Samples to show what can be done with images from the RIcoh GX100.

Sized to print at 12 by 18 inches at 300 DPI, so they are BIG.

Produced from 9 MP 3:2 format DNGs. Converted using Camera Raw 3.6, upsampled to 4096 px on the long axis during conversion. Imported at 16 bits, manipulated using Photomatix hdr plugin, curves, smart sharpen and a midtone sharpening action, finished off using neat image if too noisy, followed by PTlens to correct (slight) distortion, and upsampled to final size using Genuine Fractals.

(note added several years later--these days I would quite likely have used a different resizing plugin; Genuine Fractals works well for some images but introduces artifacts in others. At present I find that Resize Magic works best the most often, but on anything important I will try several different resizers and choose the one that works the best for the present image)

Please look at, download, even print the photo(s), but do not redistribute them in any way, they are, as each says, copyrighted. To see them full size you will need to select original size from the list below the photo.

By the way, remember that if you are looking at the images from DNGs on a normal 72 dpi monitor in a browser that shows them at 1:1 you are looking at each pixel at 4x linear (16x area) magnification, compared to its size in a 12 by 18 print. It is equivalent to looking at a 4 by 6 foot print from whatever distance you are at. Yes, you can see a bit of grain in the sky, and yes there is some noise in the water in the image with the trees, particularly in the tonemapped from DNG version. However, in a 12 by 18 print, neither is noticeable. To see what a print will look like, you can get a pretty good idea by displaying the digital file on your monitor at the physical size it will print at. For classic 72 dpi monitors, that means looking at the digital file at about 0.25X. For many laptops, with about 150 dpi screens, that means about 0.5X. You can calibrate this in the settings in photoshop; then the "print size" option for the magnifying glass will actually work.

I have received some comments from people saying, in summary, "why bother upsampling to make big prints. Upsampling doesn't add any real detail. Leave the images at native resolution and let the printer worry about it."

If you think that, I'd like to point out that, for example, laser photo printers like the ones at Costco that I make 12 X 18 prints on, print at 300 dpi. Other sorts of printers have their own native resolutions. If you do not upsample your files to the native resolution of the printer, it happens anyway, but the printer or printer driver does it for you, and you lose all control over what resizing algorithm is used. If you have applied sharpening, any sharpening halos get upsized, etc. etc. It is far better to control these things yourself.

Costco, by the way, prints using photographic laser printers and allows you to specify no automatic color adjustments and provides printer profiles for each printer at each store, so it is possible to make very nice prints if you exercise control. Just a satisified customer who wishes they would post them to Australia, as it is I do big batches when on trips to the USA.

I have also received comments along the lines of "my pocket camera shoots at 72 dpi. Changing that resolution (say to 300 dpi) cannot add any detail..." This reflects the fact that resolution can be a confusing topic. Any digital camera produces native images with a certain number of pixels in the horizontal direction, and (usually) a different number in the vertical direction. For example, 6 mp DSLRs tend to have about 2000 X 3000 pixels. The image is then tagged with a "resolution" which is completely arbitrary. It is just filling a field in the exif information. Many camera makers assume you will look at images on a monitor. Old CRT monitors tend to have 72 dpi resolution, so they put 72 dpi in the exif information. Many DSLRs have different defaults, and many let you let you change the default. This information is only a guideline to programs that are displaying the image. A 2000 X 3000 pixel image tagged as being 72 dpi has no more or less photographic information or resolution in it than a 2000 X 3000 pixel image tagged as being 300 dpi; the only difference is that a program that pays attention to the exif information will try to show you the one tagged as 72 dpi at a size of approximately 28 X 42 inches as it appears on the monitor, and will try to show you one tagged as 300 dpi at a size of 6.7 X 10 inches.

Ideally, to exercise full control over your images, you need to do most work with them at their native pixel size (the "resolution" they are tagged with is meaningless). When you are going to print them, or put them on the web, decide on what size you want, resize them using the algorithm that produces the best results for that image to the correct pixel dimensions to be that size at the native resolution of the device they will be displayed or printed on (for example 12 X 18 inches on a 300 dpi Fujitsu laser/photochemical printer means a 3600 X 5400 pixel file), do final output sharpening, set the file's exif information to reflect the dpi of the output device, and you're done. That resizing is going to happen, whether you do it or you leave it up to the device to do it. If you are leaving final resizing up to a printer or screen driver, the absolute best case is that the output is as good as you could make it if you optimized it, but it usually will not be as good.