I think the software is worth more than the actual physical scanner. There must be software out there that already does this kind of thing, flatten out curved pages and OCR to convert to PDFs and Word. I mean, I can already flatten images taken with my smart phone with Photoshop and there's an abundance of OCR software out there, some with even better accuracy than this. The trouble is when you buy something like this, you're at the mercy of this company's updates and delivery schedule for bug fixes and new features if any. For $300 for a over glorified lamp and 1990 smart phone camera I think I'll take a pass.
As I watched the video, my guess was ABBYY FineReader, judging by the zany list of languages faintly discernible on the OCR dialog. Googled it, and bingo.
I spent twenty years as a service engineer in the microfilm industry and it's things like this that killed the industry stone dead. Scanners will struggle when there are areas of different contrast on the same page but nowhere near as badly as film. It was the constant bane of my life trying to explain to people that I couldn't put contrast where no contrast existed and I couldn't make 35mm film have different sensitivity from one side of a shot to the other. I could cheat and change the exposure by holding black card in front of some lamps and not others but very few people wanted to pay my hourly rate for that. It was shame the whole thing died on its arse and took a lucrative living with it but scanners vs microfilm?.... Scanners win and deserve to win.
That said, I do thank you and other technicians of the microfilm "era". I still use the old machines to this day, and frankly, PREFER older microfilm readers to the digitized controls of today. I've used both, and it is far easier to manipulate older machines than newer digitized ones. For the thousands of microfilm rolls that I've viewed, I found it best to use the camera on my phone to digitize documents I want. Then I can manipulate those at my own leisure instead of being rushed at a library. Thank you again to you and others that serviced the machines.
@@kurtfrancis4621 It was fun while it lasted. I saw the end coming when the first affordable document scanners appeared with an automatic feed and from that point to when I pretty much threw in the towel was less than five years. The actual demise of microfilming documents reached a tipping point and went from a slow decline to a total collapse in probably about six months. The company I was most closely associated with went from a filming turnover of about £70,000 per month to about £5,000 so fast it was unbelievable and it wasn't really the scanners that caused it although they helped. The final nail in the coffin was the ever cheaper and ever larger hard drives giving even the smallest office a "server" and documents that were "born digital" and stayed that way for most purposes.
...and now there are micro-film scanners to archive the IBM size mounted cards digitally. All of the original Hammond organ drawings are on card mounted microfilm in a long wall of metal file cabinets.
But once you buy that, you no longer have a reason not to scan all those things you've thought about scanning. Good luck with all the work you just created for yourself.
I would say this thing is good enough for everything that isn't art or Da Vincis handwriting- especially manuals are much better off to be archived in any quality rather than sitting in a box waiting for the perfect scan
Having tried to read or use some improperly scanned documents (even some of the Google scanning project output) I'm not totally convinced of that, but I understand your meaning. What scares me about bad scans is that the owner of the paper will decide that a bad scan is just perfect and throw the original document away. The Feds did that with thousands of tons of documents back in Bill Clinton's era, and the remaining 200 DPI badly-lit scanned documents are pretty much unreadable. But there are no originals anymore to be rescanned better.
@@lwilton I'll be honestly the feds likely pulled that shit on purpose and these isnt too bad. I think it would be fine for stuff that doesn't matter too much but would nice to have backed up online Well and it could be used a temp scan of stuff intill s much better version can be done. I think a poor scan is better than nothing or losing it forever. Hoenstly this neat little thing has uses long as you're accepting this isnt going best but its going go be quick and easy.
The Feds very much did do that on purpose. It was called the "Paperwork Reduction Act". Some of that meant that you had to fill out fewer forms in triplicate to deal with government agencies, but most of it was aimed at reducing the amount of paper the government was storing.
Much older pre melanium manuals use standardized text. Even hand drawn tech manuals from mid century had to stick to conventional standards. So modern OCR software should not have huge error issues in processing it.. The only real issue I see is how the information is layed out on paper. The page format for diagrams and tables maybe different across different archives due to the internal preference of individual companies. This would require a more human touch to help the software recognized small details. Also older manuals with hand written margin notes or corrections might also be come problematic.
Marc, very cool. Speaking of OCR and scanning, I just acquired a bunch of the PCB's containing the CCD's from the cameras Google used (allegedly) for their massive document scanning project. The PCB's are nice self contained modules with a Xilinx FPGA, all the hardware to drive the CCD, amplifiers, variable gain amplifiers, local power regulators, and FFC connectors for the FPGA pins and power. The CCD is a very high spec & high resolution Kodak unit that has been used in Astronomical CCD cameras. I'm currently implementing a board that will power and interface it to my dev environment. Unfortunately I don't know what happened to the rest of the camera, the optics, the main PSU, and the other logic boards which would have contained the CPU and external interfaces. I think I'll buy one of these scanners for myself for Christmas :-)
@@maicod Somewhere later in this comments section someone said they abandoned the project. I guess Google has got better things to do with their resources since a document scanning project seems like a totally worthwhile thing.
@@vincei4252 Apparently they haven't abandoned it completely, but slowed down, here's an interesting article about it: www.wired.com/2017/04/how-google-book-search-got-lost/
Interesting review, looks like the thing really works! For the ALDs and fold-outs, it may be possible to construct a vacuum table. Say 20mm thick, white laminex surface, drilled with a square pattern of holes like a reverse Air Hockey table in a sense. Have two adjoining edges raised a little to allow quick positioning, then use a small vacuum cleaner to suck it down. The vacuum switch could be as simple as a hole on the side of the table you momentarily place your hand on to effect the vacuum.
Considering where we have come from over the years and these days the ability for easy archiving at a decent price, it is time for people to start digitising for the future. Thanks Marc for sharing your experiences.
Here in South Korea, there are a number of companies of dubious legality that scan books - real, proper, bound books - hardback or soft bound. I had some books scanned by one of these companies, and had a discussion with the owner about how they did it and a brief tour. They physically remove the binding and slice and/or open all the leave for automated sheet feeding scanners. Then they reassemble the book using standard book binding techniques. To the layman you can’t tell the difference, before and after. Such a system wouldn’t work at all on the types of loose-leaf and fragile documents you demonstrated, but for a broad class of books this works. And it was remarkably cheap.
Joe, if you’re located in Korea, I can try to find a couple. I’m not so sure about other countries. As for security, you typically leave your book with them for a couple days. I suppose if it was a super-secret book you could make arrangements to hand-walk the book through the process?
The problem I have found is some old bound books cannot be opened flat. It would be great to have a scanner with two sheets of glass or moulded acrylic at a 30 to 45-degree angle so that the spine of the book is not damaged yet the pages are still pressed flat against the glass. I once operated some Klimsch large-format (broadsheet newspaper) cameras. I'd love to try putting in an object lens and a hi-res CCD.
I can't think of the name right now, but I've seen an open-source project like this on UA-cam years ago. You put the book in a cradle that opened it to maybe 45 degrees and flattened the pages with a hinged piece of acrylic (?) that was formed into a matching shape. Two regular digital cameras were mounted above, one pointing at each page. You'd take the photos, lift the hinged acrylic v-shape, turn the page, lower it, take the next set of photos, repeat.
Search UA-cam for 'Easy Book Scanner - an Introduction to this 1000 pages per hour scanner' and 'Using my newly built DIY Book Scanner!' for the general concept.
Q. Let's say your manual has images of PCBs. Could you also use the CZUR to take hi-res colour pictures of the actual PBCs to be pasted into the scans?
Love that idea. A lot of hand drawn circuit traces from mid century PCB's through late 70's look more like blob-ish works of modern art. Tasking software to automatically reconstruct circuit traces from an image might cause serious issues.
Sidelights help with the contrast, especially on glossy paper. What also helps is using the fingerpads that the scanner software automatically edits out to flatten the book pages and even out the bright spots on the curved part of the paper.
Enjoyed the video. Suggestion for papers that do not fit in a single scan: there's a software called Hugin that does automatic stitching given two or more images. That might work, but it's an additional step.
I'm always wary of OCR-ing technical documents, because short of meticulously proofreading the whole thing, you'd never be 100% sure if a wrong symbol or number creeps in somewhere. What I would really like such a system to do, is to create a document format that keeps the original image, and puts the OCR text invisibly in the background. So if there is an error, you may not be able to find the term with a search, but someone reading the document still has the original image.
ABBYY Finereader will do that for you as a two layer PDF, image on top with OCR'ed searchable text behind. OCR software just received with my CZUR ET18 Pro will also do that (April 2020).
My college physics professors used to hound me endlessly about writing in my lab book with pencil... now I think I know why lol. All jokes aside, thank you for such a thorough demo of this scanner & software! Very detailed, with enough use cases to really let me know whether or not this product is worth my time (sounds like it is!!). Amazing!
It's actually extremely simple. Just need a large suction cup, a small vaccum pump and an arm which just moved by a wheel in the back. You split the arm and add a hinch plus a spring to adapt for different heights 😙🎶
@@TidusleFlemard well, the suction cup would just need to build enough force to lift the page and over a bigger area than some fingers. So everything able to move by hand should be safe to turn by such a simple apparatus. But sure, you don't want to lift old pages at a high speed.
Some people don't appreciate how much work it took someone to do this task and make it available to the masses via Internet. I find it somewhat Mind Boggling how much information has made this trip in the little time we have had the Internet available. It would seem you have a rather good handle on this thing. I'm IMPRESSED with You.
this would have been perfect when I worked at our uni's history department. it's a real pain scanning and aligning/rotating pages to make a good pdf afterwards. Even for taking pictures of small items in a collection I think it could be very useful.
Thank you. I've been getting adds for this thing on Facebook for weeks. Normally I ignore Facebook adds, however, I actually have a use for this. Old lab notebooks, including a lot of rapidly fading printouts on thermal paper in very awkward formats not fit for a regular scanner. If I cannot find an actual book scanner to rent, this might be the next best thing.
wow...this scanner is rally amazing...seems to me that it is in fact a high megapixel digital camera more than a scanner. It is very helpful to have such a device.
It looks to not even be that high megapixel. 14 megapixel CMOS sensor, 4320*3240 pixels. each pixels has RGBG subpixels, using Bayer filter. This translates to about 240 DPI. Nothing crazy, but not unusable, and definitively possible to use. If it would have just a monochromatic sensor, maybe 25Mpx, it would really improve the quality a lot (~600DPI) for monochromatic stuff.
Overall, a very nice and informative video watch and I'm really impressed by the results of this scanner. But it would nice have a comparison with a more professional high-end scanner use in museums. Keep up the great work!
That is actually pretty good, the OCR to Word / OpenOffice, worked really well. It recognized vertical text, formating, weird stuff, tables and diagrams as images. This is essentially perfect. I wonder which OCR engine they use. By quality of the OCR I would guess they are licensing some libraries from Abbby FineReader to do it.
It would be nice if it had a feature to concatenate pages. This would solve the issue you had with the larger schematics. If the scanner doesn't have it a secondary program could be written using OpenCV to concatenate the scans into one image.
pretty cool hardware software combo, now, for books an auto page turn would be an awesome addition in future. You can probably think of some sort of arduino project to do the job and communicate with the software + auto save to image and .pdf + ocr !
At 4:38 you say there are "no errors" in the OCR, but I see two (OK, technically the same error twice): the source document shows "1" (numeral one) but the OCR text shows "I" (capital letter i). Granted, that is a difficult OCR case, but to say "no errors" is not true.
Just ordered one... dont ever think of scanning books with a scanner or take images and crop them manually.. been there done that...its a pain... this will make my life so much more fun - and digital :) Thanks for the review
Scanning them is way, way more tolerable with a good scanner. I wonder what the resolution of this device is though. It doesn't really look like it can hold a candle to my (proper) book scanner. But it does look so, so much more convenient.
@@VJFranzK I have used the scanner and software while disconnected and it performs exactly the same. It does not need to use servers in China for its operation.
Great review! Really like the product. Regarding the schematic that's too large to scan in one shot: can you stitch the images together? Better can you get the scanner software to stitch the images together? I have Heathkit schematics that I want to scan!
Altough a bit slow, you can use hp smart to scan the documents from a phone. It really does an amazing job. It can even flatten images that are taken from an angle
Did the flash ever happen when you scanned? I never saw the side lights come on in scan mode in the video for the one washed out hand-written book you had. That also doesn't seem as if it is the highest DPI a scanner like that should be able to achieve.
Seemed to work for you. Am looking to scan some old magazines before I recycle them, would you recommend this product? Archive and store locally to HDD and or CD / DVD. Thanks
you mentioned that this was a Indego transaction. They now have a CZUR Shine Ultra in the "shipping stage". Is it safe to purchase this way? Thanks in advance for any help
CuriousMarc Did you try scanning any photos? I'm trying to decide between one of these and the usual flatbed. This is great for 60-70 of my materials but I still have lots of photos and want to find out how well a CZUR would handle them.
Imagine a document scanning phone app that was able to utilise; 1) Apple FaceID depth map technology to establish a better surface topology of the scanned paper for later computational flattening 2) dynamic pixel shift style resolution enhancement - increased resolution scan through natural camera shake 3) the ability to dynamically stitch imagery on the fly for larger document captures - I saw a limited version of this once in a hand held scanner, you just waved it round the document until you got it all!
You just need a sheet of glass or plastic to flatten book. As long as light is in the right spot not to cause reflection. Or a plastic frame to border book and flatten. Two ruller's top and bottom on a hinge like a gilotine could work too.
The OCR software is ABBYY Finereader integrated engine, and Irish is one of the 193 languages they list for document recognition. I went to the ABBYY site rather than Czur's.
CURIOUSMARC, I'm looking for a book scanner that has an automatic scanning where I can just flip the pages in real time and it will snap the pictures scanning the pages without manually clicking a mouse button or a footswitch to snap is scan, Do you know any book scanner that does this good quality? Most of my College Books are glossy pages very glossy prints and are 300pages to 500 pages. How can I book scan these books really fast without manually clicking a mouse button or using a footswitch. I just want to quickly flip the pages and the bookscanner will snap the pictures scans in real time in very good quality. Do you know what I mean and can this be done? What kind of laser lights does the CZUR bookscanner use compared to other Book Scanners?
You can’t press a button with your feet ? It is much easier than clicking,that is why they made the foot switch. Then you have your hands free to flip the pages. And to hold the pages in position they have included some extra clippers that function as page holders.
Hey Marc, I know this is an old video but if you ever need help with any OCR stuff let me know! I think we could get these pages looking much cleaner with some of my code.
If there was extra lights the edges probably wouldn't be dark. When I do vids I have 2 lights opposite each other at a flat angle to the object so there is no glare reflection. The Word one was good, can it convert to PDF.
I don't understand... In the end it's nothing more than a smartphone on a holder... Cause I can do exactly the same with my phone and even with an old one which would cost only 50 bucks
The problem with scanners is their "optimization" or "auto-lightning" features. Instead of including high lumens lamps with edge scanning sensors and AI position learning, they go with tech that is over a decade old...
I saw your video in my recommended 5 minutes after you uploaded. So I thought, "Hey!, no one is sleeping. Let's ask the questions.". So I will be trying to make a computer, an elementary one,, using RTL after my exams are over. Meanwhile I was trying to figure out the working of the circuits given in IBM 1401 SMS book. I liked the symbols they have used for transistors. But I cannot figure out some blocks, which are having the same symbol as the transistors but are having something written on them. Here is an image of the same: imgur.com/a/dScrCrz What does this T1-N-101 means? Is it the part no. of some transistor? I think that that is not the case for they are having labels from T1 to T6. So what do these blocks do? If they are transistors, why aren't the labelled simply npn or pnp like in some other circuit diagrams? Also in diagrams of some SMS card, voltages were marked +6M and 6 , and +12M and -12. What do these mean? What is the reference point? imgur.com/a/ldAD29T Are there any manuals online where I can learn about these notations? Sorry for being weird if I was at any part of this long comment, but given that you and your team posses such an encyclopedic knowledge of old computers, I thought that maybe you can help me out :-) I liked this video. It is different and I can relate to it. Because the toughest part of sharing my lecture notes with my friends is scanning them. I use an app on Play Store: Notebloc app. It's clumsy but does the job.
Scanning old records should be a job for robots, or at least unpaid interns. Or, get an unpaid intern to cobble together a scanning robot. I recall the 1960 census data being stored on Univax media, and thus inaccessible shortly afterwards. I think someone figured it out.? Have fun.
So uhh..... I'm not being a twat here, I really honestly mean this.... I use Adobe Mobile Doc Scanner, an Application for Android. And it literally does more than this does in software, while also, quite literally, producing better quality "scans" as well. The only drawback is that it's up to you to hold the phone in the best placement as to not cast shadows, etc... though, the software does do a VERY good job at mitigating even the worst of those when merely in B/W document mode. This thing is neat and all, but truly... give Adobe MDF a shot, it's free.... and you'll get better results, with a richer post processing suite at your disposal. Once more, for free... free.
"The only drawback is that it's up to you to hold the phone in the best placement as to not cast shadows, etc." Do you not see why this would be worth $300 when scanning hundreds or thousands of pages?
@@jackoreilly3479 Well, no... I'd just make a jig for the phone... I mean, that's what I'd do. Either out of metal, wood, 3D print, etc. Given that the results are far and away better with the phone than this, it saves a ton of money, etc... I believe I'd avoid the expensive, worse performing option more or less overall.
I think this may have been the reason the Xerox 813 did not have a platen glass and the Xerox 2400 had a curved glass, to keep people from copying (or scanning) copy-written material from books, haha.
It's too much for a glorified camera mounted on a lamp. Most current android phones with free scanning software can do better job with better post processing. Only thing the lamp does better is OCR export to word.
I was actually thinking of buying the newer model ever since I saw it being advertised on Instagram, but I think that might be possible with an old phone I have lying around, if I can plug in a mouse and video out, and mount it on a stand, but then again, I would only really need it for a research project and nothing much else.
@@kbhasi Definitely save the money. If it's an android phone you can just plug in a mouse and set up "camera uploads" function via Dropbox or Google Drive, it will then upload every image from the phone to cloud and you can access it instantly on the PC. On android there are also various free apps for using camera as a scanner with flattening effect and correcting perspective.
10:38 That's genuinely impressive how the OCR engine handled the mixed text, tables and vertical text.
actually the scanning was a tad underwhelming (the page with the light right side) compared to that OCR-engine
I think the software is worth more than the actual physical scanner. There must be software out there that already does this kind of thing, flatten out curved pages and OCR to convert to PDFs and Word. I mean, I can already flatten images taken with my smart phone with Photoshop and there's an abundance of OCR software out there, some with even better accuracy than this. The trouble is when you buy something like this, you're at the mercy of this company's updates and delivery schedule for bug fixes and new features if any. For $300 for a over glorified lamp and 1990 smart phone camera I think I'll take a pass.
As I watched the video, my guess was ABBYY FineReader, judging by the zany list of languages faintly discernible on the OCR dialog. Googled it, and bingo.
Uuh, your profile picture and name do not agree
@@rogervanbommel1086 Hey, at least it's a real cardiac cycle.
All it needs now is Blade Runner voice control and sound effects. "Enhance... enhance.... enhance... give me a hard copy right there."
"Yuck! Hard copy."
I spent twenty years as a service engineer in the microfilm industry and it's things like this that killed the industry stone dead. Scanners will struggle when there are areas of different contrast on the same page but nowhere near as badly as film.
It was the constant bane of my life trying to explain to people that I couldn't put contrast where no contrast existed and I couldn't make 35mm film have different sensitivity from one side of a shot to the other.
I could cheat and change the exposure by holding black card in front of some lamps and not others but very few people wanted to pay my hourly rate for that. It was shame the whole thing died on its arse and took a lucrative living with it but scanners vs microfilm?.... Scanners win and deserve to win.
thank you for sharing that¡
That said, I do thank you and other technicians of the microfilm "era". I still use the old machines to this day, and frankly, PREFER older microfilm readers to the digitized controls of today. I've used both, and it is far easier to manipulate older machines than newer digitized ones. For the thousands of microfilm rolls that I've viewed, I found it best to use the camera on my phone to digitize documents I want. Then I can manipulate those at my own leisure instead of being rushed at a library.
Thank you again to you and others that serviced the machines.
@@kurtfrancis4621 It was fun while it lasted. I saw the end coming when the first affordable document scanners appeared with an automatic feed and from that point to when I pretty much threw in the towel was less than five years. The actual demise of microfilming documents reached a tipping point and went from a slow decline to a total collapse in probably about six months. The company I was most closely associated with went from a filming turnover of about £70,000 per month to about £5,000 so fast it was unbelievable and it wasn't really the scanners that caused it although they helped. The final nail in the coffin was the ever cheaper and ever larger hard drives giving even the smallest office a "server" and documents that were "born digital" and stayed that way for most purposes.
...and now there are micro-film scanners to archive the IBM size mounted cards digitally. All of the original Hammond organ drawings are on card mounted microfilm in a long wall of metal file cabinets.
But once you buy that, you no longer have a reason not to scan all those things you've thought about scanning. Good luck with all the work you just created for yourself.
Add to that the film scanner I just got at Christmas....
I would say this thing is good enough for everything that isn't art or Da Vincis handwriting- especially manuals are much better off to be archived in any quality rather than sitting in a box waiting for the perfect scan
Having tried to read or use some improperly scanned documents (even some of the Google scanning project output) I'm not totally convinced of that, but I understand your meaning. What scares me about bad scans is that the owner of the paper will decide that a bad scan is just perfect and throw the original document away. The Feds did that with thousands of tons of documents back in Bill Clinton's era, and the remaining 200 DPI badly-lit scanned documents are pretty much unreadable. But there are no originals anymore to be rescanned better.
@@lwilton I'll be honestly the feds likely pulled that shit on purpose and these isnt too bad. I think it would be fine for stuff that doesn't matter too much but would nice to have backed up online
Well and it could be used a temp scan of stuff intill s much better version can be done. I think a poor scan is better than nothing or losing it forever. Hoenstly this neat little thing has uses long as you're accepting this isnt going best but its going go be quick and easy.
The Feds very much did do that on purpose. It was called the "Paperwork Reduction Act". Some of that meant that you had to fill out fewer forms in triplicate to deal with government agencies, but most of it was aimed at reducing the amount of paper the government was storing.
Much older pre melanium manuals use standardized text. Even hand drawn tech manuals from mid century had to stick to conventional standards. So modern OCR software should not have huge error issues in processing it.. The only real issue I see is how the information is layed out on paper. The page format for diagrams and tables maybe different across different archives due to the internal preference of individual companies. This would require a more human touch to help the software recognized small details. Also older manuals with hand written margin notes or corrections might also be come problematic.
I'm actually impressed with how well that OCR engine handled the text and diagrams as usually they choke a bit on mixed content.
Marc, very cool. Speaking of OCR and scanning, I just acquired a bunch of the PCB's containing the CCD's from the cameras Google used (allegedly) for their massive document scanning project. The PCB's are nice self contained modules with a Xilinx FPGA, all the hardware to drive the CCD, amplifiers, variable gain amplifiers, local power regulators, and FFC connectors for the FPGA pins and power. The CCD is a very high spec & high resolution Kodak unit that has been used in Astronomical CCD cameras. I'm currently implementing a board that will power and interface it to my dev environment. Unfortunately I don't know what happened to the rest of the camera, the optics, the main PSU, and the other logic boards which would have contained the CPU and external interfaces.
I think I'll buy one of these scanners for myself for Christmas :-)
is that google scanning project still ongoing ?
@@maicod Probably, maybe the hardware was upgraded at some point - this hardware seems to be circa the early 2000's.
@@maicod Somewhere later in this comments section someone said they abandoned the project. I guess Google has got better things to do with their resources since a document scanning project seems like a totally worthwhile thing.
thanks Vince I, I guess Marc's team is acquired to take over from Google now :)
@@vincei4252 Apparently they haven't abandoned it completely, but slowed down, here's an interesting article about it: www.wired.com/2017/04/how-google-book-search-got-lost/
Interesting review, looks like the thing really works! For the ALDs and fold-outs, it may be possible to construct a vacuum table. Say 20mm thick, white laminex surface, drilled with a square pattern of holes like a reverse Air Hockey table in a sense. Have two adjoining edges raised a little to allow quick positioning, then use a small vacuum cleaner to suck it down. The vacuum switch could be as simple as a hole on the side of the table you momentarily place your hand on to effect the vacuum.
Pegboard, a box and a vacuum cleaner ;)
Considering where we have come from over the years and these days the ability for easy archiving at a decent price, it is time for people to start digitising for the future. Thanks Marc for sharing your experiences.
Here in South Korea, there are a number of companies of dubious legality that scan books - real, proper, bound books - hardback or soft bound.
I had some books scanned by one of these companies, and had a discussion with the owner about how they did it and a brief tour. They physically remove the binding and slice and/or open all the leave for automated sheet feeding scanners. Then they reassemble the book using standard book binding techniques. To the layman you can’t tell the difference, before and after.
Such a system wouldn’t work at all on the types of loose-leaf and fragile documents you demonstrated, but for a broad class of books this works. And it was remarkably cheap.
Chris Lott, how secure would the process be and where would I look for a reliable company that does this?
Joe, if you’re located in Korea, I can try to find a couple. I’m not so sure about other countries. As for security, you typically leave your book with them for a couple days. I suppose if it was a super-secret book you could make arrangements to hand-walk the book through the process?
The problem I have found is some old bound books cannot be opened flat. It would be great to have a scanner with two sheets of glass or moulded acrylic at a 30 to 45-degree angle so that the spine of the book is not damaged yet the pages are still pressed flat against the glass.
I once operated some Klimsch large-format (broadsheet newspaper) cameras. I'd love to try putting in an object lens and a hi-res CCD.
Please check out this open-source machine diybookscanner.org/en/intro.html
I can't think of the name right now, but I've seen an open-source project like this on UA-cam years ago. You put the book in a cradle that opened it to maybe 45 degrees and flattened the pages with a hinged piece of acrylic (?) that was formed into a matching shape. Two regular digital cameras were mounted above, one pointing at each page. You'd take the photos, lift the hinged acrylic v-shape, turn the page, lower it, take the next set of photos, repeat.
Search UA-cam for 'Easy Book Scanner - an Introduction to this 1000 pages per hour scanner' and 'Using my newly built DIY Book Scanner!' for the general concept.
I remember there already is a scanner which can compensate that, but I cannot remember its name...
something like the "ScanRobot"? ua-cam.com/video/SdipuAuWsEs/v-deo.html
Q. Let's say your manual has images of PCBs. Could you also use the CZUR to take hi-res colour pictures of the actual PBCs to be pasted into the scans?
Love that idea. A lot of hand drawn circuit traces from mid century PCB's through late 70's look more like blob-ish works of modern art. Tasking software to automatically reconstruct circuit traces from an image might cause serious issues.
Sidelights help with the contrast, especially on glossy paper. What also helps is using the fingerpads that the scanner software automatically edits out to flatten the book pages and even out the bright spots on the curved part of the paper.
Enjoyed the video.
Suggestion for papers that do not fit in a single scan: there's a software called Hugin that does automatic stitching given two or more images. That might work, but it's an additional step.
Did you say Bell Labs?
Very interesting, did some scanning & ocr work between 1999-2004 with massive Minolta book scanner and ordinary flatbed scanners.
I'm always wary of OCR-ing technical documents, because short of meticulously proofreading the whole thing, you'd never be 100% sure if a wrong symbol or number creeps in somewhere. What I would really like such a system to do, is to create a document format that keeps the original image, and puts the OCR text invisibly in the background. So if there is an error, you may not be able to find the term with a search, but someone reading the document still has the original image.
ABBYY Finereader will do that for you as a two layer PDF, image on top with OCR'ed searchable text behind. OCR software just received with my CZUR ET18 Pro will also do that (April 2020).
My college physics professors used to hound me endlessly about writing in my lab book with pencil... now I think I know why lol.
All jokes aside, thank you for such a thorough demo of this scanner & software! Very detailed, with enough use cases to really let me know whether or not this product is worth my time (sounds like it is!!). Amazing!
Thanks for this review. Could you please try to scan a book with photos?
7:40 Nice bit of Verilog spotted there, and an old-skool Xilinx FPGA. ;)
Keen eye! That’s what it is.
Thanks for the video review! I'm interested in grabbing their new scanner model coming out next year.
@curiousmarc's next project: a page-turning robot ;-)
It's actually extremely simple. Just need a large suction cup, a small vaccum pump and an arm which just moved by a wheel in the back. You split the arm and add a hinch plus a spring to adapt for different heights 😙🎶
@@RubenKelevra wouldn't a suction cup possibly damage documents in already bad shapes tho?
@@TidusleFlemard well, the suction cup would just need to build enough force to lift the page and over a bigger area than some fingers. So everything able to move by hand should be safe to turn by such a simple apparatus.
But sure, you don't want to lift old pages at a high speed.
Some people don't appreciate how much work it took someone to do this task and make it available to the masses via Internet. I find it somewhat Mind Boggling how much information has made this trip in the little time we have had the Internet available. It would seem you have a rather good handle on this thing. I'm IMPRESSED with You.
@ungratefulmetalpansy likely a auto correct thing
@ungratefulmetalpansy Just to Emphasize . I don't even remember doing it. LOL
Very interesting! I always wondered about a better way for scanning old manuals etc for various subjects. love these videos
this would have been perfect when I worked at our uni's history department. it's a real pain scanning and aligning/rotating pages to make a good pdf afterwards. Even for taking pictures of small items in a collection I think it could be very useful.
The OCR demonstration was pretty cool!
Thank you.
I've been getting adds for this thing on Facebook for weeks. Normally I ignore Facebook adds, however, I actually have a use for this. Old lab notebooks, including a lot of rapidly fading printouts on thermal paper in very awkward formats not fit for a regular scanner.
If I cannot find an actual book scanner to rent, this might be the next best thing.
I like the idea of some kind of air device to flatten documents
Sounds like something that can be done computationally, too. I believe the Hubble scope has to correct similar flaws.
I'm impressed... That thing works really quite well
I like the idea of the footpedal. I’m severely visually impaired and I want a way of getting my old book collection into electronic form.
Wow!! The OCR was incredibly impressive! I've dealt with bad OCR software before, and it's a nightmare. This really looked amazing!
wow...this scanner is rally amazing...seems to me that it is in fact a high megapixel digital camera more than a scanner.
It is very helpful to have such a device.
It looks to not even be that high megapixel. 14 megapixel CMOS sensor, 4320*3240 pixels. each pixels has RGBG subpixels, using Bayer filter. This translates to about 240 DPI. Nothing crazy, but not unusable, and definitively possible to use. If it would have just a monochromatic sensor, maybe 25Mpx, it would really improve the quality a lot (~600DPI) for monochromatic stuff.
This video reminds me of the time I dreamt about working with you, and you'd have your physical books, while I'd be using a paper tablet.
Overall, a very nice and informative video watch and I'm really impressed by the results of this scanner. But it would nice have a comparison with a more professional high-end scanner use in museums. Keep up the great work!
That's my Christmas present sorted
That is actually pretty good, the OCR to Word / OpenOffice, worked really well. It recognized vertical text, formating, weird stuff, tables and diagrams as images. This is essentially perfect. I wonder which OCR engine they use. By quality of the OCR I would guess they are licensing some libraries from Abbby FineReader to do it.
It would be nice if it had a feature to concatenate pages. This would solve the issue you had with the larger schematics. If the scanner doesn't have it a secondary program could be written using OpenCV to concatenate the scans into one image.
pretty cool hardware software combo, now, for books an auto page turn would be an awesome addition in future. You can probably think of some sort of arduino project to do the job and communicate with the software + auto save to image and .pdf + ocr !
That is alarmingly clever!
At 4:38 you say there are "no errors" in the OCR, but I see two (OK, technically the same error twice): the source document shows "1" (numeral one) but the OCR text shows "I" (capital letter i). Granted, that is a difficult OCR case, but to say "no errors" is not true.
Dang. Busted.
Just ordered one... dont ever think of scanning books with a scanner or take images and crop them manually.. been there done that...its a pain... this will make my life so much more fun - and digital :)
Thanks for the review
Scanning them is way, way more tolerable with a good scanner. I wonder what the resolution of this device is though. It doesn't really look like it can hold a candle to my (proper) book scanner. But it does look so, so much more convenient.
I read a security concern - the data is sent to their server in China for image processing?
The data is processed on a PC or Mac using software they provide, not remotely.
@@carlclaunch793 It was mentioned in a Facebook discussion.
@@VJFranzK I have used the scanner and software while disconnected and it performs exactly the same. It does not need to use servers in China for its operation.
Some cheap 6axis robot arm with vacuum suction cup at the end for flipping pages?
And some scripting in linux to make it fully automated?
There are commercially available machines doing exactly this :D
Is that the camera angle or does that thing not correct rotation/skewing automatically?
Great review! Really like the product. Regarding the schematic that's too large to scan in one shot: can you stitch the images together? Better can you get the scanner software to stitch the images together? I have Heathkit schematics that I want to scan!
Altough a bit slow, you can use hp smart to scan the documents from a phone. It really does an amazing job. It can even flatten images that are taken from an angle
Did the flash ever happen when you scanned? I never saw the side lights come on in scan mode in the video for the one washed out hand-written book you had. That also doesn't seem as if it is the highest DPI a scanner like that should be able to achieve.
I would love to work at that archive. Scanning documents all day.
how does this compare the scansnap sv-600?
Fujitsu scansnap has something similar tho last time I checked it cost around $1500. For loose pages the ix500 is a beast for around $450
I have the ix500, it's awesome. Soo much better than my HP AIO that can't scan more than 2 pages before the ADF boogers up
Seemed to work for you.
Am looking to scan some old magazines before I recycle them, would you recommend this product?
Archive and store locally to HDD and or CD / DVD.
Thanks
you mentioned that this was a Indego transaction. They now have a CZUR Shine Ultra in the "shipping stage". Is it safe to purchase this way? Thanks in advance for any help
CuriousMarc Did you try scanning any photos? I'm trying to decide between one of these and the usual flatbed. This is great for 60-70 of my materials but I still have lots of photos and want to find out how well a CZUR would handle them.
I did try. It's not good for photos. I would not recommend using it for that.
The lab book (and other lower-contrast sources) might also benefit from the side lights.
Yes, because the pencil marks are also reflective.
Imagine a document scanning phone app that was able to utilise;
1) Apple FaceID depth map technology to establish a better surface topology of the scanned paper for later computational flattening
2) dynamic pixel shift style resolution enhancement - increased resolution scan through natural camera shake
3) the ability to dynamically stitch imagery on the fly for larger document captures - I saw a limited version of this once in a hand held scanner, you just waved it round the document until you got it all!
You just need a sheet of glass or plastic to flatten book. As long as light is in the right spot not to cause reflection. Or a plastic frame to border book and flatten. Two ruller's top and bottom on a hinge like a gilotine could work too.
NICE !!! I was thinking to get a plate of glass and it should look flat.
if i lived closer i would of volunteered
Wow! A lot of value for the money!
I wonder what ocr engine it uses. Tesseract?
Does if produce as an original pdf, capable of selecting letters/ words ?
Yes, it has the option to make a pdf with OCR character recognition.
What website are you making these documents available on? Number 5 needs input!!!
Scans at the Computer History Museum usually end up posted at bitsavers.org
Where can I find a list of the languages it can handle? I need it to be able to OCR Irish-language texts.
The OCR software is ABBYY Finereader integrated engine, and Irish is one of the 193 languages they list for document recognition. I went to the ABBYY site rather than Czur's.
Maybe try using the side lights for pencil as well
CURIOUSMARC, I'm looking for a book scanner that has an automatic scanning where I can just flip the pages in real time and it will snap the pictures scanning the pages without manually clicking a mouse button or a footswitch to snap is scan, Do you know any book scanner that does this good quality?
Most of my College Books are glossy pages very glossy prints and are 300pages to 500 pages. How can I book scan these books really fast without manually clicking a mouse button or using a footswitch. I just want to quickly flip the pages and the bookscanner will snap the pictures scans in real time in very good quality. Do you know what I mean and can this be done?
What kind of laser lights does the CZUR bookscanner use compared to other Book Scanners?
You can’t press a button with your feet ? It is much easier than clicking,that is why they made the foot switch. Then you have your hands free to flip the pages. And to hold the pages in position they have included some extra clippers that function as page holders.
All that scanning sounds like a great job for a college intern. ;^}
Hey Marc, I know this is an old video but if you ever need help with any OCR stuff let me know! I think we could get these pages looking much cleaner with some of my code.
If there was extra lights the edges probably wouldn't be dark. When I do vids I have 2 lights opposite each other at a flat angle to the object so there is no glare reflection. The Word one was good, can it convert to PDF.
How much is this device please?
Quality review
I was so expecting you to say "And when you're done, you just go *boop* and return it to the store." instead of "*boop* and close it" XD
Old version of Word? I'm still using Word 2003 =) Love youre channel very much. Greetings from Germany.
Me to Word 2003. If it works, use it.
Is the software Windows only, or are there Mac OS and Linux versions?
WOW! Those Archives! The smell of vintage paper must be amazing!
Tried to scan a book once, I got 10 pages in and got so bored I gave up
I guess this is about as good as you can get with a camera vs. an actual scanner
Why is there no Android software so you can also work with smartphone + portable book scanner on the go? And then you need no laptop any more.
There is in the newer version (see the Pro18 review)
Quite smart, especially for this price.
One comment, you may want to upgrade the Windows on your laptop :)
You mean, downgrade to Windows 10 ;-)
@@CuriousMarc haha, there's the point :)
I think It would be much better to put a glass plate on the book to flatten the curves.
I don't understand... In the end it's nothing more than a smartphone on a holder... Cause I can do exactly the same with my phone and even with an old one which would cost only 50 bucks
No need for glass to straighten paper - you can use your fingers to strech a paper a bit...
Unfortunately, apple products don't come with so many things included
The problem with scanners is their "optimization" or "auto-lightning" features. Instead of including high lumens lamps with edge scanning sensors and AI position learning, they go with tech that is over a decade old...
No Linux version. :(
I saw your video in my recommended 5 minutes after you uploaded. So I thought, "Hey!, no one is sleeping. Let's ask the questions.". So I will be trying to make a computer, an elementary one,, using RTL after my exams are over. Meanwhile I was trying to figure out the working of the circuits given in IBM 1401 SMS book. I liked the symbols they have used for transistors. But I cannot figure out some blocks, which are having the same symbol as the transistors but are having something written on them. Here is an image of the same: imgur.com/a/dScrCrz
What does this T1-N-101 means? Is it the part no. of some transistor? I think that that is not the case for they are having labels from T1 to T6. So what do these blocks do? If they are transistors, why aren't the labelled simply npn or pnp like in some other circuit diagrams?
Also in diagrams of some SMS card, voltages were marked +6M and 6 , and +12M and -12. What do these mean? What is the reference point? imgur.com/a/ldAD29T
Are there any manuals online where I can learn about these notations?
Sorry for being weird if I was at any part of this long comment, but given that you and your team posses such an encyclopedic knowledge of old computers, I thought that maybe you can help me out
:-)
I liked this video. It is different and I can relate to it. Because the toughest part of sharing my lecture notes with my friends is scanning them. I use an app on Play Store: Notebloc app. It's clumsy but does the job.
Scanning old records should be a job for robots, or at least unpaid interns. Or, get an unpaid intern to cobble together a scanning robot. I recall the 1960 census data being stored on Univax media, and thus inaccessible shortly afterwards. I think someone figured it out.? Have fun.
CZUR = seizure?
as for $300 indiegogo product it's really impressive.
Pretty sure it's pronounced like Caesar.
So uhh..... I'm not being a twat here, I really honestly mean this.... I use Adobe Mobile Doc Scanner, an Application for Android. And it literally does more than this does in software, while also, quite literally, producing better quality "scans" as well. The only drawback is that it's up to you to hold the phone in the best placement as to not cast shadows, etc... though, the software does do a VERY good job at mitigating even the worst of those when merely in B/W document mode. This thing is neat and all, but truly... give Adobe MDF a shot, it's free.... and you'll get better results, with a richer post processing suite at your disposal. Once more, for free... free.
"The only drawback is that it's up to you to hold the phone in the best placement as to not cast shadows, etc." Do you not see why this would be worth $300 when scanning hundreds or thousands of pages?
@@jackoreilly3479 Well, no... I'd just make a jig for the phone... I mean, that's what I'd do. Either out of metal, wood, 3D print, etc. Given that the results are far and away better with the phone than this, it saves a ton of money, etc... I believe I'd avoid the expensive, worse performing option more or less overall.
It looks like CZUR would pronounce SEIZURE
It's packed like an Iphone! :)
I think this may have been the reason the Xerox 813 did not have a platen glass and the Xerox 2400 had a curved glass, to keep people from copying (or scanning) copy-written material from books, haha.
I it scans better into a Word document than anything else.
It's too much for a glorified camera mounted on a lamp. Most current android phones with free scanning software can do better job with better post processing. Only thing the lamp does better is OCR export to word.
I was actually thinking of buying the newer model ever since I saw it being advertised on Instagram, but I think that might be possible with an old phone I have lying around, if I can plug in a mouse and video out, and mount it on a stand, but then again, I would only really need it for a research project and nothing much else.
@@kbhasi Definitely save the money. If it's an android phone you can just plug in a mouse and set up "camera uploads" function via Dropbox or Google Drive, it will then upload every image from the phone to cloud and you can access it instantly on the PC. On android there are also various free apps for using camera as a scanner with flattening effect and correcting perspective.
Scanner Pro on iOS does a pretty good job too even free hand, but for an optimal result you would need a gig for the phone and proper lighting
upload it all to archive.org :)
Xcanex is automatic though maybe worse picture quality
ScanSnap SV600 is much better the quality of Azur picture in depth was crap when i used it the last time
Intereasting
Amazing. Can you scan something for me? Seriously
Honestly not that amazing.
hingeslevers its all relative
i've seen countless "museum archival" scans way worse than this. i am impressed with the price to quality ratio.
I guess it is pronounced like Caesar.