Retaining images/figures in RTF/DOC derived from PDF?

[haguilar]

Hi everyone - I bought the PRS-505 primarily to view scientific papers in PDF format on it (I know now this may not have been the ideal device for this purpose, but I'm hopeful!), but every type of conversion I've tried yields less than optimal results.

I've tried converting to html/lrf/doc/rtf directly from the PDF, I've used practically every piece of software I could find by reading numerous threads on this site, and I've tried using a combination of different steps (pdf to html to lrf for example) to get to the final file format, and while I've found a couple of transformations that seem to work in terms of the actual text in my PDF files, I can't seem to find a way to get the images to stay integrated with the text. Is this not possible with the sony reader, or is there a key way to do this that I'm overlooking?

Thanks,

H

[soilwork]

Did you try PDFLRF?
https://www.mobileread.com/forums/showthread.php?t=13135
This tool will directly convert PDF to LRF and makes PDF documents quite readable on Sony Reader. Actually, I got my PRS 505 to read journal articles in PDF and I found most of the articles are quite readable after conversion using PDFLRF.

[haguilar]

yep, tried that.

But perhaps I wasn't using the right settings. I did at one point attempt to use ghostscript (did you use this? I'm not even really sure what it does to be honest) but after installation I kept getting a bunch of error dialog boxes relating to ghostscript being displayed after every page that was processed.

Can you tell me which boxes you check/uncheck when using this tool for your papers? portrait or landscape? (EDIT - after a little more playing with it, I can see landscape could be promising but optimally I'd like to see it in portrait mode for easier reading).

The one thing I could not figure out is how to remove some intervening pages that are generated when converting PDFs that have both 2 column formatting and one column formatting on the same page (so for example, the top half of a page could be the title/author/university/abstract information and it spans the entire page horizontally, while the bottom of the page begins in the 'usual' two column format with the introduction to the paper. - I found in this situation what the PDFLRF tool does is it begins chopping up the PDF into 'quarters' but it doesn't quite get the order of things right, so there are some intervening pages which make no sense - a white page or a page containing part of the title/author info inserted between pages containing the first/second columns of the actual text of the paper.)

I guess I'm just looking for the optimal settings so I can convert a large number of PDFs, load them onto my 505 and read them when I get a chance without having to individually convert/optimze settings (maybe that's asking for too much, I don't know, this is my first reader).

H

[soilwork]

Originally Posted by haguilar

I did at one point attempt to use ghostscript (did you use this? I'm not even really sure what it does to be honest) but after installation I kept getting a bunch of error dialog boxes relating to ghostscript being displayed after every page that was processed.

I guess you may not have Ghostscript installed on your computer. You can the program from here.
http://pages.cs.wisc.edu/~ghost/
However, from my experience, I don't see any difference in quality whether you use ghostscript or not. I usually use non-ghostscript option. So, I would recommend not to bother with 'Ghostscript' unless you need it for another reason (LaTeX for example).

Can you tell me which boxes you check/uncheck when using this tool for your papers? portrait or landscape? (EDIT - after a little more playing with it, I can see landscape could be promising but optimally I'd like to see it in portrait mode for easier reading).

My rules of thumb are as follows.

1) With 2 column articles, I use '2col' option (naturally!)
1-1) If the 2 column articles have tables or figures spanning across columns, I just create another copy of the article in 'landscape' without 'join page' option.
1-2) If the article has such problem only with the abstract page, I just convert the 1 page in 'landscape' as a separate file.
[edit: When you create multiple LRF copies of PDF, it is good idea to give different 'title' depending on whether it is 2 col or landscape version. Without different title, it is hard to distinguish between the two without opening both files. For example, if the title is "Discrete Choice", I make the title for 2column version as "Discrete Choice - 2c" and for landscape version as "Discrete Choice - ls" ]

2) With 1 column articles,
2-1) If the fonts are relatively big and/or the margins on the sides are wide, I used 'portrait'. So typical LaTeX documents with wide margin will be good candidates. (e.g, http://elsa.berkeley.edu/choice2/pre.pdf )
2-2) If the fonts are small or the margins on the sides are narrow, then I use 'landscape'

Overall, other than documents with extremely small fonts and very narrow side margins, the documents are quite readable.

The one thing I could not figure out is how to remove some intervening pages that are generated when converting PDFs that have both 2 column formatting and one column formatting on the same page (so for example, the top half of a page could be the title/author/university/abstract information and it spans the entire page horizontally, while the bottom of the page begins in the 'usual' two column format with the introduction to the paper. - I found in this situation what the PDFLRF tool does is it begins chopping up the PDF into 'quarters' but it doesn't quite get the order of things right, so there are some intervening pages which make no sense - a white page or a page containing part of the title/author info inserted between pages containing the first/second columns of the actual text of the paper.)

Well, I have the same problem. So I use my rule of thumb 1-1) and 1-2) and create multiple versions of the documents. On the reader, I switch back and forth using 'bookmark' function. (Keep pressing 'bookmark' function will bring up all bookmarks of all the documents on the reader & SD card) I know it is quite inconvenient but at least this way, I do not have to pay extra 400$ to get iliad.

I guess I'm just looking for the optimal settings so I can convert a large number of PDFs, load them onto my 505 and read them when I get a chance without having to individually convert/optimze settings (maybe that's asking for too much, I don't know, this is my first reader).

Well, it would be great if it is possible but, judging from the inconsistent quality of PDF files, I guess the conversion should be one by one. Even articles from one vendor (e.g. Proquest) has varying quality. Sometimes articles have clear black border from scanning but not always. In such cases, I had to apply pre-trim option to get it properly converted.

Good luck on reading papers on PRS!

[haguilar]

Originally Posted by soilwork

I guess you may not have Ghostscript installed on your computer. You can get free from here.
http://pages.cs.wisc.edu/~ghost/
However, from my experience, I don't see any difference in quality whether you use ghostscript or not. I usually use non-ghostscript option. So, I would recommend not to bother with 'Ghostscript' unless you need it for another reason (LaTeX for example).


My rules of thumb are as follows.

1) With 2 column articles, I use '2col' option (naturally!)
1-1) If the 2 column articles have tables or figures spanning across columns, I just create another copy of the article in 'landscape' without 'join page' option.
1-2) If the article has such problem only with the abstract page, I just convert the 1 page in 'landscape' as a separate file.

2) With 1 column articles,
2-1) If the fonts are relatively big and/or the margins on the sides are wide, I used 'portrait'. So typical LaTeX documents with wide margin will be good candidates. (e.g, http://elsa.berkeley.edu/choice2/pre.pdf )
2-2) If the fonts are small or the margins on the sides are narrow, then I use 'landscape'

Overall, other than documents with extremely small fonts and very narrow side margins, the documents are quite readable.


Well, I have the same problem. So I use my rule of thumb 1-1) and 1-2) and create multiple versions of the documents. On the reader, I switch back and forth using 'bookmark' function. (Keep pressing 'bookmark' function will bring up all bookmarks of all the documents on the reader & SD card) I know it is quite inconvenient but at least this way, I do not have to pay extra 400$ to get iliad.



Well, it would be great if it is possible but, judging from the inconsistent quality of PDF files, I guess the conversion should be one by one. Even articles from one vendor (e.g. Proquest) has varying quality. Sometimes articles have clear black border from scanning but not always. In such cases, I had to apply pre-trim option to get it properly converted.

Good luck on reading papers on PRS!

Ok, thanks very much for the advice. It sounds like there is no magical combination of options/conversion tools that makes all PDFs entirely readable, but I'll give some of these a try and see what happens.

H