On Tue, Nov 23, 2010 at 11:10 AM, Eric Weir <[email protected]> wrote: > Regarding converting to DNG, I will now expose my naiveté by [1] admitting > that, again for the time-being, I am shooting jpeg, [2] asking what's the > advantage of DNG. > > Though my mastery of file management is minimal I sense that I may be making > my first forays into editing soon. E.g., I did a lot of shooting at my > camera's highest iso setting at a community pot-luck the other night; the > images are fine till you look close; I'm now wondering if the instructions in > the help file for cleaning up noise in high-iso images might help me improve > the quality somewhat.
DNG is just a standardized raw file format using a specification published and maintained by Adobe. Native Pentax raw files are .PEFs. So what you're asking is "what's the advantage of raw vs JPEG?" In two words: editing potential. In the context of a Pentax DSLR, a JPEG file is comprised of 12-bit sensor data (the primary contents of a raw file) which has been demosaic'ed, gamma corrected, chroma-interpolated into RGB channels, had various filters for white point, sharpening, contrast, and saturation applied, then reduced to 8bit per component channel and compressed. Each of those steps from capture to final file is a one-way transformation that discards some (hopefully an unimportant some) of the original data ... by the time the file is a JPEG file, more than 50-60% of the total data has been discarded. If all the capture and image processing settings on the camera are correct, you get a fine finished image in a compressed, 8-bit per RGB component file to print, view and distribute. The issue is that the image processing settings on the camera are somewhat coarse in the adjustment they provide and most of the processing algorithms have only limited leeway so it is difficult to exploit the most that the sensor can do when faced with difficult conditions. However, JPEGs out of the camera can be very good if everything lines up correctly and you make the right settings. But if you are intent on learning how to process images, 8bit compressed RGB data is "fragile": it's easy to make even a couple of small adjustments and run out of the range of adjustment possibilities. That's because each channel can only represent 256 different luminance values, slicing the intensity scale from darkest to lightest values only 256 ways. Raw data, on the other hand, allows you to take the much more powerful computer at your disposal and set all the adjustments for the processing steps above with much more fine-grained control. The data represented has 4096 different luminance values in any channel, the processing curve can be bent and shifted a great deal farther before data clipping happens, and you can simply get a heck of a lot more out of your exposures. You also have much finer and more specific control over the sharpening and noise filtering operations. When you capture raw format and do the image processing after the fact, you're in essence shooting more like what you did when you used a film camera: you worry about focus, framing and getting the exposure right, not how much sharpening to apply, what white point you want to be using, how much saturation and contrast you want, etc. Aside from the greatly expanded image processing potential of raw capture, it lets me focus on the image taking tasks much more single-mindedly and leaves the rendering tasks and decisions to another time. Dealing with underexposed and noisy image files isn't easy in either case, but you have lot more data to work with and a lot better control of how it is rendered if your original data is a raw capture. -- Godfrey godfreydigiorgi.posterous.com -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net to UNSUBSCRIBE from the PDML, please visit the link directly above and follow the directions.
