Please test these files on Paperwhite

[rostamiani]

I'm planning to buy a Kindle Paperwhite. Then asked someone to test my PDFs on his kindle and this is the answer :

"cannot select the text. whole page gets selected as its a image. beleive me its a pain to read this pdf on e-ink screen."

Wow ! Can anyone else test attached files on paperwhite ? These files are optimized using K2PDF Opt.

Is navigation fast ? Is texts selectable ?

Thanks

[DSpider]

No need for a Paperwhite. It has problems even on a computer, with Adobe Reader. The text on pages 3, 9, 12, 19, 22 (and maybe more) of the first attachement isn't selectable. It's like the program forgot to OCR it or something. Judging by the misalignment of some of the text, and the low quality (not scaling upon zooming), I'd say that these documents are just images with positional OCR behind them. It even OCR'ed the first figure on page 10.

I probably wouldn't put these on a reading device... The text looks very jagged and I can see why he/she would think it's painful to read. Opening it in Adobe Reader and going to File -> Properties, it says that the document is 3.37 x 4.43 inches, which InDesign translated as 242.64 x 318.96 pixels! That's very, very low on a Paperwhite (758×1024, according to Wikipedia), or even the rest of the Kindles (600x800).

[rostamiani]

Originally Posted by DSpider

No need for a Paperwhite. It has problems even on a computer, with Adobe Reader. The text on pages 3, 9, 12, 19, 22 (and maybe more) of the first attachement isn't selectable. It's like the program forgot to OCR it or something. Judging by the misalignment of some of the text, and the low quality (not scaling upon zooming), I'd say that these documents are just images with positional OCR behind them. It even OCR'ed the first figure on page 10.

I probably wouldn't put these on a reading device... The text looks very jagged and I can see why he/she would think it's painful to read. Opening it in Adobe Reader and going to File -> Properties, it says that the document is 3.37 x 4.43 inches, which InDesign translated as 242.64 x 318.96 pixels! That's very, very low on a Paperwhite (758×1024, according to Wikipedia), or even the rest of the Kindles (600x800).

Thanks
I Created this PDFs with K2PDF Opt to optimize it for Kindle:
http://www.mobileread.com/forums/sho...d.php?t=144711
The original file is here :
http://www.autodeskresearch.com/pdf/eg01.pdf

I know that the resolution is low because I created it for my Nook Simple Touch

1 . With Nook I cannot select text .even on selectable pages ! Can Paperwhite handle this ?
2 . And I don't know if navigation speed is enough on Paperwhite

[DSpider]

The "Nook Simple Touch" also has a resolution of 600x800, so it probably doesn't look very good on it either.

Source looks very good except that, well, you're trying to convert a PDF to a PDF... And that's very bad. PDF is meant as a final destination. For a proper conversion, you would need the original Word/OpenOffice/LibreOffice document (or whatever the author used), but then again, the equations aren't helping either - they will need to be converted to GIF/PNG, because Kindles don't do SVG. The rest of the images should be kept as colour images, in case someone reads it on a Kindle Fire, not automatically converted to 1 bit images (black).

Heh. If you were hoping for a one click conversion solution, there is none. Personally, I would extract the text with ABBYY FineReader, proofread it, save as RTF, run a custom Word macro that gets rid of the styles (but keeps italics, bolds, superscripts, places the footnotes as inline text, etc), import into InDesign, redo the layout, save it as an ePub, convert the ePub to Mobi, then proofread the final product and adjust as necessary if something doesn't look right.

[markom]

Originally Posted by rostamiani

Thanks
I Created this PDFs with K2PDF Opt to optimize it for Kindle:
http://www.mobileread.com/forums/sho...d.php?t=144711
The original file is here :
http://www.autodeskresearch.com/pdf/eg01.pdf

I know that the resolution is low because I created it for my Nook Simple Touch

1 . With Nook I cannot select text .even on selectable pages ! Can Paperwhite handle this ?
2 . And I don't know if navigation speed is enough on Paperwhite

With Sony Prs-T1 I can select word but it then goes to highlight the whole line, not just that particular word, so I had to enter word manually in dictionary or delete other parts.

I've done OCR again in Acrobat and then I was able to select single words.

Speed was normal.

[willus]

Originally Posted by rostamiani

Can anyone else test attached files on paperwhite ? These files are optimized using K2PDF Opt.

Is navigation fast ? Is texts selectable ?

It looks like you used OCR on the k2pdfopt conversion of the first attachment, because most of the text in that one is selectable using Sumatra PDF reader on my PC (except for the text inside of a figure, which k2pdfopt does not OCR at this point). I'm not sure how paperwhite will handle it--hopefully it will also be selectable there. The second one will not be selectable as k2pdfopt has no OCR for that language/font (yet)--I've only done English. But that's an issue with how you converted it, not with the paperwhite. Is the second attachment selectable in the source file? You can try using the native PDF output option of k2pdfopt if it is, and that should keep it selectable in the output file, but then you won't be able to use text wrapping / re-flow in the conversion. (You can also try the native PDF option on the first attachment--use the -mode 2col option).

The paperwhite looks pretty fast at handling PDFs--starting to approach tablet capability, with "pinch to zoom" feature. Check out this excellent "triparison" video review at about the 13-minute mark, where the author shows how three "glow" e-readers compare at reading PDF files (paperwhite on the left).

[willus]

Originally Posted by DSpider

...The text on pages 3, 9, 12, 19, 22 (and maybe more) of the first attachement isn't selectable. It's like the program forgot to OCR it or something.

Most of the text on these pages (3, 9, ...) is selectable using Sumatra PDF on my PC. K2pdfopt will not OCR text within a figure (at this time).

Originally Posted by DSpider

Judging by the misalignment of some of the text, and the low quality (not scaling upon zooming), I'd say that these documents are just images with positional OCR behind them.

K2pdfopt does OCR by using PDF text directives to put text invisibly behind the bitmap-rendered text (it can be made visible by selecting the right options), just like most scanning software which uses OCR. The resolution of the converted file is entirely selectable. Using "-dr 2", for example, will double the resolution of the converted file.

Originally Posted by DSpider

Opening it in Adobe Reader and going to File -> Properties, it says that the document is 3.37 x 4.43 inches, which InDesign translated as 242.64 x 318.96 pixels! That's very, very low on a Paperwhite (758×1024, according to Wikipedia), or even the rest of the Kindles (600x800).

I think it's probably 242.64 x 318.96 points, not pixels. They are not the same. There are 72 points per inch (hence 3.37 inches = 242.6 points--pixel resolutions would not typically be listed as fractional values, in any case). The page images in the PDF file attachments are 560 x 735 pixels because that is the Kindle 2 display resolution. Like I said, using "-dr 2" would increase the resolution to 1120 x 1470 pixels, but would also cause the output file to take up more room (bigger file size in bytes).