[Conversion Output] KFX Output

[oops]

Originally Posted by alessandrokindle

Hi, I have an EPUB file that contains different types of fonts. After using your tool, the fonts are not preserved...

Hi. I ran into the same problem and have root-caused it and have a fix. First of all, the KFX Output plugin is not at fault. Kindle Previewer/KindleGen are the ones that are misbehaving. In my case, there would be messages in "KPR.Log" where the font files would be invalidated. The reason is that "FontValidator.exe" is a) a 32-bit application and b) requires the x86 2013 VC redistributable DLLs. It might complain only about "msvcr120.dll" and you'll be tempted to just copy it alongside the executable, but looking at the executable in Dependency Walker, it needs its supporting libraries as well. In my case the redistributable was not installed--I don't play games on this machine and Steam tends to be the one installing these nowadays. Installing the redistributable fixed the problem.

For reference here is what the error looks like in "KPR.Log":

Spoiler:

[ERROR] Invalidating the font:%LOCALAPPDATA%\Temp\2f029474-7d9d-47a1-b485-786b93621966\conv_out\book\res/rsrc55B

In addition, some fonts might require conversion to TrueType glyphs to display properly, so make sure that you take care of that in case it is still not working for you. This is what that warning from KindleGen looks like; note that it doesn't bubble up to Kindle Previewer's conversion log:

Spoiler:

************************************************** ***********
Amazon kindlegen(Windows) V2.9 build 0000-kdevbld
A command line e-book compiler
Copyright Amazon.com and its Affiliates 2015
************************************************** ***********
...
Warning(prcfile):W14029: CFF/Type1 (Postscript) embedded font included in your source may not render clearly on all Kindle readers. Please refer to Kindle Publishing Guidelines on the recommended usage of embedded fonts. Font File : ...

Originally Posted by jhowell

If that doesn't work I would need to see an example of an EPUB the loses its fonts when converted in order to diagnose this further.

Hopefully @jhowell, this saves you some sleuthing.

[jhowell]

Originally Posted by oops

Hi. I ran into the same problem and have root-caused it and have a fix.

Thanks for the useful information.

[alessandrokindle]

Originally Posted by oops

Hi. I ran into the same problem and have root-caused it and have a fix. First of all, the KFX Output plugin is not at fault. Kindle Previewer/KindleGen are the ones that are misbehaving. In my case, there would be messages in "KPR.Log" where the font files would be invalidated. The reason is that "FontValidator.exe" is a) a 32-bit application and b) requires the x86 2013 VC redistributable DLLs. It might complain only about "msvcr120.dll" and you'll be tempted to just copy it alongside the executable, but looking at the executable in Dependency Walker, it needs its supporting libraries as well. In my case the redistributable was not installed--I don't play games on this machine and Steam tends to be the one installing these nowadays. Installing the redistributable fixed the problem.

For reference here is what the error looks like in "KPR.Log":

Spoiler:

[ERROR] Invalidating the font:%LOCALAPPDATA%\Temp\2f029474-7d9d-47a1-b485-786b93621966\conv_out\book\res/rsrc55B

In addition, some fonts might require conversion to TrueType glyphs to display properly, so make sure that you take care of that in case it is still not working for you. This is what that warning from KindleGen looks like; note that it doesn't bubble up to Kindle Previewer's conversion log:

Spoiler:

************************************************** ***********
Amazon kindlegen(Windows) V2.9 build 0000-kdevbld
A command line e-book compiler
Copyright Amazon.com and its Affiliates 2015
************************************************** ***********
...
Warning(prcfile):W14029: CFF/Type1 (Postscript) embedded font included in your source may not render clearly on all Kindle readers. Please refer to Kindle Publishing Guidelines on the recommended usage of embedded fonts. Font File : ...

Hopefully @jhowell, this saves you some sleuthing.

This fixed the problem, thank you so much!

[tomsem]

Originally Posted by alessandrokindle

This fixed the problem, thank you so much!

I take it that this is a Windows-specific issue?

[tomsem]

I've been using CLI to convert KPF (from Kindle Create) to KFX, but just tried using Convert in calibre after importing the KPF. The job ran for a couple of minutes without progressing beyond 67%, so I killed it (the CLI does this in a couple of seconds).

I assume this workflow is not supported. Is it possible to support it?

[jhowell]

Originally Posted by tomsem

I've been using CLI to convert KPF (from Kindle Create) to KFX, but just tried using Convert in calibre after importing the KPF. The job ran for a couple of minutes without progressing beyond 67%, so I killed it (the CLI does this in a couple of seconds).

I assume this workflow is not supported. Is it possible to support it?

calibre's conversion workflow always converts to EPUB as an intermediate format, so in your case it will convert KPF to EPUB and then EPUB to KFX using the Kindle Previewer. Additionally calibre's conversion does not properly support fixed layout books such as comics. So for now it is best to use the CLI to convert KPF to KFX.

At some point I may add a toolbar button to this plugin to perform the equivalent conversion of the CLI within the calibre GUI, bypassing calibre's conversion pipeline. But that won't be coming in the near future.

[tomsem]

Originally Posted by jhowell

calibre's conversion workflow always converts to EPUB as an intermediate format, so in your case it will convert KPF to EPUB and then EPUB to KFX using the Kindle Previewer. Additionally calibre's conversion does not properly support fixed layout books such as comics. So for now it is best to use the CLI to convert KPF to KFX.

At some point I may add a toolbar button to this plugin to perform the equivalent conversion of the CLI within the calibre GUI, bypassing calibre's conversion pipeline. But that won't be coming in the near future.

Got it, thanks!

[tomsem]

I have an interesting issue.

I've started doing Add from folders and subfolders for the first time, as I'm doing some housekeeping on my calibre library and re-doing the earliest ones added when I didn't know as much - and also I prefer KFX to AZW3 for side loading and when these are fixed content, the conversion to ePub with From KFX is more of a no brainer (in past I had to remember to use KindleUnpack's conversion).

I'm downloading with Windows app, pointing it at the My Kindle Content folder and selecting the aforementioned option. It asks if I want to combine books in folders as one book, or treat them as separate.

If I chose to 'combine' it creates books with goofy titles (ASIN?) that end with .azw extension and inside that are the imported book in KFX, usually. And that goofy title is used for the book file (.kfx) and another with the same name with .MD extension.

These .MD files are in binary format, no idea where they came from but it was not in the source folder, and .MD is supposed to be plain text.

I thought I could exclude .MD from calibre's supported file preference but it had no effect, and even if there was no .MD file the title/filename is not as expected.

This instance of calibre is only used to de-drm books, then I import them to my Mac via cloud folder. But Saving to this folder propagates the goofy book name, making it hard to merge with the previous book record in order to update with the refreshed content.

The workaround is to treat them as separate books. Then the MD's have the goofy names and the KFX (AZW3 or MOBI) have the expected names. I can delete the MD entries, select the books and Save to the transfer folder.

I don't know where these MD files are coming from (it's not in the source folder) but I believe it is only there with KFX imports so that's why I'm posting here.

[jhowell]

Originally Posted by tomsem

If I chose to 'combine' it creates books with goofy titles (ASIN?) that end with .azw extension and inside that are the imported book in KFX, usually. And that goofy title is used for the book file (.kfx) and another with the same name with .MD extension.

The combine option tells calibre to treat each of the files in the folder as a separate instance of the same book. Books in KFX format are composed of multiple files that together make up only one instance of the book.

The Kindle for PC app uses the file extension .md for KFX metadata files. Calibre will interpret that incorrectly as being a separate copy of the book in markdown format.

See the section "Importing from Kindle for PC" under "Importing KFX books into calibre" in the first post of the KFX Input plugin thread for information on how to ignore those .md files and make the import work better.

Originally Posted by tomsem

These .MD files are in binary format, no idea where they came from but it was not in the source folder, and .MD is supposed to be plain text.

They are there in the folder that makes up a book in KFX format along with .azw, .res, and .voucher files. Your screenshot shows that you have file extensions hidden in Windows file explorer. (That is the default unless changed.)