(Amazon KDP) Vectors in PDF/X-1a:2001 printing more black than PNGs, but display identical digitally

[Kyle23930129ln4r]

Hello!

Perhaps someone can help me. I am preparing a black and white paperback manuscript for printing through Amazon KDP's print on demand service.

From InDesign I export a PDF/X-1a:2001 (per Amazon KDP rec) which contains solid black PNGs. These PNGs are, for example, a logo with a transparent background, or solid black silhouettes with transparent backgrounds. (Transparency I figure isn't relevant here considering it gets rasterized away by the PDF/X-1a standard.)

These solid black images print with what appears to be some sort of grey; the ink itself is not grey, but is perforated with the tiniest white spaces, and appears to be 'sprayed' onto the page in a different manner than how everything else in the manuscript gets printed, as it's got these rather rough lines which don't seem to be the product of pixelation but rather due to the manner in which the ink gets onto the page, which I assume has been determined by the interpretation that the printer has made of the rasterized pixel data in the PDF.

Everything else in the manuscript is also solid black, but prints smooth. Everything that prints smooth is incidentally vectorized, being either actual text or outlined text.

So: the rasterized, solid black content (logos and silhouettes) is printing different than the vectorized, solid black content (text and text outlines).

I've attached an image. On the left is how it appears in the PDF, and on the right is how it appears printed. The rounded box of text in the lower right hand corner is actually an image, a silhouetted PNG. Notice how it prints differently than the vectorized content of the text above it, despite appearing indentical in its blackness in the PDF. (If you zoom in you can see the 'perforation' that I'm talking about--it's not a solid black, unlike the vector content above it.)

The export settings which produced this print are:

• PDF/X-1a:2001
• Acrobat 4 (PDF 1.3)
• All three Compression drop-downs set to Do Not Downsample
• Compression for Color, Greyscale and Monochrome set to Automatic (JPEG) (maximum), Automatic (JPEG) (maximum) and None, respectively
• Compress Text and Line Art is unchecked.

I think that I can achieve the result of the rasterized black images printing smooth if I can vectorize them. I don't know how to get the images into the PDF/X-1a standard and retain their vectorization. SVGs that I drag into indesign turn into rasterized images when exported, this way.

Could I, for example, 'outline' the images in the same way that text is outlined to produced vectorized pixels resembling text? That I think would be an effective workaround, if possible.

Or if there's another way--I'm ALL ears!

Thank you for any help!!!

[Test Screen Name]

Well, they aren't going to be solid black though. PNG does not support either greyscale or CMYK. When converted to CMYK they will be converted to a muddy mixture of inks. Since your PDF is to be converted to CMYK, maybe working in CMYK will give you more control. 

[rob day · Accepted solution]

If you zoom in you can see the 'perforation' that I'm talking about--it's not a solid black, unlike the vector content above it.)

What you are describing and showing in your capture are halftone dots. The placed PNG has to be RGB, and an RGB black would always convert to a 4-color black with a PDF/X-1a export. Looks like KDP prints the page as black only, so the halftone is showing on the edge because of the conversion from 4-color CMYK to 1-color black (grayscale) at output.

A possible solution would be to open the PNGs in Photoshop, make sure the black is absolute black (0|0|0 RGB), convert to grayscale, and save as PSD. When you place the PSDs, use InDesign’s Separation Preview to confirm that the rasterized text is reading as 0|0|0|100 CMYK

[Kyle23930129ln4r]

You're a genius--it worked.

1. opened the PNG in PS
2. control-clicked the layer thumbnail to select the PNG non-transparent content
3. applied solid color, RGB 0|0|0 which creates a new layer
4. deleted original layer
5. image > mode > grayscale > discard
6. dragged into indesign
7. separations preview--quick test to see if it's exclusively black: hide the black layer by clicking the eyeball and it completely goes away

Well!

Thank you!!

[Kyle23930129ln4r]

*6. dragged into indesign after saving the image as PSD

[rob day]

I don’t think you need so many Photoshop steps. Just open the PNG, flatten it, convert to grayscale, and adjust the Levels black Input slider until the percentage hits 100% in the Info panel—save as PSD:

[Kyle23930129ln4r]

Okay, cool.

But after flattening it--and removing the transparency--won't there be white ink in the image when printed, considering the background is now white?

[rob day]

There’s no white ink—it’s just black ink on paper

[Kyle23930129ln4r]

Wild, thanks

[Kyle23930129ln4r]

Okay--so the process I'm running is basically what you said, without flattening it

1. open PNG in PS
2. convert to grayscale
3. levels > drag left and middle sliders all the way to the right, which makes every pixel in the image K : x / 100%
4. save as PSD, etc.

[Luke Jennings3]

Do you have black only vector versions of the art? (svg) Open the svg in Illustrator and save as pdf (using the Illustrator default settings) and place the pdf into InDesign. If you absolutely must use raster png files, open them in photoshop and save as grayscale, adjust the curve so the black is 100% and as little "soft edge" as possible (which would explain the less than 100% ink coverage), or re-create the type in InDesign in black.

Check your exported pdf in Acrobat using the Output Preview tool, the placed pdf files should be 100% black only. 

[James Gifford—NitroPress]

Looks as if you've gotten good answers, but I will add a point: PNG is best for web/electronic display and should be avoided for print no matter how many layers (such as PDF) it passes through.

Use JPEG optimized for print for more consistent, reliable images. Best is to convert to grayscale before placing; PDF export using a grayscale profile works almost as well.

Or, best of all for illustrations involving text, try to keep the text live/vector instead of part of a rasterized image. Lay the text in as ID text boxes, or create and export PDFs from ID, Illustrator, etc.

—

[rob day]

Hi James, What would the quality problems be with PNG? In this case a different format is needed because PNGs are limited to RGB, but they can be saved as 24- or 48-bit depth. Starting with CC2015 they can be fully color managed, and wouldn’t have JPEG like compression artifacts. If I place the same image saved as PNG and PSD and export to PDF, Acrobat’s Object Inspector shows identical image attribute specs, and Separations show identical output numbers.

[James Gifford—NitroPress]

A lot of that comes down to "can be." I know we've seen a lot of issues when PNG is used for print projects. The format was designed and optimized for digital display.

My impression is that it's not the standard or the technical "can be" but that they tend to be generated and edited by sloppy tools, and the demands of print, especially higher-end processes, expose the flaws.

We'll do this all over again with WEBP in the next few years. 🙂

—

[rob day]

 I know we've seen a lot of issues when PNG is used for print projects

Can you provide an example of a PNG that caused a problem in print?

It wouldn’t be too difficult to mishandle PNG color and get an unexpected result, (e.g. let it export as DeviceRGB—RGB with no profile—or expect an indexed PNG to output like a 24-bit PNG), but that would be a problem with any file format.

You’re not the only one to warn about using PNG in print, but I’ve never been able reproduce or force a problem with PNGs on export. It feels more like a cousin of 90’s conventional wisdom where you would never place an RGB image in a print document—now it’s the reverse because placed RGB image formats can have embedded source profiles, and color managed CMYK conversions can happen later in the printflow.

[James Gifford—NitroPress]

I won't make any passionate arguments here; pretty much any image format can be used successfully if some guidelines are followed and you have enough skill to select, format, place and export them.

But PNG was developed expressly for online use, without consideration of anything but home-office printing. That it can be used for print, and (with above caveats in mind) just as successfully, doesn't mean it's a good choice for print work.

Because it's a "web" format, as well, most PNGs I encounter are on the crappy side—created by substandard tools and processes, small in resolution, erratic in color management, and rarely taking advantage of its one strong feature, lossless compression. When I have to use a PNG I get from somewhere, I invariably have to work on it in Photoshop, and always save to an optimal JPEG for further use.

I also have encountered very few examples of that lossless/superior compression actually producing a better result than a well-optimized JPEG.

Anyone who wants to use PNGs and work with the format's strengths and around its limitations (many of which come from the image creator/providers) can certainly do so. Maybe I'm too conservative, or have battled with those crappy images too many times, but it feels like the wrong tool for the job.

As noted, I'm waiting for this argument to cycle around to why WEBP is just a fine format for color print work. 😛 But then, I remember the vehement arguments about why GIF was superior to that JPEG crap, too, with the gods of TIFF sniffing in the background.

—

[rob day]

As noted, I'm waiting for this argument to cycle around to why WEBP is just a fine format for color print work.

I haven’t paid much attention to it, but according to its Wiki it can be saved with an ICC profile, contain transparency, and (I assume) be at least 24-bit. So seems like the only thing to worry about, will be compression artifacts. In the end either you are going to properly color manage or not, the format doesn’t save you from that.

For me the test is what happens to the image when it is exported to PDF. There is no record of the original format over in the PDF, but Object Inspector shows all the important info—color space, color profile, bit depth, effective resolution, and how it’s been compressed on export