> This is not surprising because of the way files are rasterised for > printing. For a 150 lpi (lines per inch) screen a 150 ppi (pixels per > inch) image is needed BUT when sampled for rasterising it is inevitable > that some pixels will be sampled twice and some missed! Result is poor > reproduction. Make the original image 300 ppi and you guarantee that > good sampling will result in the best possible resolution of detail in > print. In fact I think it take a very fussy printer with a loupe to tell > the difference between the result from a 225 and a 300ppi original when > printed at 150lpi. My own belief is that the 'quality' of an original > makes more difference than simple dpi.
But it is still a very subjective method of work for prepress and printers, who have to go spot on with results, specially to standardise their workflow and to save on costs. It makes an interesting set of tests though. Going down from 300 to 225 makes a difference in image size, specially with large files, and potential quality damage due to upsizing is lowered . > > Assuming the printers will use a 150line screen have you considered > preparing the file at 150ppi and then enlarging to 300ppi using PS > 'nearest neighbour' setting. My tests in PS suggest that this will avoid > the 'softening' of some details that bicubic can cause as it > interpolates. I have assumed as a law of nature that Bicubic is the only way to go and admit I have never even tried the other alternatives in PS. Don't even know what is the theoretical difference, andy hints will be appreciated, and I will no doubt run some tests about this. Photo Retocuh Pro also includes those options besides its proprietary upsizing engine, and I just wondered why, if no one use them? All the best. Jorge Parra APA/ASMP www.jorgeparra.com =============================================================== GO TO http://www.prodig.org for ~ GUIDELINES ~ un/SUBSCRIBING ~ ITEMS for SALE
