Bloated epub file sizes?

[theducks]

Originally Posted by Joques

theducks, that is interesting. I have a few books that are especially sluggish. I will try that out.

More likely is:
1) horrible coding (unnecessary spans and cr*p. Usually from Word )
2) huge segments. a File per chapter is more nimble than one the reaches the 260K: must break (to be usable on older devices like mine) point.

[j.p.s]

Originally Posted by Joques

The ones that are in the ebook are 502x714, almost exactly half of your already shrunk files, and bad enough that several of the location names are really hard to make out. Edit: And those are the ones that take up 300kB of space.

I was skeptical when you had written "No compression at all, it seems.", but the uncompressed size for 714x502 is 350KB, not much larger than the 300KB for the files in the book. Uncompressed size for the files on the web site is 2.8MB, also not much larger than the 2.7MB JPEGs that are there.

I made 714x504 (same 0.3486328125 scaling to both dimensions) JPEG and PNG images from the 2048x1448 JPEGS on the web site, and got 142KB and 125KB sizes respectively. Both look about the same. Some of the text is difficult on a 96dpi monitor. I hate to think about trying it 167 dpi or 200dpi e-ink. 250 dpi e-ink or LCD is likely impossible. In any case, the publisher could have reduced the size of the book by almost a megabyte by going with PNG, maybe even a full megabyte starting with a clean image with no JPEG artifacts.

Since all 5 maps have a huge amount of identical content, an even more drastic bloat reduction will be possible when SVG support in EPUB readers gets decent. The same underlying image with only the constant text could be used for all the maps with the shaded regions and changed text as SVG overlays.

It looks like maybe the quality setting on the JPEG files was cranked well into diminishing returns territory with the result of all the worst features of JPEG with essentially none of the compression.

JPEG does not do well on regions of solid color or sharp edges. These figures are nothing but sharp edges and solid colors.

Originally Posted by Joques

Like I said - having hi-res maps on a tablet beside the ereader is ideal for me, more important really than having hi-res maps in the ebook file as it is so much work going back and forth between map and text.

It has been so long since I played with putting bare image files on a reader, that I had forgotten about the possibility. That does seem like a good workaround when such images are available.

[Joques]

theducks, Gravity's Rainbow is just four chapters for ~800 real pages - so the chapters are unreasonably big. Especially the third chapter, the biggest one. I like having a single page count for a single chapter, so I've tried keeping it intact, but my HD is just struggling too much now. The html file for chapter 3 is ~900kB.

In truth, it would have been better if the publisher had divided the chapters by all the smaller subdivisions within each chapter. Say, is there a simple way for me to do that? I tried playing around a little with the Calibre epub editer, but couldn't find a simple "split the file at this location" tool.

[Joques]

Wait! Sigil has a "split chapter" option. Very good, I'll do that. Probably make the damn thing easier to digest anyway.

[doubleshuffle]

I'm sure the Calibre editor must also have a function to set a split marker, but I don't know how it works there.

In Sigil you can set split markers with Ctrl+Shift+Return and then split at those markers with F6.

ETA: Too late...

[Joques]

I just used ctrl-return to split, didn't have to mark and execute.

[DomesticExtremis]

Splitting the chapters won't decrease the size of the epub, it will
in fact increase it due to all the header data and extra toc and opf
entries, though only a little bit.

[Joques]

Domestic, now we're not talking about the size of the epub any more, but the sizes of the individual html files (usually one per chapter) inside it :-) If the chapters get too large (like in GR), the ereader struggles, so then it is better to split them. Calibre will do this automatically, but usually in a less pleasingly aesthetic way.

[eschwartz]

There is a split button in calibre editor, at the bottom of the preview.

Also, I believe you can set an xpath to specify split locations in bulk.

[JSWolf]

One way to reduce the size of an eBook is to use jpegtran and pngout to reduce the size of the graphics without affecting the image quality at all. These programs remove the junk in the images and keep the pixels exactly as is to make the images smaller. The link below is for Windows and there are versions for Linux and OS X.

http://blog.stationfour.com/automati...on-in-windows/

[theducks]

Originally Posted by Joques

theducks, Gravity's Rainbow is just four chapters for ~800 real pages - so the chapters are unreasonably big. Especially the third chapter, the biggest one. I like having a single page count for a single chapter, so I've tried keeping it intact, but my HD is just struggling too much now. The html file for chapter 3 is ~900kB.

In truth, it would have been better if the publisher had divided the chapters by all the smaller subdivisions within each chapter. Say, is there a simple way for me to do that? I tried playing around a little with the Calibre epub editer, but couldn't find a simple "split the file at this location" tool.

The split tool is ONLY in Book view window. Lower corner: looks like a folded page.
Clicking that mags a Green line that you drag to the proper piont, when clicked sets and does the split. All you need to do then is Save.

[eschwartz]

Originally Posted by eschwartz

Also, I believe you can set an xpath to specify split locations in bulk.

^^ Accessible by right-clicking in a file editor (not preview) and selecting "Split at multiple locations". Uses xpath, and offers the calibre xpath wizard (with a link to the tutorial).

You can also set a shortcut key in Preferences.

[Joques]

Originally Posted by theducks

The split tool is ONLY in Book view window. Lower corner: looks like a folded page.
Clicking that mags a Green line that you drag to the proper piont, when clicked sets and does the split. All you need to do then is Save.

Ok, thanks. Still seems easier in Sigil, I have to say.

[eschwartz]

Originally Posted by Joques

Ok, thanks. Still seems easier in Sigil, I have to say.

What about Sigil is easier than using xpath? You can easily split on any marker you create. As well as any other describable element path.

[Joques]

I'll just put the cursor where I want the split, hit ctrl-return, and the split happens.