thanks pickles... I mean eforer!
that was very educational Probably ought to copy & paste it into it's own thread, but I'm going to remember how to get back here.
I violate several of your taboo procedures, like exporting .ai files (with embedded rastor componants) as .tiff files for 90% of my printing.
When you bring in rastor images as linked files, tweak all your sizes to your satisfaction & then record those final sizes...
If you resize with GF plug-ins in PS, do you have to carefully place each image exactly where the old one was, & then delete the old one... or do you just resave with the same name & let Adobe "update links"?
WOW!! 720x720 at 16 pass... and then you compromise all that quality onto
coroplast?? ...for such an obviously bright professional... that just sounds crazy to me. I guess I'm just not a fan of the coroplast flute texture
Thanks again for the time spent explaining all that. I'm going to go online & check ouit the cost of genuine fractals print pro plug-in now. I'm convinced that you know what you're talking about... so even if I don't understand it all, better tools will lead to better work, even if I don't understand everything about it at first.
Cloroplast was the customers bag, I just make the prints. You don't have to change the placement when you update links, it will leave your layout in tact.
Yes PPI is correct, DPI is the smallest dot a nozzle can fire and its a useless measurement more or less (made even more useless by variable dot size machines). The reality is, most people (incorrectly) use them interchangeably. Ask a graphic designer what PPI. Ya got me. That said
Javvila: A
raster image processor (RIP) is a component used in a
printing system which produces a
bitmap. The bitmap is then sent to a printing device for output. The input may be a page description in a high-level
page description language such as
PostScript,
Portable Document Format,
XPS or another bitmap of higher or lower resolution than the output device. In the latter case, the RIP applies either smoothing or interpolation algorithms to the input bitmap to generate the output bitmap. -Thanks Wikipedia!
While the above post simplifies some of the elements and ignores some. The basic principal holds true. Your printer has a finite hardware determined resolution. There is the "LPI" (somewhat of a misnomer but an old industry accepted term) meaning lines per inch. This is actually the number of discreet cells made of of the dots of ink a printer uses to describe a single area of tone. There is an inversely proportional relationships between the LPI (again cells of tone per inch) and range of tone. This is because you can fit fewer individual dots of color from each ink reservoir in each cell as the cell size decreases and the density increases. For example, if you have a printer with a true fixed dpi (and I mean real dpi) of 1440, and your LPI is 1440, you will only be able to express 7 tones with a 6 color system as only 1 dot will fit per cell. A larger cell (read lower LPI) can contain more dots to express that region of tone and therefore is afforded more range of color for each zone. Variable dot size is another means of increasing tonal range per cell.
Looking at the printer in terms of a pure density resolution (ignoring the aforementioned relationship between LPI or discreet cells of tone and dots of color, the colors that define the tone of the cell) the goal is to create the most direct relationship between the images density (ppi) and the printers density. When these numbers aren't one to one, or worse yet share no common factors, the RIP has to flat out do more work interpolating the file data to resolve it for the printer. I'm not saying a 1440 PPI image will rip faster than a 360, but if your printers LPI is the same as the images ppi, the rip has less work to do, so re-sampling your image (which like it or not the rip is doing if this ratio isn't 1:1) even 60 ppi can yield benefits in ultimate print quality due to removing interpolation by the rip and improve speed as the rip is doing less interpolation (work) to create the instructions for the machine. Furthermore, all RIPs due a crummy job of re-sampling as they pretty much all use a nearest neighbor method of filling in or removing data. The result, noise, jaggies artifacts etc etc. The methods used in Genuine Fractals are much more sophisticated and the re-sampling produces noticeable differences.
Finally, as far as the illustrator thing to rip goes. If you send it to rip as an AI file with embedded images resized in the file, the RIP is going to re-sample/interpolate it, and rips don't excel at that. Even a photoshop bi-cubic does a better job. Doing a resize with a non-re sample on raster art and sending it to the rip is more of the same. The RIP will interpolate it to meet the density of the printer. If you feed your printer something that is a 1:1 ratio in terms of density, or evenly divisible, you'll get better results requiring less processing power. End of story
1 addendum that just occurred to me. Even with the relatively standard wide format heads in most machines, the actual density of the machine is not super cut and dry. It is highly dependent on the range of dot sizes on variable dot machines, the number of colors available etc. What it comes down to though is that there will be a common factor that can be used that will work better. The cell size in terms of dots must be an integer as you can't have half a dot, so that limits things too. I have found that 90, 180 and 360 produces the best results (the factor of 12 thing breaks down at 96/90 as that droplet size when fixed can't yeild an integer from the heads on a Mimaki so interpolation uglies ensue). The trade off is sharpness and detail for transitional color smoothness. One of the reasons I only do 6 color is it lets me eek out a little more tonality from a higher density print. On crappy banners with big long gradients, I like 90 dpi as the tonal range per cell is higher and gradients look way smoother. Keep in mind the trade off is exponential too.
