Could this be a new version of KFX?

[5aLIVE]

Hello to everyone, I recently purchased a Kindle Print Replica ebook on Amazon.

It's been about 5 years since I last used K4PC so I had to update from a dusty old v1.16.0 (44025) on an old laptop to 2.8.0 (70980) on a new laptop to be able to view the newer file format. I've also disabled autoupdates in the new version.

I also had to update Calibre (portable) from v1.48 to v9.5.
I also installed the DeDRM v10.0.18 and KFX Input v2.29.0 plugins.

I've been attempting to using KFXKeyExtractor282 to extract the key for this new ebook but I see the following error:

BookOpen error 25 DRM envelope refers to an unknown voucher

Has anyone seen this before?
Could this be a new version of KFX in the wild? This particular ebook was published in Jan 2026 if this is relevant. Is there anything I can do to remedy this issue?

In the meantime, I moved my attention to an older title from my modest collection.

I was able to extract the secrets from one of the old Print Replica books I purchased back in 2020 and was able to import the KFX file and extract the PDF successfully using Calibre. So that at least proves I have a working setup.

side note:
One observation I thought I'd mention here is that the original PDF was produced with MSWord 2013 and the new download of this contains a PDF produced by pypdf. The file size has gone from ~5MB to ~25MB which seems ridiculous to me. Is this normal for a PDF Print Replica KFX files or an outlier?

Thanks in advance,
5aLIVe

[jhowell]

Originally Posted by 5aLIVE

I've been attempting to using KFXKeyExtractor282 to extract the key for this new ebook but I see the following error:

BookOpen error 25 DRM envelope refers to an unknown voucher

I suggest raising this in the Discussions section of the site hosting the DeDRM plugin. Detailed assistance with DRM removal is not allowed here on MobileRead.

Originally Posted by 5aLIVE

One observation I thought I'd mention here is that the original PDF was produced with MSWord 2013 and the new download of this contains a PDF produced by pypdf. The file size has gone from ~5MB to ~25MB which seems ridiculous to me. Is this normal for a PDF Print Replica KFX files or an outlier?

The KFX Input plugin uses pypdf to produce a PDF file for fixed-layout books in cases where the KFX format either contains a PDF that was split into multiple pieces or contains images instead of PDF. Both cases are fairly common.

Without knowing the specific book is is difficult to say more about it.

[5aLIVE]

Originally Posted by jhowell

I suggest raising this in the Discussions section of the site hosting the DeDRM plugin. Detailed assistance with DRM removal is not allowed here on MobileRead.

I'm sorry about that, I didn't realise I was breaking forum rules here.

Originally Posted by jhowell

The KFX Input plugin uses pypdf to produce a PDF file for fixed-layout books in cases where the KFX format either contains a PDF that was split into multiple pieces or contains images instead of PDF. Both cases are fairly common.

Without knowing the specific book is is difficult to say more about it

I understand, I mistakenly thought that it was Amazon were re-rendering the original book that had been uploaded by the original author.
This particular book I recently downloaded for testing has an .azw file extension
and includes 9 files with .azw.res files.

Am I correct in saying that these files in essence are separate PDF files each containing pages from the book and need to be joined together to create a complete document?

Can you tell me if it is possible to restructure the PDF using some tool to restore the file size close to the original document? I'd be grateful if you could advise me on this.

What do you need to know regrading this specific book? I'll try to help as best I can.

Thankyou,
5aLIVE.

[jhowell]

Originally Posted by 5aLIVE

I understand, I mistakenly thought that it was Amazon were re-rendering the original book that had been uploaded by the original author.

Amazon does a varying amount of processing and reformatting of books depending on what was originally provided by the publisher and the specific format being delivered to the user (MOBI, KF8, KFX, etc.).

Originally Posted by 5aLIVE

This particular book I recently downloaded for testing has an .azw file extension and includes 9 files with .azw.res files.

Am I correct in saying that these files in essence are separate PDF files each containing pages from the book and need to be joined together to create a complete document?

The .azw.res files may contain either PDFs or just page images. It varies by book.

If you add up the total size of .azw.res files along with the .azw file it should give you a rough idea of the size of the content Amazon is providing for that book.

Originally Posted by 5aLIVE

Can you tell me if it is possible to restructure the PDF using some tool to restore the file size close to the original document? I'd be grateful if you could advise me on this.

Without knowing more about this specific book it is hard to say whether or not anything can be done. If those .azw.res files contain highly processed PDF files or only images then the original content will not be recoverable from KFX format.

Originally Posted by 5aLIVE

What do you need to know regrading this specific book? I'll try to help as best I can.

If you provide the ASIN of the book (ten character ID starting with "B") I may be able to tell more by looking at the product listing and the the free sample for it, as long as it is available in my region.

[5aLIVE]

Thankyou very much for the detailed additional information, so my initial thoughts regarding Amazon reformatting books may well prove to be true afterall.

I can confirm that the sum of .azw and azw.res files is a fair estimate of the size of the extracted PDF.

The ASIN for this book is B076ZX6CGB which I see matches a folder name in My Kindle Contents folder. This book is available in the US as the author is a fellow American. Hopefully the sample file will give you the information you need.

[jhowell]

Originally Posted by 5aLIVE

The ASIN for this book is B076ZX6CGB which I see matches a folder name in My Kindle Contents folder. This book is available in the US as the author is a fellow American. Hopefully the sample file will give you the information you need.

I was able to examine the book in question and what I see is similar to what you stated except that I get a PDF from KFX of 15MB instead of 25MB.

That book in KFX format has been split by Amazon into 26 individual PDF files, each holding up to 20 pages. The main reason that the split PDFs total to a size larger than the original is that they each contain a copy of the fonts that they use, which are many in this book. The recombining of the individual files back into a single PDF done by KFX Input using pypdf does not eliminate duplicate fonts and so the resulting PDF contains multiple duplicate fonts, making up most of its size.

I will research options to better handle this situation for a future KFX Input plugin release. Thanks for bringing it to my attention.

[kovidgoyal]

@jhowell: There is existing code in calibre for combining PDF files while removing duplicate fonts, used by its own PDF output. Dont know how suitable that is for you, but worth a look. IIRC the functions are called merge_fonts() and dedup_type3_fonts()

[5aLIVE]

Originally Posted by jhowell

I was able to examine the book in question and what I see is similar to what you stated except that I get a PDF from KFX of 15MB instead of 25MB.

Thank you for looking at this for me, that is intesting that your output file is 10MB smaller than mine. I assume you were working the sample file rather than the complete book in this instance yes?

Originally Posted by jhowell

That book in KFX format has been split by Amazon into 26 individual PDF files, each holding up to 20 pages. The main reason that the split PDFs total to a size larger than the original is that they each contain a copy of the fonts that they use, which are many in this book. The recombining of the individual files back into a single PDF done by KFX Input using pypdf does not eliminate duplicate fonts and so the resulting PDF contains multiple duplicate fonts, making up most of its size.

That is curious that you are seeing 26 individual files here compared to my total count of 10. Could different versions of the same book exist for different markets? I'm in the UK.

Originally Posted by jhowell

The main reason that the split PDFs total to a size larger than the original is that they each contain a copy of the fonts that they use, which are many in this book. The recombining of the individual files back into a single PDF done by KFX Input using pypdf does not eliminate duplicate fonts and so the resulting PDF contains multiple duplicate fonts, making up most of its size.

I follow you, the duplicate embedded fonts explains the redundancy in the output PDF.

Originally Posted by jhowell

I will research options to better handle this situation for a future KFX Input plugin release. Thanks for bringing it to my attention.

That sounds promising, I'll look forward to seeing that some day.
You are most welcome and my thanks to you for continuing to develop such a useful plugin.

Kind regards,
5aLIVE.

[jhowell]

Originally Posted by kovidgoyal

@jhowell: There is existing code in calibre for combining PDF files while removing duplicate fonts, used by its own PDF output. Dont know how suitable that is for you, but worth a look. IIRC the functions are called merge_fonts() and dedup_type3_fonts()

Thanks. I will look into that.