Riddle Me This

[Tech]

Can anyone explain why is it, when printing to postscript [in Quark and Distiller but not in Indy], resolution is measured in "dpi" and not "lpi". For example 2400 dpi for X-1 preset seems excessive when magazines print at 133 lpi.

I did a test, as soon as I reduced this 2400 dpi to say 1200 or 600, I can generate a smaller file size pdf and still keep images downsample to 300 dpi. I'm concluding the 2400 dpi controls the range of data being pack similar to a color profile? Can anyone confirm this or tell me I'm out of my mind?

 

[WiseGuy]

dpi is the resolution your printer is using to make dots on the plate (or paper). LPI is the number of dot lines it is making for your press. Each LPI dot is made up of thousands of printer dots (or spots). The more resolution, the smoother or crisper the LPI dot.

Resolution of the image is typically 300dpi ( or about 2x the LPI) so that you have enough data there to make good LPI dots. Much more than 2x is really a waste.

So, printer dots are made of spots. LPI halftone dots are made of printer dots.

I hope this helps.

Michael

 

[gordo]

Can anyone explain why is it, when printing to postscript [in Quark and Distiller but not in Indy], resolution is measured in "dpi" and not "lpi". For example 2400 dpi for X-1 preset seems excessive when magazines print at 133 lpi.

I did a test, as soon as I reduced this 2400 dpi to say 1200 or 600, I can generate a smaller file size pdf and still keep images downsample to 300 dpi. I'm concluding the 2400 dpi controls the range of data being pack similar to a color profile? Can anyone confirm this or tell me I'm out of my mind?

It's been a long time, but, if I remember correctly the dpi setting controls the "flatness" of graphics in PostScript. I.e. when a vector graphic is rasterized (RIPped), curved lines and edges are approximated by a series of short straight lines. Ideally, the length of the lines would be only one device dot long, but this is seldom the case. If it were, every curve would be perfectly smooth, but would require a huge amount of processing time and memory to rasterize. By using short straight lines to approximate the curves, Postscript saves a bunch of time and memory at the RIP stage. The mechanism for controlling the length of these short lines is called the Flatness Control.

Flatness is defined by the mathematical equation:

Flatness = Printing device resolution / Output resolution setting

This may be why the dpi info is required when printing to PostScript.

best, gordo

 

[Lukas Engqvist]

I remember it as Gordo says. If you had a too low dpi you would get octagonal "o"s etc. The flatness setting is still there in Photoshop if you look at clipping path. There is a way to get to it in PDF, but it's ages since I saw any PDF that suffered from flattness, but guess that's cause it mostly comes from Indd

 

[Tech]

When was the last time someone told ya both are awesome? Gordo's explanation makes perfect sense and at 250dpi setting, I could start to see a hint the "o" pattern Lukas mentioned.

We are experimenting with pdf ebooks for tablets but the biggest problem so far has been controlling image quality and file size. As I discovered, using this lower "flatness" setting via distiller does a better job than the typical jpeg down sampling [even better than export from Indy...I know it's hard to believe]. The hint of pattern is a none issue right now on 7-10 in tablets. Even if one notices it, when zoomed into the image, it still appears better than the sad artifacts from jpeg compression.

** found this webpage with same explanation for anyone interested**
http://www.ps.missouri.edu/ps2/support/TutorialFolder/flatness/index.html