"Robert G. Earnest" <robert@earnestphoto.com> writes: > Who shit in your Sake, Chandler? > Let me rephrase the question. Well, OK, I'll try to explain as patiently as I can. But you're about to rephrase a (fundamentally) meaningless question, so the result is likely to be equally meaningless. > I have a Canon G2 which I have been very happy with. My purposes for the > camera have never once included actually printing one of the images. Right, so if for example it produces [don't know actual figure] images which are 2400 x 1600 pixels, you can display these on a big enough screen, or display them on a street display of 2400 x 1600 coloured light bulbs. In any event, as long as you display all the pixels in the image, the amount of detail in the hair (or whatever) doesn't change. It doesn't make any difference whether the 2400 x 1600 screen is CRT or LCD, or the precise pixel spacing, and even if the street pixels are each the size of a television screen, you see exactly the same amount of detail. (Modulo factors like the lighting quality.) This should immediately suggest to you that it is not possible to measure image quality by how far apart the pixels are. > It had occurred to me that 72 ppi was probably not the optimum resolution > to print with as all the scans I have ever had anything to do with were > professional drum scans done at 300 dpi. >From the end: "drum scans at 300 dpi"??? Am I totally misunderstanding? I thought drum scanners were for negatives. Scanning a 35 mm negative at "300 dpi" gives you an image about 450 x 300, which doesn't sound very professional to me. Trouble is, perhaps they *do* say "300 dpi", by which they mean at a certain printed image size, the dots would be spaced roughly 10 to the mm. Well, if the size they have in mind happens to be 4 furlongs by 6 furlongs, this is a stupendously high quality image; if it happens to be 1 x 1.5 cm, it's a stupendously low quality image. This would be a stupid way to say how you're scanning, particularly since a 35mm negative is a physical object with physical dimensions, so it makes semi-perfect sense to say scan the 35mm neg at 2400 dpi. Only semi-perfect, since the negative is 36 x 24 mm, and it's pretty silly to insist on clinging to obsolete units for measuring it in. > So, I noticed, in a quick look that if I simply go under "Image Size"... Yes, but only "Image Size" in Photoshop. Photoshop may have some good features, but it has a brain-dead notion of what image size is. Photoshop isn't really an image manipulation program at all, it's a print preprocessing system, intended for graphic designers and the like, who are working entirely on (real, physical) paper. Even so, "Image Size" is the wrong term, since it seems to mean - "For the particular (notional) physically sized piece of paper in your hand, the size of the file representing it in information content terms." Or was it "change the size of the piece of paper, while not affecting the image in any way"?? Anyway, it's only there to confuse you. > A change to 300 dpi will afford me an image 7.573x5.68 inches with a > 2272x1704 pixel count. > > Now I would normally assume that this is the best way to do it. No. It might be the Photoshop way to do it, but that is never going to be the best anything. > So, with no regards to how long your Pi is or big your pixels are... I > will ask again. > Will not an image print better at 150 or 300 ppi than at 72 ppi? If you mean "Will an image which is 2400 x 1600 pixels print better at 300 ppi than at 72 ppi?" then the answer is "No, it will print smaller." If you mean "If two images are of different sizes and when printed, one at 300 ppi, the other at 72 ppi, come out the same print dimensions, will one be of higher quality?" then the answer is "Yes, the one which is a larger resolution - (300/72)^2 times as many pixels as the other one - will print with higher quality because it is a (pixel-)larger image." But this is really only tautology: the higher quality image will print with higher quality. > Is this in fact the best way to change the ppi of an image? Or dpi if > you are considering, as I do, each pixel to be a digital dot. So, again, we have to ask: What do you mean by "changing the ppi of an image". Clue: the jpeg format (for example) has (UIMM) absolutely no information in it called "ppi" or "physical dimensions", any more than my signature at the end of this plain text email has a font size. I actually set PaintShopPro to a "pixel spacing" of 1 pixel per cm; this eliminates confusing and meaningless numbers everywhere. No-one has yet complained that my images are (n/2.54)^2 times worse than theirs are. Actually, I could change the "ppi" of my last image submitted to the gallery by telekinesis: I just _think_ it 25400 ppi, or 0.7 ppp (pixels per parsec). Here's a useful web reference: http://manual.gimp.org/manual/GUM/Migrate.html It's a comparative review of The GIMP and Photoshop, but the part of interest is where they simply have to explain that Photoshop terminology and interface presentation is out of kilter with GIMP (and with everyone else, of course) everywhere it talks about "image size" and related issues. Brian Chandler ---------------- geo://Sano.Japan.Planet_3 Jigsaw puzzles from Japan at: http://imaginatorium.org/shop/
