Confused about PPI, 300 and 72

[AFY7]

IngramSpark requires print-ready PDFs be exported at 300 PPI. But IngramSpark's guide also says images of 72 PPI are acceptable! To me, this is confusing. If the export is at 300 PPI, shouldn't all images be 300 PPI too?

[v_kyr]

Well it would have helped if you would have supplied a reference/link here, so people don't have to search after that topic ...

• https://www.ingramspark.com/hubfs/downloads/file-creation-guide.pdf

They talk about inner text pages there and say "at least 72 ppi", also for book covers then "at least 200 ppi". The whole is based on their internal processing workflow and tells the minimum requirements here. - So if you export everything as 300 ppi (images too) you should be fine!

[AFY7]

v_kyr,

So, are you saying that if the placed PPI of an image in the project file is 72, and I export at 300 PPI, then the image in the exported PDF will be at 300 PPI?

[v_kyr]

No, I said you should setup everything (all and thus the embedded images too here) as 300 ppi (so images too) and export as 300 ppi so you are on the secure side!

[AFY7]

v_kyr,

That is what I have always done, but I received a new job that had been started by someone else. Oddly, all the images in her file are at 100 PPI. I will meet this person tomorrow and discuss this issue.

Thank you for your assistance!

[thomaso]

2 hours ago, AFY7 said:

if the placed PPI of an image in the project file is 72, and I export at 300 PPI, then the image in the exported PDF will be at 300 PPI?

Sort of it will. Affinity resamples the jpg-compressed images on PDF export, so your 72 dpi will get upscaled to the document / export DPI/PPI.
Which is not really e.g. 300 but smoother, less jaggy than pure 72

[AFY7]

thomaso,

When I tried this I got a different result. I placed an image at 100 dpi into a blank page and exported as PDF/X-1a:2003, 300dpi. Opening the resulting PDF in both Affinity and a third-party app, the image dpi is shown as 100. Is there a setting that I missed that resamples ("up samples") the image at export?

[lacerto]

On 1/19/2022 at 6:02 AM, thomaso said:

Sort of it will. Affinity resamples the jpg-compressed images on PDF export, so your 72 dpi will get upscaled to the document / export DPI/PPI.

I was afraid that something has changed, but happily this is not true, or is only conditionally true. It is bad enough that upscaling happens for vector cropped placed images, but it does not happen for non-cropped or vector-clipped images. Upscaling should never happen inadvertently.

[EDIT: I say "happily" because it must be possible to use placed images with low PPI value (e.g. screenshots) without upscaling them, because mechanical (nearest neighbour kind of] pixel addition done by the RIP produces typically better results than bilinear upscaling (especially when it happens inadvertently.]

 

[AFY7]

Lagarto,

Thanks.

I looked at your retest.afpub. I don't understand why the bottom image was upscaled but the top image was not.

[lacerto]

5 hours ago, AFY7 said:

I looked at your retest.afpub. I don't understand why the bottom image was upscaled but the top image was not.

The bottom image is cropped with the Vector Crop tool. Affinity upscales such images using the document DPI value (in this case 300 dpi). If cropping is done by clipping the image into a vector shape (or if the image has not been cropped), this does not happen. Affinity apps also upscale raster layers that have (Pixel) layer id appended (in the Layers panel), whenever you manually resize them.  

[AFY7]

Lagarto,

Great information. Thank you.

[AFY7]

Lagarto,

I follow you with regard to the vector crop tool. However, I don't understand how to append a layer ID.

[AFY7]

Lagarto,

Never mind! I just found the Add option for adding a layer.

Thanks again for your assistance.

[lacerto]

4 hours ago, AFY7 said:

However, I don't understand how to append a layer ID.

Sorry, I used an inaccurate term. I mean the words appended in parentheses in the Layers panel by the app after a layer name that show the kind of a layer in question. E.g., for image layers, "(Image)" is appended after the layer name, and for pixel layers, "(Pixel)" is appended. These identifiers are not user-changeable (unlike the layer names themselves).