Minor question GIF in zip output

Mohan-V-Allied · 05-16-2022, 07:40 AM

Hello,
Have a rather minor issue concerning converting PDF's. Have already found a work around; but it means using another program and adds time which adds up if all the files are converted.

Want to convert PDF's but as the pages are typewriten cannot convert to any format, other than to a zip file which produces a zip folder containing images. Can then take these images and put them through a OCR program.

The issue is the image output is PNG. Having run some tests (saving text as various image formats) found that BMP has the most errors. PNG has a lot less but GIF has half the amount of PNG.

So is it possible to have the images from the PDF in the zip folder out as Gif? This would be helpful as it would save some considerable time.

Thanks in advance.
Mohan

kovidgoyal · 05-16-2022, 09:14 AM

No, I'm afraid not, PNG is a strictly superior format to GIF (apart from animations) as such calibre has no code for converting to GIF only from GIF.

Sarmat89 · 05-16-2022, 09:42 PM

All these are lossless formats, so it cannot affect the OCR.

aborel · 05-16-2022, 10:28 PM

Quote:

Originally Posted by Sarmat89

All these are lossless formats, so it cannot affect the OCR.

But GIF is limited to 256 colors, so converting an image with a larger palette into this format would actually be lossy. This could have an impact, I suppose.

Mohan-V-Allied · 05-19-2022, 07:27 AM

Thanks for your input.

For converting the typewritten PDF's the fewer colours the better. By converting a image with a greater palette such as BMP to an image with a smaller palette you are basically removing extraneous details leaving just the basic text. At least this is what happened in the tests. For the OCR this works the best.

Mohan-V-Allied · 05-19-2022, 07:28 AM

Thanks for you input.

By running a few simple tests found that the OCR output is affected by the higher quality of details. BMP having the greater details, picks up everything. This includes marks, smudges, text which has been tipexed out. As these are typewritten the letters are not printed sharply, so rn would be shown as m or l shown as "i etc. Hence more errors.

Mohan-V-Allied · 05-19-2022, 07:29 AM

Thanks for your time, this helps make these forums amongst the better ones. Have found a lot of information here from a number of posters.

Agree, that the PNG images are superior, but it was just that the GIF produces less errors in conversion. But as this is only the typewritten PDF files and have another program which will convert the PNG images to GIF images thought it would not hurt to ask.

05-16-2022, 07:40 AM	#1
Mohan-V-Allied Member Posts: 14 Karma: 10 Join Date: Jul 2013 Device: Kobo touch	Minor question GIF in zip output Hello, Have a rather minor issue concerning converting PDF's. Have already found a work around; but it means using another program and adds time which adds up if all the files are converted. Want to convert PDF's but as the pages are typewriten cannot convert to any format, other than to a zip file which produces a zip folder containing images. Can then take these images and put them through a OCR program. The issue is the image output is PNG. Having run some tests (saving text as various image formats) found that BMP has the most errors. PNG has a lot less but GIF has half the amount of PNG. So is it possible to have the images from the PDF in the zip folder out as Gif? This would be helpful as it would save some considerable time. Thanks in advance. Mohan

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Zip support question	MaxStirner	KOReader	3	03-08-2021 03:24 PM
Question: Multiple files in one zip folder	BMaloney	Library Management	8	03-24-2013 08:53 AM
Zip output format numbering	xlx	Conversion	2	06-14-2011 12:48 PM
650 First Impressions (synopsis, very good, one minor question)	sgtpokey	Sony Reader	5	10-25-2010 06:31 PM
iLiad Newbie Question on updates (V2.12 and zip file)	Teneflin	iRex Developer's Corner	2	02-10-2008 09:05 PM

05-16-2022, 09:14 AM	#2
kovidgoyal creator of calibre Posts: 45,253 Karma: 27110894 Join Date: Oct 2006 Location: Mumbai, India Device: Various	No, I'm afraid not, PNG is a strictly superior format to GIF (apart from animations) as such calibre has no code for converting to GIF only from GIF.

05-16-2022, 09:42 PM	#3
Sarmat89 Fanatic Posts: 515 Karma: 2268308 Join Date: Nov 2015 Device: none	All these are lossless formats, so it cannot affect the OCR.

05-19-2022, 07:27 AM	#5
Mohan-V-Allied Member Posts: 14 Karma: 10 Join Date: Jul 2013 Device: Kobo touch	Thanks for your input. For converting the typewritten PDF's the fewer colours the better. By converting a image with a greater palette such as BMP to an image with a smaller palette you are basically removing extraneous details leaving just the basic text. At least this is what happened in the tests. For the OCR this works the best.

05-19-2022, 07:28 AM	#6
Mohan-V-Allied Member Posts: 14 Karma: 10 Join Date: Jul 2013 Device: Kobo touch	Thanks for you input. By running a few simple tests found that the OCR output is affected by the higher quality of details. BMP having the greater details, picks up everything. This includes marks, smudges, text which has been tipexed out. As these are typewritten the letters are not printed sharply, so rn would be shown as m or l shown as "i etc. Hence more errors.

05-19-2022, 07:29 AM	#7
Mohan-V-Allied Member Posts: 14 Karma: 10 Join Date: Jul 2013 Device: Kobo touch	Thanks for your time, this helps make these forums amongst the better ones. Have found a lot of information here from a number of posters. Agree, that the PNG images are superior, but it was just that the GIF produces less errors in conversion. But as this is only the typewritten PDF files and have another program which will convert the PNG images to GIF images thought it would not hurt to ask.

Advert

Advert