enlarge the math?

[stumped]

Originally Posted by DiapDealer

With EPUB3, they could have gone full-blown MathML to display the formulas. How did you do your EPUB3 to 2 "reduction"? .

i probably just did a calibre epub to epub convert to force epub2. or a simple edit and save in sigil ( those are the only 2 skills I have, for getting out of epub3 . i only knew my source is epub3 because sigil says so !

FYI, was poking inside of a different book that seemed to have a lot of hard to follow mathML stuff, but it all shows in green in sigil,as if it's commented out .
green means ignored?

[DiapDealer]

Sigil supports MathML (with Mathjax javascript), so if the code is commented out, that's in the epub itself.

[RbnJrg]

Originally Posted by stumped

my install is configured to be well behaved - easily done.

those SVG images display as intended in Moon +. no missing/empty boxes.

IF you actually read the whole thread and not just jump in to bitch about an android app you don't like, you would have seen that.


but they display at the authors' intended size and I'd like them about 20% bigger. that's what this thread was about

( they seem deliberately designed to appear inline as a bit smaller and greyer than the black text- no problem is, say, sigil book view on a 27 inch monitor but not ideal for my eyes on 10 inch or less devices.)
I have never seen svg defined like this though - where's the actual image ?
...

Well, after studying your .svg image, I can tell you the following:

1) Your .svg is an image made of "paths".
2) Your .svg is not a plain .svg, so it can produce issues with some ereaders.
3) Your .svg, even after converted as plain .svg, it has some instruccions not supported by ADE.

I attach an. epub with two .svg images: your original converted as plain .svg (test2.svg) and a new one made by me (test3.svg), also a plain one.

As you can see in the following screenshots, your image is not displayed under ADE but it looks ok under Sigil. So, that is a source of issues because if ADE has troubles to display your .svg, all ereaders based on RMSDK won't be able to show properly your formulas.



Since you have so many .svg images, to correct them in order that ADE can display them is not an alternative. So your best bet IMHO is to convert them (batch conversion) as .pgn using a conversor like Irfanwiew or ImageMagick.

Regards
Rubén

[stumped]

I found a better workaround, see next para. But I am puzzled by the amount ofnwork that someone-the publisher? - has devoted to creating a huge set of not very compatible images, brus stroke by brush stroke.

The workaround. Checking Amazon Kindle store ,they offer the book cheaply. I grabbed the free sample into kindle for android on my same tablet. The kindle version has equations what scale up and down when you change text size, and are blacker and at the same size as the text.
And presumably, if I bought it and converted to epub I would have something that works on all my devices.
Will look at the kindle sample book code later.

Or

[RbnJrg]

Originally Posted by stumped

But I am puzzled by the amount of work that someone-the publisher? - has devoted to creating a huge set of not very compatible images, brus stroke by brush stroke.

No, no many work. If the original formulas were written -for example- in latex (a common task if the ebook is about cientific matter and also was printed as .pdf), to convert the formulas as .svg it doesn't take a lot by using -for example- a tool like dvisvgm.

The workaround. Checking Amazon Kindle store ,they offer the book cheaply. I grabbed the free sample into kindle for android on my same tablet. The kindle version has equations what scale up and down when you change text size, and are blacker and at the same size as the text.
And presumably, if I bought it and converted to epub I would have something that works on all my devices.
Will look at the kindle sample book code later.

Amazon utilzes a different render engine. For .azw3/.kf8, employs Webkit (the same engine as Sigil); for .KFX Amazon only supports partially .svg images and converts them as .jpg; and it doesn't have any support for the .kf7 format.

Your workaround (convert a Kindle ebook into an epub) won't work for all epub reader devices, especially if the Kindle ebook you are going to download is a .kf8 one because Amazon won't modify your .svg (maybe you are a lucky man and the downloaded ebook is in .KFX format -I doubt- and all your formulas were converted in .jpg). If you, now, without any change, open your epub with PocketBook, your formulas will display well. But that is because PocketBook has two render engines: RMSDK and WebKit. Another reader device not based in WebKit will have issues with your .svgs. Convert your .svg as .pgn (a better format than .jpg for formulas), it's your best bet.

[stumped]

in the kindle sample - there are no .svg at all. they have been replaced with .gifs
( that's looking at the sample downloaded to K for PC then moved into calibre and converted for examination )
seems someone at amazon , or the publisher, realised what a pain in he posterior those svg constructs are, and had them replaced before putting the book up for sale.
so I shall buy the full version from amazon and move on.
( it was an interesting lesson ( in SVG, not in calculus! )

https://smile.amazon.co.uk/Calculus-...al-text&sr=1-1

[RbnJrg]

Originally Posted by stumped

...seems someone at amazon , or the publisher, realised what a pain in he posterior those svg constructs are...

Well, .svg images are the best option for formulas (if you can't use MathML). The problem here is that the .svg images in your epub are not in the standart format (plain .svg) and so, many ereaders can't display them properly. But you must not be with the idea that .svg images are always a bad option.

[KevinH]

FWIW, it is possible to use the command line and the stream editor "sed" to automate mass find replace upon any set of files (even 500 of them). So the option of fixing the svg is possible.

[st_albert]

Originally Posted by KevinH

FWIW, it is possible to use the command line and the stream editor "sed" to automate mass find replace upon any set of files (even 500 of them). So the option of fixing the svg is possible.

Aha! A 50+ year old solution to a modern problem! Somewhere, John Culliton is smiling:

If you are running any version of linux and don't use the command line
then you are hopping on one leg.

Albert

[JSWolf]

Originally Posted by RbnJrg

Well, after studying your .svg image, I can tell you the following:

1) Your .svg is an image made of "paths".
2) Your .svg is not a plain .svg, so it can produce issues with some ereaders.
3) Your .svg, even after converted as plain .svg, it has some instruccions not supported by ADE.

I attach an. epub with two .svg images: your original converted as plain .svg (test2.svg) and a new one made by me (test3.svg), also a plain one.

As you can see in the following screenshots, your image is not displayed under ADE but it looks ok under Sigil. So, that is a source of issues because if ADE has troubles to display your .svg, all ereaders based on RMSDK won't be able to show properly your formulas.

Attachment 182028Attachment 182029

Since you have so many .svg images, to correct them in order that ADE can display them is not an alternative. So your best bet IMHO is to convert them (batch conversion) as .pgn using a conversor like Irfanwiew or ImageMagick.

Regards
Rubén

There is a new problem. The sample formula is not the same once it's converted.

[Tex2002ans]

Originally Posted by KevinH

FWIW, it is possible to use the command line and the stream editor "sed" to automate mass find replace upon any set of files (even 500 of them). So the option of fixing the svg is possible.

Or Notepad++ has a "Find in Files", where you are able to Find/Replace over entire folders + filters.

So what I would do to handle the "gray maths issue":

1. Unzip the EPUB.

2. Notepad++ (or similar) to replace all rgb(0%,0%,0%) with rgb(0,0,0).

3. Repack EPUB. (Manually, or just open/save the EPUB in Calibre/Sigil.)

Originally Posted by RbnJrg

Well, after studying your .svg image, I can tell you the following:

1) Your .svg is an image made of "paths".
2) Your .svg is not a plain .svg, so it can produce issues with some ereaders.
3) Your .svg, even after converted as plain .svg, it has some instruccions not supported by ADE.

[...]

So your best bet IMHO is to convert them (batch conversion) as .pgn using a conversor like Irfanwiew or ImageMagick.

Agreed.

IF you want to stick with SVG, I know many times over the years we've discussed "Plain SVG" output from Inkscape.

Or you could convert to PNG formulas, while not as nice/crisp, at least it'll display in horrible non-standards-compliant readers such as Moon+ Reader.

All depends on what you're trying to accomplish, and what your source material is.

[AlanHK]

Originally Posted by stumped

I found a better workaround, see next para. But I am puzzled by the amount ofnwork that someone-the publisher? - has devoted to creating a huge set of not very compatible images, brus stroke by brush stroke.

Unlikely anyone ever looked at or edited the SVG code.

The printed book probably used Postscript/EPS/PDF graphics, and that was then exported automatically to SVG, which theoretically should give the same result.
In practice, not guaranteed. Many publishers just use the most primitive and also most compatible format, and convert to images (usually JPEG). That however balloons the size to get decent resolution.

As for the size of inline images (e.g. Chinese characters, symbols, etc.), of any kind, I make a class for these image with "height=1em" for a start and if that looks too small, bump it up a bit. Maybe some more complex ones need 2em, or even more.

Originally Posted by Tex2002ans

Or you could convert to PNG formulas, while not as nice/crisp, at least it'll display in horrible non-standards-compliant readers such as Moon+ Reader.

I generally use Kindle, and kindlegen converts PNG to either GIF or JPEG, enlarging it and usually degrading it. So for linework, I convert to GIF directly, which passes through Kindlegen.
I use Irfanview, usually reduce the number of colours to 4 (or 8 or 16 if there are greys), then adjust the palette so black is actually (0,0,0). Gives me very compact and sharp files.