Publisher PDF file size and export time questions

[Granddaddy]

I am delighted to find that I can open PDF files and reformat them to my liking. Here I'm trying to better understand how AFPub works and whether this is the kind of performance I should expect from AFPub when working with PDF files.

I opened an 891 page PDF document containing a favorite novel (Novel.pdf). The first page contains a small image of the original book cover. The remaining 890 pages are plain text in what appears to be Courier font. My goal is to reformat it with a more readable font and more pleasant format as a learning exercise in applying paragraph styles and all that.

Original PDF file is 5,468 KB (Novel.pdf)

Making no modifications at all I saved it as an afpub file whose size is 114,156 KB (Novel.afpub)

Making no modifications at all I exported the afpub document to PDF. The export took 8 minutes, producing a PDF file of 3,315 KB (Novel-AFPub.pdf).

QUESTIONS 1 - 3

1.) Is the enormous size of the afpub file expected?

2.) Is the export time on my relatively fast computer as expected?

3.) What might account for the much smaller size of the PDF produced by Affinity Publisher (Novel-AFPub.pdf) when compared to the original PDF (Novel.pdf)? I really do not know what affects PDF file sizes for simple text documents.

I copied the text from the original PDF file (Novel.pdf) into a Word document that saved as 3,665 KB (Novel-ExtractedToWord.doc). I printed the Word document to PDF using Foxit PDF Printer to produce PDF file of 6,123 KB (Novel-MSWord-FoxitPDFPrinter.pdf). At this point I was wondering what is in the AFPub file that makes it so much larger than the Word file for the same text?

I also wondered how much overhead is involved in storing text in an MS Word file as compared to a pure text file. I copied and saved the contents of (Novel-ExtractedToWord.doc) and saved as a text file of 3,085 KB (Novel-ExtractedToNotepad.txt). That's not much smaller than the Word file itself. This implies to me that Word adds very little overhead to the original 890 pages of text.

ADDED 6/23/2019

It's good that no one has answered yet as I woke up thinking of more questions. Also I have added a screen capture of the file names and their sizes to try to avoid confusing myself as well as any readers. I've added the file names to the text above also and added some clarifying text.

The only image in the original PDF file (Novel.pdf) is a scan of the book cover. The remaining 890 pages are plain text. AFPub says that image is 458x680 pixels.  I wondered how much this small image was contributing to the size of the afpub file produced by simply opening the original PDF file and saving as afpub.

Using PDFShaper, I deleted that first page with the image from Novel.pdf. Saving with PDFShaper increased the file to 6,054 KB (Novel-CoverPageDeleted-PDFShaper.pdf), slightly larger than the original PDF that contained the image.

I then opened the PDF file saved with PDFShaper (Novel-CoverPageDeleted-PDFShaper.pdf) in AFPub, did nothing at all to the file, then saved as afpub (Novel-CoverPageDeleted-PDFShaper.afpub). That file is 113,843 KB.

CONCLUSION

The 890 page novel is about 6 MB in PDF format and about 114 MB in afpub format whether or not the image is included.

QUESTION 4

Question 1.) above can be restated:

4.) Why is the AFPub file 19 times larger than the original PDF file?. Is the AFPub format really adding information equivalent to 16,000 pages of text even when no editing or styles or formatting or anything else has been applied to the text?

[Granddaddy]

I thought this one would be easy, but still no reply after 5 days. Perhaps someone will have a go at it now that the surge of activity following the AFPub launch is over.

I thought my Questions 1.)  and 2.) would be easy and require no more of an answer than Yes or No, though my restatement of Question 1.) as Question 4.) might require a more complicated answer.

I am interested in enormous Affinity file sizes because when some of us reported the ballooning sizes of afphoto files some changes were quickly made that yielded much smaller afphoto file sizes.

[Callum]

Hi Grandaddy,

1. it isn't expected but I couldn't say for sure without seeing the original PDF and the image included within it. We have a bug at the moment with our native files caching a high res version of any image within the document so this could be caused by this.

2. Again without seeing the file I can't say wether this is to be expected as I haven't tried exporting a 891 page document myself.

3. I can't say exactly what would make the file smaller but our PDF exporter is unique to us so its expected for it to produce different file sizes to other exporters.

4. See answer to question 1.

If you could provide some of the files in question I would be more than happy to look into this further with you.

Thanks

C

[Granddaddy]

Thanks for replying, Callum

As shown above in my screen capture from Windows File Explorer, The afpub file size is almost unaffected by the small graphic included on page 1 of the original PDF. The enormous afpub file results just from the text.

I just created a new blank document in AFPub using default settings. I created a Text frame on the first page and pasted into that text box the contents of Novel-ExtractedToNotepad.txt. I then saved that text as afpub. The resulting file is 165,382 KB. That's 54 times larger than the 3,085 KB plain text file I started with.

It's clear that the small image in the original file is not causing the afpub file to be so large. It is simply that pasting a relatively small amount of plain text without any formatting  into a text frame yields a monstrously huge afpub file.

I'm just curious as to why that should be so. Why can't an afpub file be saved in a more efficient way when it contains so little textual data.

If you provide a private upload location I'll upload Novel.pdf including the graphic on page 1 for you to play with. I cannot distribute this file to the public. 

[Callum]

Hi Grandaddy,

1) Id say the size is to be expected with the number of pages in the document mixed with the bug I mentioned previously it should be solved soon

2) Exporting the 891 page PDF took 20seconds for me. Please could you try exporting again to see if it is any quicker now?

3) See original answer.

C

[R C-R]

From what I can tell from my own tests & what has been mentioned before about what is stored in a native Affinity format file, could it be that one of the things that results in such large native file sizes is because the files are storing 3 pre-rendered thumbnails/icons (or mipmaps?) for each master & spread for the Pages Studio panel? Even without the cached high rez image bug, & even though each thumbnail by itself might require only a small amount of file space, the cumulative effect of storing them for each page of a several hundred page document could account for a considerably larger than expected file size.

Does that make any sense?

[Granddaddy]

12 hours ago, Callum said:

1) Id say the size is to be expected with the number of pages in the document mixed with the bug I mentioned previously it should be solved soon 2

OK. I don't have any idea what should be expected, it just seemed very, very large. I'd still like to know what AFPub adds to the text to make it so huge compared to the original text. I don't have enough experience with DTP software to speak further. R C-R above knows way more about what AFPub does than I do. I did just run a test with a 12 page Word document that has nothing on each page but the words Page x of xx. The Word 2000 document is 21 KB, which prints to PDF as 28 KB. Opening that PDF in AFPub and saving as afpub file increases the size to 40 KB. That's not an enormous difference. Perhaps sometime I'll make the Word file 60 times larger and see what happens to the exported PDF from AFPub. 

Another thread has just brought up enormous file sizes in AfPub and @stokerg has been replying. Perhaps @Callum and he should discuss this together.

https://forum.affinity.serif.com/index.php?/topic/91257-publisher-slow-saves/

 

12 hours ago, Callum said:

2) Exporting the 891 page PDF took 20seconds for me. Please could you try exporting again to see if it is any quicker now?

Still takes 8 minutes on what is a pretty high-end Dell XPS Tower Special Edition that is not quite two years old. I have been very satisfied with its speed in rendering Blu-ray discs. My export taking 24 times longer than yours seems unusual.

[Callum]

15 hours ago, Granddaddy said:

OK. I don't have any idea what should be expected, it just seemed very, very large. I'd still like to know what AFPub adds to the text to make it so huge compared to the original text. I don't have enough experience with DTP software to speak further. R C-R above knows way more about what AFPub does than I do. I did just run a test with a 12 page Word document that has nothing on each page but the words Page x of xx. The Word 2000 document is 21 KB, which prints to PDF as 28 KB. Opening that PDF in AFPub and saving as afpub file increases the size to 40 KB. That's not an enormous difference. Perhaps sometime I'll make the Word file 60 times larger and see what happens to the exported PDF from AFPub. 

Another thread has just brought up enormous file sizes in AfPub and @stokerg has been replying. Perhaps @Callum and he should discuss this together.

https://forum.affinity.serif.com/index.php?/topic/91257-publisher-slow-saves/

 

Still takes 8 minutes on what is a pretty high-end Dell XPS Tower Special Edition that is not quite two years old. I have been very satisfied with its speed in rendering Blu-ray discs. My export taking 24 times longer than yours seems unusual.

I have logged the file that is taking a long time to export with our developers we will see if they have any information.