I know someone’s going to bark at me to read other posts, so before you do, know that I have. Most of them deal with people not understanding resolution, resampling, and rescaling.
I have a simple question which (in my perfect little world) should — i hope — have a simple answer.
If i’m outputting to an imagesetter, I ask my print vendor what line screen they’re printing at, multiply it by 2, add a little extra, and VIOLA, I know what to set my Photoshop resolution at:
2 x LPI + a little extra = PPI
2 x 133LPI + a little extra = 300PPI
However, what about if I’m outputting to an inkjet printer? It doesn’t use halftone screens, so I have no idea what I need to res-up to. There’s gotta be a threshold where all I’m doing is adding excessive pixel data without a visible increase in image clarity.
Technically, I’d be outputting to an Epson 1280 or 2200. I imagine the answer will have something to do with the DPI capabilities of the printer itself, but I need to know the mathematical relationship between the DPI of the printer and the PPI of my Photoshop doc.
I teach this stuff, so, please, tech jargon is welcomed.
Learn how to rescue details, remove flyaways, add volume, and enhance the definition of hair in any photo. We break down every tool and technique in Photoshop to get picture-perfect hair, every time.
G Ballard has some pretty good advice, too, I suppose: just test it out.
So, as a general rule of thumb, anything lower than about 175ppi is going to be noticeably pixelated, and anything over about 300 is going to be overkill?
As far as over 360 ppi being overkill, if for some reason I had an image that was over 360 ppi, I certainly wouldn’t down-sample it. 🙂 But there is no reason to up-sample an image beyond 360 ppi for output on the Epson 2200.
With the new Bicubic Softer capabilities of Photoshop 8, I’m finding myself up-sampling old images with more confidence than before. I guess it depends on how the up-sampled image looks on the screen.
It’s not being used to print billboard tiles heheh
I’m using it to print pages for my design / layout portfolio, so a good mix of text, line art, and photography. Viewable at normal reading distances.
Ramón — agreed. The shots of my stuff were taken with an 11-megapixel Nikon, so they’ve got plenty of pixel data, but I’m going to be blowing them up to fit 13 x 19 pages, so I think I’m going to have to do a bit of upsampling.
You can prepare your files at 300 ppi for Epson inkjets and not worry about other resolutions unless you really want to split hairs.
I don’t know why this topic gets some workers so passionate, but there have been real flame wars about this. One "authority" claimed that 260 ppi was the maximum EVER needed for 1270s, 1280s and the 2200 and that no further improvement could be seen using higher image resolution. In fact, he insisted, 240 ppi was generally all the Epsons could utilize.
I can see improved quality at 360 over 300, but it’s because I know which prints are which. I have NEVER been able to get anyone else to distinguish between prints made at 360 vs 300. So I use 300 most of the time. 360 takes much longer to print. 300 ALWAYS beats lower input in my prints, so I don’t go lower.
As far as resolution for press, there is nothing to be gained by adding any extra resolution to the 2x lpi formula. Since most commercial work is 150 lpi, 300 ppi is all you need to prepare and that covers the inkjet also. Very convenient. In fact most imagesetters can produce the same quality at 175 lpi from 300 ppi images.
If, and this is a big if, you prepare your files correctly in terms of sharpening, you can indeed tell the differences between various output resolutions. Determing the EXACT amount of output rez you need is dependent upon the native resolution of your file-uprezing will not help output, your image type-strong high frequency texture will use higher resolution to better keep detail, and you intended viewing distance-which is generally 2X the diagonal of your print size.
Also, it’s useful to send images to the printer that are exact divisibles of the print head resolution such as: 180ppi, 240ppi, 288ppi, 360ppi, 480ppi.
The resolution you need also depends on your media. Matte or watercolor paper can not hold the detail of glossy or semigloass coated papers. Printing small prints (5×7) on glossy with proper sharpening, you can tell the difference between 240, 360 & 480ppi on a high quality coated paper printed out at 2880dpi on an Epson 2200. You will need to sharpen matte output more/different than glossy output.
Again, you need to test out YOUR images at the various resolutions and YOUR images to determine of the higher resolution of the images and the printer is worth the larger files and longer print times.
On the other hand, I often use 180ppi @ 720dpi for 30" x 40" prints on my 9600 on watercolor because the high rez on watercolor paper is fine-unless you look at the print from 6 inches away. . .
"As far as resolution for press, there is nothing to be gained by adding any extra resolution to the 2x lpi formula."
Every printer I’ve ever talked to has asked me to include a bit more pixel data than actually needed. I’ve always chalked this up to a few variables:
1) In case a client wants to make a slight enlargment at the last minute, the is pixel data is there to use as a buffer.
2) In case a photo needs to be moved to a different position and needs to have room for a bleed, you can enlarge it slightly and, again, use the extra pixel data to do so.
3) A LITTLE bit extra never hurt anyone… especially with the fast computers in production houses these days. I mean, I’m not advocating scanning at 500ppi for 133lpi or anything, but … I’ve always got a few extra bucks in my wallet just in case.
* shrug*
Splitting hairs, I suppose.
Thanks for the advice on my ink-jet question. You guys are really helping.
I use 720 when I want the best quality. The number came from Epson and I can see the difference between a 2880 print that was sent to the printer at 360 and a 2880 print that was sent to the printer at 360. While the difference isn’t spectacular it is visible.
I agree regarding preparing images at exact divisions of the printer’s resolution (288, 360, 720). I KNOW I can see improvement all the way to 720 if I look VERY close and hard at my own images. (But it DOES take a long time to print at 720!)
I agree that prints on glossy stock should show more detail than on matte stock, but in practice I don’t find that actually plays out comparing Epson matte stocks (several types) and glossy stocks (several types). And in any case, proper sharpening for the matte print removes any visible difference.
While higher (real) resolution DOES result in finer detail in images that have the quality in the first place, can anyone really claim that viewing prints at a "normal" viewing distance for the size, you can see improvement beyond 300 ppi?
I just can’t. And I’ve conducted all kinds of tests with coworkers and customers. At normal viewing distances (10 inchies or so for an 8.5 x 11 and greater as the size increases) no one has ever been able to consistently pick out the higher ppi image.
Yes, I print all my favorite images at 360 ppi on a 7600 (derived from much higher initial resolution). But that’s just because I want to and I can honestly say to the customer that the product is high quality (at least in that regard).
When trying to squeek the max out of the SP2200 for some special situation I use 240ppi, 288ppi or 360ppi depending on size/viewing distance. However that is very much the exception. 99% of the time images print just fine using 300 ppi as a default for almost all projects. The workflow benefits in house of a standard 300 ppi choice throughout are huge, and most print shops like it as well.
My testing corroborates Venicia’s observations. IMO the workflow benefits transcend tiny ppi technicalities.
I recall reading an article in Digital Photo Pro(?) stating that the need to use a ppi that is divisible by the output resolution to be a "myth". The article made sense at the time but I’ll be damned if I can remember anything in it… maybe someone else?
FWIW, I could never see any improvement in my prints beyond 300. (890 @ 1440 and 2200 @ 2880).
In theory, you should see no benefit sending more than 360 ppi to an Epson through the Epson driver360 ppi is the ‘native’ resolution of the printer. In practice, you may see benefits at certain ‘magic’ higher resolutions480 definitely being one of them*if* you sharpen to take advantage of that resolution.
With RIPs that bypass Epson’s screening such as the Colorbyte or Colorburst RIPs, I very much doubt that you’ll see any difference at resolutions higher than 360 other than the image taking longer to print. ImagePrint downsamples the incoming data to 360 ppi if it’s higher than that, pretty sure that the other RIPs do the same.
I do a fair amount of ink jet prints—8×10 and smaller—as my final pieces. I get fine results printing at 300ppi on Epson 1270.
From what I’ve seen and heard there are more effective means, but ’til I have the time to research and learn them "Unsharp Mask" is my prefered method of sharpening.
For expert level sharpening techniques tutorials in Photoshop, try here: <http://www.pixelgenius.com/sharpener/> Jeff Schewe and Bruce Fraser, both helped you here, are pixelgenius.
I recall they offer a free demo trial there for download.
The problem with output sharpening has been that it’s pretty much impossible to determine, on screen, exactly how much and what manner of sharpening is required for printing. The common wisedom in the past has been to look at the image @100% and make it look "a little too sharp". The problem is, what does a "little" look like?
I do look at an image @100% to evaluate WHAT is being done to an image’s pixels, but a better gauge of how much to apply is a 50% zoom in Photoshop. There, Photoshop is doing pixel dithering-not unlike a printer will do-to show the image on screen. At 50% there are four image pixels dithered to make one screen pixel. As to "how much is enough" that’s a tougher call.
There are a variety of sharpening methods, from edge protected USM to luminosity USM to High Pass/Overlay. All these methods come into play for a properly sharpened image. How you sharpen for print is dictated by how you did your first rounds of sharpening, local sharpening and how you’ll be preparing the output.
Some people think that just providing a recipe will help. Actually not. It’s the exact numbers at specific resolutions and for specific media that you need. Those numbers are a lot harder to determine.
The problem with output sharpening has been that it’s pretty much impossible to determine, on screen, exactly how much and what manner of sharpening is required for printing.
I share the same sediment.
I pretty much over sharpen images, (visually) at the pre scanning stage, to push the image because, sharpening is first off, subjective, and 2nd, sharpening, in general, doesn’t "visually" replicate digitally compared to actual print.
Things that contribute to this are output devices, monitors as well as software.
yes,
even applications show sharpness in a different light. Captures that I thought were true junk, faired pretty well when printed to a contract proof, such as a Creo or Kodak Approval.
To get back to the main topic – best resolution for Epson output – I can add some specific information for having quite thoroughly tested the whole thing. I deal with lenticular pictures, and a lenticular is just like a lens, like a magnifying glass placed on the print. SO, I can vouch : 1/ Definitely best output resolution is a multiple of 90 : 180 – 360 -720 -1440 ppi. Other resolutions will cause banding if you have a nice gradient ……… 2/ On lenticular, you can very clearly see the constant improvement as you res. up from 360 to 720, .. and even to 1440 ppi. Paradoxally, the best Epson printer I know is also the cheapest : I print small prints on a 915, with files at 1440 ppi for an output at 5680 dots of ink per inch .. and sorry, its damn slow, but the difference is visible. 3/ General rule I use is deviding inkjet announced output in "dots of ink per inch" by four to get the file resolution in "pixels per inch". So best output on a 7600 or 9600 at 2880 dpi will lead to a 720 ppi file. Mind this understands you do have the existing information in the original file at these resolution outputs. I dont res up files just for the pleasure. I just avoid resizing down. But I admit we lenticular picture addicts may be the only ones to see the difference.
I don’t really know what all this fuss is about. Resolution, schmezolution! I just print all my stuff on an original Apple Laserwriter. If it was good enough for Woz, it’s good enough for me! My customers don’t complain.
Like the good folks at Apple said about the SE when Amiga came out with their model 1000 and its color graphic monitor, "Designers don’t need a color display to do design."
Inkjets do use the equivalent of a screening technique, Epson uses a proprietary stochastic. This is the reason for the high dpi numbers, they use the dpi to create a mini-image setter. you can get a rough idea of the equivalent line screen by calculating how many dots form a cell and how many ink levels or dot sizes they use. I don’t think any of them go much above the equivalent of 133 lpi, although Jeff or Bruce may have a better idea.
Jeff and Bruce and VL, thanks very much for your useful input.
I do look at an image @100% to evaluate WHAT is being done to an image’s pixels, but a better gauge of how much to apply is a 50% zoom in Photoshop.
Right on Jeff, I’m surprised more people haven’t talked about this. The 50% viewing ratio on CRT displays for images that are already sized to print at 300ppi is perfect! Because this has been such a dead reliable formula for over ten years in my workflow I hate to leave 300 ppi even though other higher resolutions will provide better print detail.
I’m trying to come up with a similar formula for LCD displays (ie what viewing percentage simulates print sharpness best for files at X resolution) …
Some people think that just providing a recipe will help. Actually not. It’s the exact numbers at specific resolutions and for specific media that you need. Those numbers are a lot harder to determine.
How about recipes that include a resolution variable? For example:
Amount: variable, use 400 as a starting point for film scans, 180 for digital capture Radius: resolution in ppi divided by 360 threshold: variable, use 4 as a starting point
The key portion of this recipe is the radius. For 150 linescreen printing you would have 300ppi files. 300 divided by 360 = 0.8 For 85 lpi newsprint printing 170ppi. 170 divided by 360 = 0.5 72ppi images for the web? 72 divided by 360 = 0.2 406.5ppi Lightjet prints? 406.5 divided by 360 = 1.1 360ppi inkjet prints? 360 divided by 360 = 1
This may not be a perfect formula but it can be a helpful guide for new users.
I don’t mean to attract attention from the excellent Photo Kit sharpener which we should probably be talking more about.
Henri – I agree! Try it and look at the results. And I can also see the difference with my naked eye. Run some paper through your printer and get to know it!
I recently downloaded about 300 jpgs from a commercial site. The only thing that had been done to them for the web was that they were all at 72ppi. I adjusted about fifty of them in Photoshop as a sample. I had to use Curves, Color Balance , and USM on all 50. They required increased midtone contrast by about the same amount, and similar color balance changes(less red, a little green, and some yellow). I attribute this to the monitor calibration of the source. The USM adjustments were all over the place, though. This I attributed to the originals and not the scans. I used everything from .25 to 2.0 radius and 100 to 500 %. I kept the threshold at 4. The adjustments were made at 50% zoom. A few of the jpgs had to be gaussian blurred because they were jaggy on smooth curves at 50% zoom. I am guessing that these were underscanned. The physical sizes at 72 dpi were all over the place also. If I make them 240 dpi for inkjet, I will do the USM after changing the size.
So lets say I have an image that I want to print out at 8×10. The resoluiton is 380ppi. Since it is recomended to print at a res evenly divisable by the printers res, is it best to down sample to 360ppi or to just go ahead and print it at 380? So which is worse, downsampleing or printing at some odd res such as 325.125ppi?
Since Epson, HP and the rest of the print manufacturers won’t publish their proprietary algorithms it is probably printer and photo dependent. I always assume that PS especially CS has the best algorithms and I’d probably up sample a 380 dpi photo to 720, do some final sharpening if needed and print it.
Since it is recomended to print at a res evenly divisable by the printers res, is it best to down sample to 360ppi or to just go ahead and print it at 380?
Down-sample to 360ppi, or aforementioned "magic" resolution like 480 and do final sharpening at this resolution.
But… but… but… the ink costs like $80/microliter and the paper is $10/sheet!
First lesson in Photo school: The cost of paper and chemistry is irrelevant, the results you get are. Same goes for inkjet supplies.
If you want good results, you have to test. It is better (and in the long run, cheaper) to run a whole bunch of tests that tell you how the printer reacts to given inputs rather that trying to tweak every single image individually for optimum resolution and sharpness.
Learn how to optimize Photoshop for maximum speed, troubleshoot common issues, and keep your projects organized so that you can work faster than ever before!
Related Discussion Topics
Nice and short text about related topics in discussion sections
