The April 8, 2002 issue of Sports Illustrated magazine offers a chance to examine the printed quality of photos from the Canon EOS D60 and EOS-1D. More importantly, the printed appearance of each offers a chance to ponder the importance of workflow in the life of a digital image destined for ink-on-paper reproduction.
EOS-1D Photo by David Bergman
EOS-1D, ISO 200, Color Matrix 4, Raw .TIF original
The crisp, colourful photo that graces pages 48-49 of the magazine is a testament to the image quality of the EOS-1D. Image quality that's obtainable if the camera is configured optimally and in the hands of a photographer who knows digital. David Bergman followed Sports Illustrated's standard EOS-1D protocol, shooting RAW+JPEG on the camera's wider-gamut AdobeRGB colour processing setting, Color Matrix 4. For printed workflows, best quality comes from locking the EOS-1D on Color Matrix 4, despite the fact that the resulting images don't have the same punch on screen as those shot on, for example, the popular Color Matrix 3 (sRGB) setting.
Normally, SI photographers are expected to transmit the RAW .TIF originals, using the JPEGs only to speed the editing process in the field. At this event, deadline pressures forced Bergman to convert the RAW originals to JPEG (he prefers the quality of a Photoshop-generated JPEG over an in-camera JPEG) for transmission.
D60 Photo by Mike Sturk
D60, ISO 1000, Large Fine JPEG original
Cowboy poker at Rodeo Royal here in Calgary, Canada is the subject of a great Mike Sturk photo shot on the Canon EOS D60. Sturk, whom readers may know as the moderator of the robgalbraith.com forums, shot the Leading Off frame at ISO 1000 on the D60's Large Fine JPEG setting. The lighting conditions for the annual event are notoriously bad, with fill flash a must. The original, 18MB (when decompressed into Photoshop) image is fairly sharp and fairly noisy with okay colour.
The photo followed a circuitous route through Canadian Press (CP) and AP before it arrived on the magazine's doorstep. In Photoshop, Sturk downsampled the picture to the CP-mandated dimensions of 10 inches wide at 200 dpi before transmitting it as a Photoshop-saved JPEG. The resampling knocked down the uncompressed size of the photo to about 7MB. It also knocked down the detail level of the image to well below what the camera captured. The processing steps after that are unknown, but may have included additional JPEG compression steps, heavy image massaging to reduce the visibility of noise and more.
Whatever those steps may have been, the result is a version of the photo in print that is softer and murkier than the original. As a result, it may not be a fair representation of the D60's capabilities.
Taken together, Bergman's 1D photo and Sturk's D60 photo illustrate the importance of optimizing the digital workflow when the goal is to match, on the printed page, the quality captured by the camera.