Bloated epub file sizes?

Joques · 07-17-2014, 06:42 AM

I am curious, and a quick search on the forum didn't yield an answer:

Why is the so little correlation between the length of a book and its file size? I am not talking about books containing lots of pictures here, just formatted text in a regular novel. One book of average length might weigh in at 400kB, while another, of similar length and also having no images in there, might be 2,5 MB.

What is it that bloats these file sizes so?

shalym · 07-17-2014, 06:47 AM

Quote:

Originally Posted by Joques

I am curious, and a quick search on the forum didn't yield an answer:

Why is the so little correlation between the length of a book and its file size? I am not talking about books containing lots of pictures here, just formatted text in a regular novel. One book of average length might weigh in at 400kB, while another, of similar length and also having no images in there, might be 2,5 MB.

What is it that bloats these file sizes so?

I'm not sure, because I don't edit epubs, but I would say probably bad formatting? I would think that if the styles are declared for every page instead of just once for the whole book it would cause bloating. There could also be embedded fonts in the book.

I'm sure someone who DOES edit epubs will come along soon and either confirm what I said or ridicule me for not knowing what I'm talking about

Shari

HarryT · 07-17-2014, 06:56 AM

Please post in the correct forum. Moved to the ePub file format forum.

HarryT · 07-17-2014, 07:01 AM

You need to edit the book in an ePub editor such as Sigil or Calibre and see what's causing the bloat. It's probably caused by an over-proliferation of CSS styles.

Toxaris · 07-17-2014, 07:45 AM

If there is a cover, than I should look at the cover size for sure. An average book with only text can not be 2.5MB. Using a different compression level will have some influence.
Bad formatting will have impact for sure, but for an average book to be 2.5MB it must be seriously bad formatted...

Doonge · 07-17-2014, 08:15 AM

Embedded font bloat books.

doubleshuffle · 07-17-2014, 08:18 AM

Yep. Especially CharisSIL, which is embedded in many retail epubs.

JSWolf · 07-17-2014, 08:21 AM

Quote:

Originally Posted by doubleshuffle

Yep. Especially CharisSIL, which is embedded in many retail epubs.

But using Calibre to subset embedded fonts really does cut down the file size. Also, if the cover is either very large and/or not all that compressed, it can be a large file size. So reducing the cover image size and recompressing can help.

doubleshuffle · 07-17-2014, 08:29 AM

Quote:

Originally Posted by JSWolf

But using Calibre to subset embedded fonts really does cut down the file size.

True. Though I usually throw it out completely. I've got CharisSIL sideloaded on my reader anyway, no need to have it in lots of different books over and over again.

Tex2002ans · 07-17-2014, 08:37 AM

Overly large/badly compressed covers/images.
A "phantom cover"
- I have come across a few purchased books where there is a useless second cover inside of the book, which isn't referenced or anything, just takes up space.
When the EPUB was packaged, they didn't compress it, or didn't compress to the maximum level (this one just baffles me).
- I have come across a few purchased books which had zero compression on them.
Bad code/formatting could bloat the EPUB (oh the horrors I have seen).
Fonts that are not subsetted
Fonts that aren't used/referenced
- For example, including a bold font when no bold is used, or a smallcaps font when no smallcaps are used.

JSWolf · 07-17-2014, 09:16 AM

Quote:

Originally Posted by Tex2002ans

Overly large/badly compressed covers/images.
A "phantom cover"
- I have come across a few purchased books where there is a useless second cover inside of the book, which isn't referenced or anything, just takes up space.
When the EPUB was packaged, they didn't compress it, or didn't compress to the maximum level (this one just baffles me).
- I have come across a few purchased books which had zero compression on them.
Bad code/formatting could bloat the EPUB (oh the horrors I have seen).
Fonts that are not subsetted
Fonts that aren't used/referenced
- For example, including a bold font when no bold is used, or a smallcaps font when no smallcaps are used.

Modify ePub can get rid of unused images and subsetting gets rid of unused fonts and reduce the used fonts in size.

Thumbnails are a common waste of space and Modify ePub deletes those no problem.

theducks · 07-17-2014, 11:28 AM

Quote:

Originally Posted by Tex2002ans

Overly large/badly compressed covers/images.
A "phantom cover"
- I have come across a few purchased books where there is a useless second cover inside of the book, which isn't referenced or anything, just takes up space.
When the EPUB was packaged, they didn't compress it, or didn't compress to the maximum level (this one just baffles me).
- I have come across a few purchased books which had zero compression on them.
Bad code/formatting could bloat the EPUB (oh the horrors I have seen).
Fonts that are not subsetted
Fonts that aren't used/referenced
- For example, including a bold font when no bold is used, or a smallcaps font when no smallcaps are used.

Your list is fairly complete, I see you have been working on this issue a few times

I will add: Retailer demanded Bloat. Thumbnail covers, Hi-Def covers.

The best gain on sub-setting a font would be for those 'Display' fonts that might be only used for Chapter titles, Initial Letters.

Is there a EASY way to later determine that the books font file has been sub-setted ( I was thinking of a Quality check PI type test)?

dwig · 07-17-2014, 02:21 PM

Quote:

Originally Posted by theducks

Your list is fairly complete, I see you have been working on this issue a few times

I will add: Retailer demanded Bloat. Thumbnail covers, Hi-Def covers.

The best gain on sub-setting a font would be for those 'Display' fonts that might be only used for Chapter titles, Initial Letters.
...

... rr any OpenType font containing multiple character sets, especially those containing unused complex sets (e.g. the various Japanese kana, Chinese, ...) which can be huge when complete. With these, subsetting a font used for the main body text can result in a massive size reduction.

JSWolf · 07-17-2014, 02:49 PM

CharisSIL is a large font because of how many extended characters is contains. Most of that is unused. So when you subset, you get rid of the fonts not used and the characters not used. Most of the time, you will get rid of the bold italic version of a font. You can cut down the size of CharisSIL (on average) to between 200-300K verses about 1.2-1.3MB per font file.

As for the large graphics, this is because people read tablets with high resolution screens and tiny 800x600 sized graphics just don't cut it all that well.

Joques · 07-18-2014, 03:07 AM

Thanks everybody for the responses! I am sorry I posted in the wrong forum - but it was because though I personally use epubs, I assumed this was a universal problem that plagued all formats.

I _do_ try to find the highest-res cover art that I can for all my books, but when converted to B/W they don't take up all that much space. I'll experiment a little with removing embedded fonts with Calibre, but cleaning up bloated CSS styles for 900 books is not going to happen :-)

07-17-2014, 06:42 AM	#1
Joques Addict Posts: 294 Karma: 107414 Join Date: May 2013 Device: Kobo Glo	Bloated epub file sizes? I am curious, and a quick search on the forum didn't yield an answer: Why is the so little correlation between the length of a book and its file size? I am not talking about books containing lots of pictures here, just formatted text in a regular novel. One book of average length might weigh in at 400kB, while another, of similar length and also having no images in there, might be 2,5 MB. What is it that bloats these file sizes so?

07-17-2014, 08:37 AM	#10
Tex2002ans Wizard Posts: 2,306 Karma: 13057279 Join Date: Jul 2012 Device: Kobo Forma, Nook	Overly large/badly compressed covers/images. A "phantom cover" I have come across a few purchased books where there is a useless second cover inside of the book, which isn't referenced or anything, just takes up space. When the EPUB was packaged, they didn't compress it, or didn't compress to the maximum level (this one just baffles me). I have come across a few purchased books which had zero compression on them. Bad code/formatting could bloat the EPUB (oh the horrors I have seen). Fonts that are not subsetted Fonts that aren't used/referenced For example, including a bold font when no bold is used, or a smallcaps font when no smallcaps are used.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Calibre ebook-viewer.exe changes EPUB file sizes?	avid01	Calibre	23	04-11-2018 04:24 AM
File sizes - why the difference?	Araucaria	Sigil	4	11-22-2011 07:52 PM
Book/file sizes in Calibre	cavgirl	Calibre	2	11-12-2010 08:15 PM
Epub file sizes	jerryleejr	Sony Reader	6	07-28-2008 03:09 PM

07-17-2014, 06:56 AM	#3
HarryT eBook Enthusiast Posts: 85,560 Karma: 93980341 Join Date: Nov 2006 Location: UK Device: Kindle Oasis 2, iPad Pro 10.5", iPhone 6	Please post in the correct forum. Moved to the ePub file format forum.

07-17-2014, 07:01 AM	#4
HarryT eBook Enthusiast Posts: 85,560 Karma: 93980341 Join Date: Nov 2006 Location: UK Device: Kindle Oasis 2, iPad Pro 10.5", iPhone 6	You need to edit the book in an ePub editor such as Sigil or Calibre and see what's causing the bloat. It's probably caused by an over-proliferation of CSS styles.

07-17-2014, 07:45 AM	#5
Toxaris Wizard Posts: 4,520 Karma: 121692313 Join Date: Oct 2009 Location: Heemskerk, NL Device: PRS-T1, Kobo Touch, Kobo Aura	If there is a cover, than I should look at the cover size for sure. An average book with only text can not be 2.5MB. Using a different compression level will have some influence. Bad formatting will have impact for sure, but for an average book to be 2.5MB it must be seriously bad formatted...

07-17-2014, 08:15 AM	#6
Doonge Connoisseur Posts: 80 Karma: 1184732 Join Date: Nov 2013 Device: Kobo Glo	Embedded font bloat books.

07-17-2014, 08:18 AM	#7
doubleshuffle Unicycle Daredevil Posts: 13,944 Karma: 185432100 Join Date: Jan 2011 Location: Planet of the Pudding Brains Device: Aura HD (R.I.P. After six years the USB socket died.) tolino shine 3	Yep. Especially CharisSIL, which is embedded in many retail epubs.

07-17-2014, 02:49 PM	#14
JSWolf Resident Curmudgeon Posts: 80,740 Karma: 150249619 Join Date: Nov 2006 Location: Roslindale, Massachusetts Device: Kobo Libra 2, Kobo Aura H2O, PRS-650, PRS-T1, nook STR, PW3	CharisSIL is a large font because of how many extended characters is contains. Most of that is unused. So when you subset, you get rid of the fonts not used and the characters not used. Most of the time, you will get rid of the bold italic version of a font. You can cut down the size of CharisSIL (on average) to between 200-300K verses about 1.2-1.3MB per font file. As for the large graphics, this is because people read tablets with high resolution screens and tiny 800x600 sized graphics just don't cut it all that well.

07-18-2014, 03:07 AM	#15
Joques Addict Posts: 294 Karma: 107414 Join Date: May 2013 Device: Kobo Glo	Thanks everybody for the responses! I am sorry I posted in the wrong forum - but it was because though I personally use epubs, I assumed this was a universal problem that plagued all formats. I _do_ try to find the highest-res cover art that I can for all my books, but when converted to B/W they don't take up all that much space. I'll experiment a little with removing embedded fonts with Calibre, but cleaning up bloated CSS styles for 900 books is not going to happen :-)

Advert

Advert