UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   General Vintage Technology Discussions (https://www.vintage-radio.net/forum/forumdisplay.php?f=29)
-   -   Scanning of Manuals & Circuit Diagrams? (https://www.vintage-radio.net/forum/showthread.php?t=181691)

Chris55000 6th Jul 2021 6:21 pm

Scanning of Manuals & Circuit Diagrams?
 
Hi!

This is a technical question, NOT a copyright one – I've no time for copyright moguls!

Can anyone recommend the best settings for VueScan Pro (full professional licence bought) that will give the best settings for archiving to PDF please? Is it possible to OCR them like World Radio History does with it's magazine scans?

For example, what would be the best page settings for a TQ coloured PCB overlay – is it possible to scan these in Vuescan in colour and still get a white background – whenever I do it I always get an off–white or yellow background tint!

Chris Williams

Uncle Bulgaria 6th Jul 2021 6:31 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
I don't know the settings, but I have made scans of text better than most downloaded manuals, and photographic prints suitable to be enlarged on our Epson all-in-one printer/scanner. The white point settings on the built-in software worked fine.

I do OCR on Adobe Acrobat, but a cursory search of the Windows Store shows there are innumerable freeware or open source options, and no doubt others if you're not on Windows.

I usually do a downsampling test after scanning a file with a reasonable number of pages. It's possible to reduce tens of MB to a couple without noticeable loss of quality when doing this through Acrobat, though I've not tried with other programmes.

Seeing your message has reminded me that I said I'd look at a pattern brush for your circuit diagrams with Affinity - I'm sorry, it'd completely slipped my mind!

Edit: Speaking of Affinity, is that where you're getting the off-white tint? I had a problem when I first installed the suite where Affinity white was yellowish. There's a thread on the Affinity forum about this.

Chris55000 6th Jul 2021 6:43 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
Hi!

I have bought an Affinity Designer Licence but the Workbook is a bit difficult, so I'm using Vue Scan with an Epson ET–4500, both default settings and I've no idea if there's any better ones!

The disc that was supplied with the ET–4500 didn't appear to any OCR/Scan software supplied, hence me buying Vue Scan!

Chris Williams

Electronpusher0 7th Jul 2021 5:53 am

Re: Scanning of Manuals & Circuit Diagrams?
 
Bobdger and I have been scanning magazine inserts, mainly PW, and posting them here:

https://www.vintage-radio.net/forum/...d.php?t=169177

I have also sent them to World Radio History.

The larger pages Bobdger scanned and I scanned the A4 ones
I use Vuescan pro with the following settings:

200dpi
24 bit RGB
output - pdf
pdf size reduction - 1
pdf compression - on
magification - 100%

I do not try and adjust the colour balance, this does mean that some pages have a yellowish background but most pages I scan have yellowed with age anyway.

I have not tried ocr but vuescan does have a tick box setting "PDF OCR text"

Peter

wireman 7th Jul 2021 7:00 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
I have been investigating software to be used for OCR treatment of scans of decades of programmes for a local event.

I'm close to purchasing a license for ABBYY FineReader.

MotorBikeLes 7th Jul 2021 7:44 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
I am using linux, so may not help you, but I use "gimagereader" software for OCR. Written and supported by one guy, who is helpful if needed. No idea if available for other op sys's.
To operate on pdfs with "select" and other tools, I use okular.
With "imagemagik" I can shrink or change format very easily.
There is a pdf operator that will build up numerous pfds into one (I would need to search to remember what it is if anybody is interested.
In the past, I have created manuals using mixed pages of scan, pdf,drawings etc with Libre Office. Text I OCR directly in, drawings I often open in L.O. "Draw", manipulate if needed, then paste into L.O. "Writer". However I find i can open with Okular then transfer directly to "Writer".
L.O.IS available for Windoze, I used it 15 years ago before I moved to Linux.
Les.

Julesomega 7th Jul 2021 8:06 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
The free paint dot net will save multiple layers as jpg's, and there are many online converters that will combine all the jpg's into one PDF

WME_bill 7th Jul 2021 8:50 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
Scanning Manuals.
I seem to have three different scanners, acquired over the past 10 years.
The main one is a Kyocera A3 printer-scanner which scans B&W with automatic feed, single or both sided. It is networked and now I know it's foibles, works well. It prints out or scans larger or smaller or almost any variation of size, and can rotate a scanned document. It was a proper commercial machine, and really a luxury for a home office, but never seems to need a laser cartridge refill, so marvellously economical to run, A3, A4, foolscap, B4, B5 etc. I use a cheap Colour Laser-Jet 1215 for colour printing A4, but the refills are so expensive, I avoid using it unless I have to have colour.

For my other scanners, I use VueScan Professional. Works very well indeed. With the Fujitsu Fi5120, scans A4, single or both sides. Very fast and easy. I use it as B/W, though it can do colour.
Will gobble up whole A4 double sided manual very quickly, and dump the whole pdf file in my computer. Does not usually jam.
The Fujitsu proprietory software was very good when I bought the machine, but it was intended for XP 32 bit. Shows how many years ago I acquired it. It does not work for Win7 or Win10, 64 bit, and no update/revised version is available. Buy a new Fujitsu scanner I am told! Whereas the VueScan works well.

I run both at 300dpi to pdf, which I can then tidy-up later in Acrobat (crop, trim black margins, invert, re-arrange pages order), not the AcrobatReader. B/W or colour as wanted.
By using a proper comb-binding machine, I can get a reasonable professional finish on any manuals I make.

The other scanner is a HP SJ5590. It does single or duplex A4 autofeed, but it is slow. I used to use it for colour. The software supplied is very troublesome. HP software always seems very tender and easily put off. Perhaps I don't buy their latest machine and software update frequently enough. Claimed to work with Win7 but does not seem to.
It does work with VueScan.
I think the VueScan Professional cost about 25.
wme_bill

Terry_VK5TM 8th Jul 2021 3:31 am

Re: Scanning of Manuals & Circuit Diagrams?
 
Quote:

Originally Posted by wireman (Post 1388373)
I have been investigating software to be used for OCR treatment of scans of decades of programmes for a local event.

I'm close to purchasing a license for ABBYY FineReader.


I bought that and can recommend it.

dsergeant 8th Jul 2021 5:53 am

Re: Scanning of Manuals & Circuit Diagrams?
 
I found the OCR option in PDFXchange Editor https://www.tracker-software.com/pro...xchange-editor very good, far better than that built into the Canon software for my DRC225 scanner. Not used it on circuit diagram scans though.

Dave

Julesomega 26th Jul 2021 1:04 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
Quote:

Originally Posted by Julesomega (Post 1388396)
The free paint dot net will save multiple layers as jpg's, and there are many online converters that will combine all the jpg's into one PDF

Oh happy day! When I posted above, I had tried in vain to find the wonderful "JPEGtoPDF" utility which I had been using for years on a laptop to generate PDFs of manuals from typically dozens of PNG or JPG images. All I could find was online converters; I'm always a little sceptical about 'free' online conversion - size limits? 'free trial'? I mean, what's in it for those outfits?

Anyway, It's reappeared with a new name "Image To PDF or XPS" from compulsivecode.com. If I were sensible I'd wait til I'd tried it but I'm too excited to wait til I can get online again

GMB 27th Jul 2021 8:26 am

Re: Scanning of Manuals & Circuit Diagrams?
 
DO NOT USE JPEG unless it is a photo of something in the real world.

AVOID FULL COLOUR as it is very bulky.

Ideally you want text and diagrams processed as 1-bit monochrome. But this may not be simple unless the original documents are nice and clear and black and scanned properly.

Pictures are tricky - you need to see how they were originally printed. They were often silk-screened so may work better than expected in mnochrome (at high resolution), but will look better if down-scaled and converted to grey-scale (or colour if it really is coloured).

Ideally you want to have separated out the different kinds of image, text/circuits and photos, so that these can be composed separately and combined back together to make your PDF. That way you can have a jpeg colour photo embedded in a page of monochrome text - and obtain optimal compression and clarity of the result.

Note that due to the way lossless compression is usually done in PDF you will still get fairly good compression if you have pasted a grey-scale photo into a monochrome page of text.

Maarten 27th Jul 2021 3:57 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
1 bit monochrome is mostly out of the question unless you want the scan either to loose lots of detail or you're willing to do pixel and palet editing by hand to reduce better source material; it's not a goal that should be targeted unless you have extensive experience in editing.

4 bit is achievable without too much trouble and 8 bit grey scale is just fine for mixed content (text and pcb layouts or other simple grey scale illustrations.

PDF compression should be set to PNG, or a PDF should be created from PNG source material.

Terry_VK5TM 27th Jul 2021 11:27 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
Quote:

Originally Posted by Maarten (Post 1393290)
PDF compression should be set to PNG, or a PDF should be created from PNG source material.


Ive also said this many times.


The CQ-DATV mag pdf version that I compile and edit only uses png images.


On average the mag comes out to around 7Mb (being pushed for time once, I used the JPEG images direct and the mag came out at over 22Mb).

GMB 28th Jul 2021 8:32 am

Re: Scanning of Manuals & Circuit Diagrams?
 
Quote:

1 bit monochrome is mostly out of the question unless you want the scan either to loose lots of detail or you're willing to do pixel and palet editing by hand
It depends on your starting material. If you have a nice clear document then if you scan it for monochrome at quite high resolution (minimum 300 dpi) then the result is usually just fine and requires no editing unless your are a perfectionist, and you have lost no useful detail. I do a lot of document reprocessing and find that this usually works just fine.

The point here is that until quite recently there was no way to print pages with proportional shades. If you get the microscope out you will find that on the paper are scattered bits of solid black ink. But even quite old printing was surprisingly high resolution, so if you exceed its resolution in scanning then the result in pure monochrome is as good as the original. What you are doing by using greyscale is blurring the black patterned areas to obtain the grey intent. For photos this is a way to reduce the dpi and increase the bits per sample and the result can look better in a PDF than the original document (because PDFs tend to be viewed on small screens).

The trouble comes with old photocopies of documents that were not flat. Or photos of them, and with electronic photos you have added JPEG fringes which defeat thresholding and ideally need to be blurred but this looses detail. Then the advice to use the smallest number of greyscale bits that works is the way to go. The point here is that the human eye and brain are superbly good at compensating for the errors software will struggle to do the same.

Julesomega 29th Jul 2021 1:42 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
Quote:

Originally Posted by Julesomega (Post 1393007)
If I were sensible I'd wait til I'd tried it but I'm too excited to wait til I can get online again

Forget this new 'image to PDF' converter, it does no such thing. It has been hijacked by NCH and all it will do is convert anything to ZIP. You can't print a ZIP !! Besides, Zipping has been done by your OS for the past few dacades. I have emailed the guy about this and will post if I get a response >((

Julesomega 29th Jul 2021 3:18 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
1 Attachment(s)
No, forget any previous excitement, this is exciting!

Desparate to generate a PDF from a batch of PNG images I tried the laptop which already had JPEGtoPDF installed but it wouldn't work, just tells me it can't cos all the PNGs are in use and unavailable etc. No problem, just this once I'll 'convert online' and looked through the first 500 sites that came up to see if any didn't have tiny size limits, watermarks, require sign up, pay now etc etc and PDF Candy showed up. Went to try it and blow me there's an option to download an offline version, which I did and it really works. Document now printed off in duplicate ready for signing. Does what it says and offers a lot more functions. Next will download on to the desktop PC

Now hurriedly looking for images of ecstasy ;D

Julesomega 29th Jul 2021 3:41 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
1 Attachment(s)
Sorry to get a bit over excited but this is really good! The OP asked for OCR and PDF Candy offers OCR - haven't tried it yet but I will. Screenshot shows the functions

Edit: you have to pay after your first try, 6 + VAT -> 7.20, I'm aiming to get my money's worth from this!

Julesomega 29th Jul 2021 4:57 pm

Re: Scanning of Manuals & Circuit Diagrams?
 
An update: I heard straight back from Jesse Yeager at compulsivecode, assuring me that nothing has changed except the name, so I tried downloading it again. Last time, I double clicked the zip file and somehow NCH installed itself and opened up offering to convert anything to a zipfile but without bothering to unzip anything, so this time I extracted the .exe manually and everything works just as I wanted :)

So back to the old faithful, and it's free!


All times are GMT. The time now is 12:15 am.

Powered by vBulletin®
Copyright ©2000 - 2021, vBulletin Solutions, Inc.
Copyright ©2002 - 2021, Paul Stenning.