Originally Posted by tkerr
The 'most monitors are 72 dpi' statement is dated, as it applies to typical CRT displays, not modern LCD panels. Your monitor is 1920 pixels wide, right? Measure the horizontal dimension and divide. My 24" desktop monitor is 1920x1200 and ~22.4" wide, so it's resolution is ~86 dpi; my 17" MacBook Pro display is also 1920x1200, but only ~14.5" wide, so it's ~132 dpi.
However, it's true that you can ignore DPI/PPI for most editing purposes, since print drivers and photo labs handle that all behind the scenes, so it doesn't matter whether you print/send a 16x10.7" image at 350 dpi, or a 78x52" image at 72 dpi - the printer will size and scale the image appropriately for the output size. The only time it really matters is when designing page layouts or submitting artwork to a publisher, who may have specific requirements (for example, in scientific publishing, figures usually need to be submitted as TIFF format at 300 dpi for images, 1200 dpi for line art/text, and 600 dpi for combinations like images with text labels).
As for the other parameters besides Output Resolution, Image Quality determines how much compression is applied to the final image. More compression means smaller files and lower IQ, you can see an extreme effect here:
The same 24.6 MB RAW file converted with the following settings result in the following file sizes:
- DPP, 16-bit TIFF - 126.3 MB
- DPP, JPG Quality 10 - 11.9 MB
- DPP, JPG Quality 9 - 8.9 MB
- DxO, JPG Quality 100 - 13.4 MB
- DxO, JPG Quality 95 - 3.9 MB
- DxO, JPG Quality 90 - 2.4 MB
So, it's pretty clear that DxO's 90% quality is not the same level of compression as DPP'squality 9. Personally, I left DPP at Quality 10 (but I no longer use DPP for routine processing). The file size decrement from 10 to 9 is pretty small. In DxO, I usually have Quality set to 95 - there's a substantial drop in file size, and for on-screen display the quality is excellent. For files I'm going to print at large sizes, I set the export quality to 100.
ICC = International Color Consortium. The 'Embed ICC Profile' checkbox adds metadata about the color space (sRGB or Adobe RGB) to the JPG file. Although previously this didn't matter much for web display since most browsers ignored the tags and just used sRGB as the color space. However, some current browsers (e.g. Apple's Safari 5) respect the embedded ICC profile and adjust the color to match your display, the idea being to represent the 'true' color (which really only works if you're also using a calibrated monitor). Click HERE for an example of the effect of embedding a profile (or you may just learn that your browser ignores the profile, if the image on the left changes as you mouse-over; clicking changes the gamma to match Apple's default 1.8 vs. Windows default 2.2). I think it's a good idea to embed the color profile - if you're having the image printed, it allows the lab to determine which color space was originally used relative to the color space of their printers, and if you're sharing it on the web, it will either allow the viewers' browser to correctly display the colors, or be irrelevant because the browser will ignore the tag, so no harm done.
The resize box just allows you to export the image at a smaller size, e.g. 800 pixels wide for posting here. If you're going to share photos, or keep them just for viewing on your computer, and you save the RAW files (as you should if you shoot RAW ;p ), you can consider exporting at a smaller size, e.g. 1800x1200, which means smaller files that are still large enough to fill most common displays (for now). Personally, I usually just export everything full size, and reduce using Photoshop for posting here or elsewhere. I would keep the aspect ratio locked, else you can independently change the width and height of the image, resulting in distortion.
If you're going to make downstream changes to exposure (brightness, contrast) or color, etc., then using a 16-bit TIFF as a go-between from RAW to Photoshop can produce better results, as you'll have the full bit depth to work with (even though as Daniel has explained in his snake oil discussion, 'full bit depth' does not mean the 14 bits that Canon advertises, it's still more than the 8 bits of a JPG image). However, 16 bit images are large, especially as youadd adjustment layers - adding 3 adjustment layers to that 126 MB 16-bit TIFF file mentioned above resulted in a 762 MB TIFF file. Editing files that largecan be taxing on your computer's RAM and CPU - my Mac's Core i5 with 8 GB of RAM does fine, but my work laptop, a Lenovo Core2Duo with 2 GB RAM, struggled when working with that 4-layer 762 MB TIFF file. Also, when I say Photoshop I mean the full Photoshop CS - you mention working with the file in PS Elements, which makes it a moot point since PSE can't really do much editing of 16-bit images - they can be resized, cropped, and have brightness/color, etc., adjusted, but you can't add layers (which you'd almost certainly want to do for editing) or even use the magic wand selection tool on a 16-bit image in PSE.
Hope that helps...
--John
Reply With Quote