The thinking is that the line screen for a litho book printed on coated stock would be around 150 lines per inch. By having the PPI at 300 you allowing two pixels per dot. In fact you can often get away with less down to around 200ppi some claim.
If you set it to 300ppi then the printer can effectively print two rows of pixels at a time? Does this mean resolution being 300 would make the printer print the document faster? Or have I missed the point completely here?
Resolution only exists in output, and is the the relative size of the pixels. The higher the resolution, the smaller the pixel size. The object is that you should not be able to make out individual pixels when viewing the output, so you need to keep the resolution high enough to keep that from happening, but there is no universal number that is correct -- it will depend to a degree on the image, and even more on the viewing distance. See Distance-Resolution Formula
The resolution at which an image is saved has no effect on the pixels themselves, AS LONG AS you don't re-sample when you change the resolution value. The only use made of that number is to determine the intended dimension and resolution when placed at 100% scale. In your example above, if you are talking about the same image pixels, one copy saved at 300 ppi and one at 10 ppi, but both scaled to the same physical scale there will be no difference at all between them. If however, you start with one image, then save a copy with a new resolution and re-sample, the results will be quite different.
In classes I used to demonstrate the relationship between resolution and image size using a balloon with a checkerboard drawn on the surface. As the balloon is inflated, the physical dimensions get bigger, the pixels (the squares of the checkerboard) get larger, and the density (number of pixels in a given area, or resolution) goes down.
I'm not sure Derek has given you a good analogy. Printers have anywhere from moderately low (300 dpi) to very high (2800 or moe DPI) resolutions, depending on the device. Litho image-setters are at the high end. Images ar printed using some sort of halftone screen because you can't print anything but the inks that are loaded into the press or printer. To get a 50% tint of gray using a solid black ink requires that you fool the eye and print small dark black dots with lots of white showing between them.
A halftone linescreen controls the density of the dots. Unlike pixels, which are discreet and you cannot fill only half a pixel, halftone dots are made from a clustered group of smaller printer dots, so some of those dots can be printed and others not, creating the illusion of many tones. The more printer dots used to create each halftone spot, the more shades can be produced. There is, however, a tradeoff here. The more of the printers available dots are used for a single spot, the fewer spots you can make per inch (a lower linescreen), so the more obvious those spots will become and the less detail you can render (think about black and white photos in an old newspaper), so you have to find the balance between spot size (linescreen) and the number of discreet gray levels that can be used in an image. 600 DPI desktop printers typically print at around 100 lpi and have acceptable tonality. The same image printed on a press on coated stock at 200 lpi will show both more detail and grater tonal range by virtue of having about two times the number of dots in each spot, and four times the number of spots in the image.
I'm not talking about a desk-top inkjet printers, they use a different system, I'm discussing litho printers – they use line screen (half-tone dots) and the amount of lines per inch used depends on the paper used. News print, for example uses a low screen of around 65 to 85 lpi and fine quality coated papers use a fine line screen 133 to 150lpi.
Below is a typical halftone – larger dots darker image, smaller dots lighter image.
That's very different from images of pixels, which are all the same size – the closer the pixels are together the higher the resolution!
Increasing resolution in Photoshop automatically tries to resample the pixel dimensions too. This makes sense to me. The more pixels, theoretically the more sharper the image will be when printed.
Increasing resolution in Photoshop generally has no positive effect. You can see from the info in the lower left corner of these two images that they have the same resolution and pixel dimensions but very different quality:
Yes, we run 175lpi and our image setter outputs at a total res of 2540dpi. That being said, Photoshop images need to be 300dpi and placed in InDesign no larger than 100%. Yes, Ricky, you can change the dpi in PS without resampling to whatever you want... but placing a 100dpi image in InDesign at 100% is exactly the same as placing a 300dpi image at 300%, which will result in a low-res fuzzy final product. An 8.5" x 11" image can be set to 300dpi or 10dpi, but it needs to be 2550pixels x 3300pixels if you want high-quality results.
See this is the thing, Robert-BPRint, you say if you print an image of 100dpi at 100% or 300pdi at 300%, it would produce a fuzzy image, but I've never seen it do this. I got a book made with an image of 10dpi and 300dpi. Both were full-page images on a 30" squared book, but the images side by side looked exactly the same. The image used was about 2400px X 3500px or so. I can understand if I resampled the pixel dimensions that one would look very different to the other, but just changing the resolution alone visually made no difference.
Again, I understand that the pixels are drawn smaller the her the resolution, but scaling the image up even well beyond its recommended size, the only degrading aspect is the picture itself, with the pixels that make up the image becoming visible as if you digitally zoomed in on it.
Today I naughtily got a new book, and a good 70% of the images used in it showed warnings when going through preflight that their resolution was below what it wanted (300ppi for the equivalent size of each image). I ignored them all and they all look just as good as the images that had no warnings again.
If your image setter is 175LPI then you only need 252ppi minimum for your images.
You're dealing with a halftone square, or a square to make it simpler.
The square is rotated 45 degrees, leaving tip to tip 1.44 x the original width(height). As any square corner to corner is exactly 1.44 x it's w or h.
Truly you'd only need 300 ppi if the image setter was set to 200 LPI
Although, you wouldn't run newspaper at 175lpi for example, so news papers can typically be 200 ppi or lower.
You talk about 10 ppi and 300 ppi, but I suspect those are the "Actual PPI'" numbers listed for the images. What are the "Effective PPI" numbers?
It also matters what sort of equipment is is being used for output. Many print-on-demand projects, for example, run on digital, rather than litho, presses and run at lower screen values (and thus require less image resolution). Many newer presses also use some form of stochastic screening, rather than conventional halftones. In a stochastic screen the spots are all the same size and 100%color value, and very small, but are spaced tighter together in dark areas and farther apart or not at all in light areas. This also requires less image resolution for good results.
There's no relationship between line screen and image resolution. 300ppi has become a magic number but it would be easy to come up with an image that would benefit from more than 300ppi of resolution and equally easy to find one that would require much less.
You could have a high contrast image with no halftone screen (all the values at or near 0 or 100%) and a lot of detail—colored line art:
Or on the other end a mid-range image, that has no clearly defined edges and details, which require a higher resolution to resolve:
300 is really only a rule of thumb that covers image types between the two extremes. If you had an image straight from a camera or scanner with an effective res of 203ppi in the layout there would be zero advantage in upsampling to a magic number other than satisfying your printer's preflight.
Ricky, when you change the dpi value without resampling, you are not actually changing anything. All you're doing is changing your reference point. In your case, you are correct, and the image you used had a high enough resolution (pixel count). What I'm saying is if you 1) place an 8.5" x 11" @ 100dpi (850px x 1100px) image into ID, it is exactly equal to 2) placing a 2.833" x 3.667" @ 300dpi (850px x 1100px) image into ID then enlarging it by 300% to fill the page... both would produce a fuzzy image. So, the rule off 300dpi is strictly as a reference point within Photoshop. It is perfectly ok to send the printer an image that is 25.5" x 33" @ 100dpi (2550px x 3300px), but if you're designing a print job targeted for Letter size paper, it would be a lot easier going with 8.5" x 11" @ 300dpi (2550px x 3300px) in PS.
Eugene is correct that technically the full 300dpi is not required, so if your preflight flags an image that is only 200dpi, it should be ok. Also keep in mind that the rules aren't quite as strict for digital printing, I have seen much lower resolution images (120dpi) print clean on a good digital press, because digital does not use lpi like an imagesetter does.
Exactly correct Rob, I'd run the foggy image at 225 ppi if it was supplied that way for litho, I'd go to 120ppi for digital for that image, although I've gone lower.
For the sharp lines I'd go roughly 300 ppi, but for digital I wouldn't go below 220 and for litho I'd insist on at least 260 ppi for this - but I'd question the printers.
Most of the time the conversation is lost on these people (not being mean) but they're trained on a preflight system and that system as 300 ppi embedded on the checks and nothing else will do, not the content of the image or the actual ppi is considered, it's not 300 it can't be printed, which is wrong.
See these kinds of images confuse me:
I've never seen a printer produce a pixelated image because its 72ppi instead of 300ppi for example. That is just an image that is scaled down in pixel dimensions surely?
So the page in that book I made is here. Looking at the image properties in InDesign, I can see that actually they are the same effective ppi as eachother, despite one being save at 1ppi, and the other at 300ppi.
Maybe I am thinking of ppi in the wrong way. When printers ask for 300ppi images, do they basically just mean the image should have 300 pixels per inch when printed at the size you put them on the page, so not to see pixelation where it is 'zoomed in' too far? Therefore, changing the ppi by itself and not resampling pixel dimensions, or not having a large enough graphic to begin with (in pixel dimensions), will make no difference at all? It makes no difference to the print quality directly or to the printers. It is just an algorithm to tell you that your graphic might look pixelated when printed at that size because there aren't enough pixels in the image to cover that area without you seeing the jagged edges of the pixels?
In other words, no matter how high or low your ppi is set to on your image, it will not change the look of your image when printed. It just tells you what physical size you can print that image at while still maintaining 300ppi.
Is this correct? Or have I still not got it?
Thanks for the help by the way. It just frustrates me that I can't understand how resolution affects images when printed!
You are correct, although placing an image at 1ppi on the page and scaling it down to .2% is not a common practice. And many print shops are going to freak if they look closely at the file. You're asking the printing software (RIP) to do a lot of work during the printing that it shouldn't have to do. Imagine having to wash an oversized sweater and wait for it to shrink in the dryer every time you want to wear it.
However, it isn't too much problem to take a file at 250 ppi and shrink it down a bit. Or a file at 350 ppi and scale it up. In the old days printing RIPs couldn't do that extra work. Now, it's not a big deal.
Next, you don't zoom in on printed images. You can take a magnifying glass to printed pages, but you're not going to see more details. You're going to see more lines per inch of the halftones.
BUT, if you take a picture at 72 ppi and then print it at that size there are several things that could happen. In the Samurai image you show, there would be different things happening in different areas. The green background could look pretty much OK. That's such a "mooshy" part of the illustration that a loss of details may not be obvious. The face could look a little "soft" or blurry. That's because many RIPs try to fill in the lack of pixels in the output. They don't necessarily do a great job. So things can get fuzzy.
You most likely wouldn't get jagged edges. Continuous tone images don't usually have the stark edges of a single color to cause that look. But if you had black type on a white background at 72 ppi, it will probably look jagged. But remember, jagged only happens on curves and diagonals. Pure horizontal and vertical lines won't look jagged at all.
You may want to look at my book "From Design Into Print, Preparing Graphics and Text for Professional Printing." It was written especially for the designer who has never been given any production education.
In your example the 1ppi Actual resolution would only be the output resolution when you print the image at its actual size—6238cm x 8750cm (or 204x287 feet!). But you aren't printing actual size, you've reduced the size of the art and its pixels so now the size of a pixel is 1/391" and not 1"—391 pixels per inch is the effective resolution.
In other words, no matter how high or low your ppi is set to on your image, it will not change the look of your image when printed
No. You always have to consider scale or the effective resolution. Lowering the effective resolution of a high quality image from 300 to 72ppi would have a clear affect on output quality.
Actual Resolution = Actual Print Size (the image and pixels at 100%)
Effective Resolution = Scaled Print Size (the image and pixels scaled)
And many print shops are going to freak if they look closely at the file.
If you are sending a PDF I don't think there's any way to know what the original or actual PPI was (unless it's buried in some preflight rule). Acrobat's Output Preview's Object Inspector only lists the image's output (effective) resolution and its pixel dimensions and doesn't care about the original's unscaled res.
So this image started as 1ppi in ID.
Things are making a lot more sense to me now thinking of it this way. So Sandeecohen, sending an image of 1ppi to a printer to scale it, process and print it now at around 300ppi, makes the printer do unnecessary work? This would be a logical reason to me to make sure things are the equivalent of 300ppi, making the print process as efficient as possible.
process and print it now at around 300ppi, makes the printer do unnecessary work?
I don't see how the degree of scaling could matter—whatever the percentage change the math is the same. It's easy enough to test—the amount of scaling isn't changing the print time from my (old) printers. Also, if you are providing PDF, the scaling happens during your export, so the difference between actual and effective resolution no longer exists in the PDF.
You need to get off the 300ppi bus... it's not accurate
all have different degrees of print quality.
In most cases 300 ppi is too much.
You're totally right about sending a PDF eliminates all scaling processing time.
All I know is print shops used to ask that images be set for their resolutions in Photoshop and placed at 100% in the page layout program. It meant less processing time.
However, today's RIPs are fast. And might not care.
InDesign exports to PDF and can downsample images for teh PDF.
I think the Resize and Place at 100% in the layout application is a leftover from Quark days.
I know, I know, I'm just using it as an example at this point.
Thank you all for the help!
Quark also resized scaled images, but I don't remember a way to find the effective res other than getting out my calculator. I can imagine some printers would have had a 100% rule because without an effective res reference one wouldn't know the output resolution.
Ricky-T, there is a lot of good information here but it talks about the small details of quality. You talk about 10 ppi and an effective resolution of 10 ppi will print as a colorful blur. The trick is to understand that only ONE resolution is important and that is the EFFECTIVE resolution. Someone mentioned this but I think it was lost.
If you have 300 ppi at 100%, the effective resolution is 300 dpi. If you have 300 ppi at 200% the effective resolution is 150 ppi. And if you have 3 ppi at 0.01% you have an effective resolution of 300 ppi. The % here is not just as typed in one place, but the actual, real, percentage, from graphic to output. Clearly if you use 200% scaling when you print, you halved the effective resolution.
In this regard the resolution you choose in Photoshop doesn't matter a bean (but getting it right may help in your keeping track of what is going on, and in default layout size). What matters is the number of pixels and the number of actual, eventual, inches. Divide one by the other and you get pixels per inch. That's all.
I just have memories of using Quark. And setting up batch photoshop actions to resize all images to 300 ppi for placing in Quark.
This was before a PDF workflow that's in place now. I was working with EPSF (or something like that - a verison of EPS).
It was between that and PS. But nobody was making PDFs for printing then.
Sorry, 3 ppi at 1% not 0.01%, is 300 ppi effective.
Yes, in MOST cases (web, digital), 300dpi is too much. But, I cannot tell you how many times inexperienced designers have started with low-resolution images for a website or digital copy run, THEN tried to go to offset press with their art! Too much is better than not enough, and 300dpi images can always be down-sized to suit your target media. Especially if your end target is a PDF file, you can work in high-res all day long then export using a target profile. But, once you have created your digital masterpiece in low resolution, there is no up-sizing it!
One benefit of having images at 300ppi actual resolution, is that once the layout has been approved, you resize (reduce) the image in Photoshop by copying and pasting the reduction percentage from ID into the Image Size dialog in PS, then sharpen at final size.
I tend to use PSDs and keep the unsharpened, original image in a smart object so it can be re-enlarged if necessary, and the Smart Filter used to sharpen will already be about right. You can even re-purpose for one-colour print and get the colour back in future. It's convoluted and greedy for storage space, but I automate it and storage isn't expensive.