps of about 10% to 20%, and resharpen the image between steps. That's old advice, based on old versions of Photoshop, before the Bicubic Smoother" feature was added. Now one big step is as good as many small steps.
But I say again, resizing may be unnessesary. If you print your original 2400x3840 image at 16x24 inches you will have 150 ppi. That's probably good enough, and will likely give you better results than if you resample to 300ppi.
Best to do your own test. Resample your image to 300ppi and save as a copy. Cut or crop out a small 4x6 portion of each image and make cheap 4x6 prints. Review each print closely and decide.
Qimage may be worth a look. It is only a print program. You still do all your edits in PS. It's claim to fame is a simpler interface than the Photoshop print dialogue, AND a much better internal resampling routine. It will do a better job of turning your 150 "pixels per inch" into the 1440 "dots per inch" that the printer requires.
redcrown, very cool, thanks. yeah my question was more theoretical and i just wanted to know the exact details, which you have provided. it sounds like to uprez, you use bicubic smoother (or 3rd party - i plan to use blow up 2) and you don't generally use output sharpening at the end for "small prints". in addition it sounds like as long as i don't go below 150 (or 180ppi), i should be good although i should strive for higher ppi if possible.
