Blog Of Rick Pepper

Here’s a tidbit related to resolution and pixels and what it all means. A mistake that can be made in Photoshop when resizing images is, if you’re just looking to change the resolution from 72 dpi to 300 dpi, then here’s how it should be done. This does not actually add more data, it just re-arranges the sizing information, which allows you to directly see what size your image is (height and width) at 300 dpi resolution, making it simpler to tell what you’re going to have in the end, if you need a 300 dpi image. For offset print (i.e. use in QuarkXPress or InDesign) you should be re-sizing your images to the desired print resolution anyway, and then scaling them to the correct print dimensions (by checking the “resample Image” option) even though RIP performance is light-years ahead of what it was in the early days of PostScript, this is still best practice. So here are some screen shots in Photoshop showing what happens when you use the proper method to re-order an images resolution, note the chain linking the inch dimensions to the resolution – and that the pixel dimensions are locked – they won’t be changing. This is done by un-checking “Resample Image”.

Our first example is at 72 as it came off the camera:

Now, changing the same image to 300 dpi:

Lastly, just to illustrate that all we’re doing is defining the number of pixels in an inch – converting from pixel dimensions into inches (in this case) – 1 dpi:

See what happens? The inch dimensions match the pixel dimensions, because, we’ve told it that there’s 1 pixel in 1 inch.
Another thought along this line, when shooting in RAW mode, and using Adobe’s RAW Import (or any other camera RAW software for that matter) there’s an area where you can define the resolution, that does not magically create more pixels, it just does the very thing I’m illustrating in these screen shots – re-ordering the dimension information. When done correctly, this operation is absolutely instantaneous. You’ve done it wrong – typically fabricating pixels, if you see a progress bar or it takes a few seconds. To change resolution, then scale an image you will use this dialog twice, the first time with “resample image” turned off, the second time with it turned on.

In just the 2 hours I’ve been writing posts for this blog I’m convinced Blogging is an out-growth of the spreading of A.D.D. in our world. It’s soooo easy to just throw some thoughts and images up on the web (which is the whole point, duh…) But, given that fact, here are some really quick thoughts on digital photography and the associated workflow, which to me is the crux of the whole topic.

If you are only planning to make basic 4X6 prints and the occasional 8X10 – or use them on a website, anything over 3 megapixels is enough for you – but hang on for some heavy math to help you figure that out for yourself ;~)

If you want to be able to submit photos for a school yearbook or anything involving traditional offset lithography (a.k.a. four color printing on a real printing press) then you’ll need at the very least 3 megapixels, and if you want to be able to print these (in said yearbook) larger than 4X6 then I’d recommend 6 megapixels or greater, and those figures don’t leave you with much room for cropping. Here’s how the math works to figure this out: For traditional screening (a.k.a. AM – the most common, still) you need twice the resolution of the line screen your printer (as in a printing company, not your desktop inkjet) is printing at. For the sake of discussion the most common is 133 lpi, so figure 300 dpi (literally; 2 X 133 = 266 dpi), at the print size. So, if you want to print a 4X6 image at 300 dpi that’s 4 X 300 for the width, and 6 X 300 for the height, in pixels. That translates into meaning you need 1200 X 1800 pixels to print said image at 133 lpi in offset print (up to 150 for best quality, 175 would be allowable.) In megapixels that’s (1200 X 1800 / 1,000,000 = 2.16) – 2.11. So, back to that 8X10 print. I’ve gotten varying answers from photo labs, but from experience I’ve gotten good results from both the Fuji Frontier systems (Walgreens, Wal-Mart and Sams) and Kodak systems with as low as 150 dpi. Translate that; 4X6 would have to be at least 600 X 900 pixels (4 X 150, 6 X 150.) Understand that regardless of what resolution your photo editing app (or camera software) is set for, if the image is 600 X 900 pixels, it’s good to go. How pixel width, height and resolution (depth) relate is for another post, but all you need to really know is the actual pixel dimensions. 8X10 would require 8 X 150 = 1200, 10 X 150 = 1500 (1200 X 1500 pixels), which comes down to 2.11 megapixels, minimum. The one photo finisher in town told me basically 200 dpi, so that would mean 1600 X 2000 pixels for an 8X10, which is 3.2 megapixels. What he said was 3000 X 2000 pixels was the max on the Fuji system, and that would be for a 15X10, equaling 200 dpi.

That’s all pretty straight forward, except; Most point and shoot digi-cams shoot using a 4:3 aspect ratio, the same as standard video. So, for the typical 2.11 mpcamera, with 1600 pixels wide (that’s the 4 in 4:3), if we divide by 4, we get 400, then 400 X 3 = 1200 (which is the 3 in 4:3.) That’s the resolution of a typical 2.11 mp cam like the Olympus C-2020Z. So, back there on our calculation for an 8X10, needing 1500 pixels for the 10 part of the 8X10 (that’s 4:5 aspect ratio) that would mean there would be next to no room for cropping – you got 100 pixels to throw away on the one dimension, none on the other – and that’s only for the bare minimum, IMO, 150 dpi. Some point and shoot digi-cams offer a 3:2 aspect mode, such as the Olympus C-730UZ. That’s a 3.2 megapixel camera, so at 3:2 you get 2048 X 1360 pixels, or enough data to do 13.65 X 9 at 150 dpi. In 4:3 mode that camera yields 2048 X 1536 pixels. 3:2 is the same aspect as 35mm film, and the same aspect of a 4X6, though to be correct in page layout speak that would be a 6X4 since the first number is always the width, and the 3 in 3:2 represents the width. As always YMMV, especially with inkjet printers.

This illustrates one other point I’ll make in closing; give just a little forethought to how you compose a shot. If you know you only want it for 4X6 prints, then drive on. But, if you want to make sure it will work for multiple sizes/aspect ratios, then be sure to leave some room in both directions, in other words, don’t crop the shot to tight (like me.) It simply means this; when cropping an image shot at 4:3 aspect for use on a 4X6, you will be cropping off material on the top and bottom of the image (assuming landscape orientation in both cases.) If you shoot using 3:2 aspect, and want to use it for video at 4:3, the reverse is true – you will be cropping off the left and right of the image. These examples are assuming that you want to use the entire image right up to the edges.

Here are some visual aids for this concept. This image was shot wide enough that it will work at almost any aspect, but note the last example where, if shot at 3:2, and you want an 8X10 out of it, how much width you lose, the source image is always 600 pixels wide before “cropping.”

Here’s the original 4:3 image (above).

Now cropped to 3:2 for a 4X6 (ghosted areas show what would be cropped.)

Now cropped for 8X10 (ghosted areas show what would be cropped.)

Lastly, here it is shown as if it was shot at 3:2 originally, then cropped for 8X10. Hopefully this gives you some idea how these all relate to each other.

So much for the quick blurb and A.D.D..