The document digitisation looks excellent. The live video less so. Big Clive's set-up with a relatively inexpensive smartphone and LED lights clamped to a bookshelf seems to be noticeably better lit, with less harsh contrast, higher resolution and better clarity in general. He seems to have just as good a control over focus, brightness, zoom, framing etc. You may want to consider using the paper in landscape mode when recording.
The processing is terribly offputting, I'd say it outweighs the positives. There's a lot of encoding 'dancing' when its interpolating an already oversharpened image. The old scottish Clive has the best balance. I really don't need a perfect top down over a more pleasing rendering... Cheers!
I have one of this scanners. The hardware seems to be quite good, but the software leafs room for improvement. The trick with the searchable PDFs is, to keep the picture and use the OCR detected text only for the search. As far as I remember this option isn't the default. The page flattening function is very usable in most cases, unless the page is really glossy. My biggest issue with this feature is, that it only works if you scan both pages of a book at once. I have a lot of stuff, that has some pages, where you can fold out a drawing, which makes each page double wide. I wasn't able to scan these properly as a single page yet. The automatic finger removal is also soso. Sometimes it works, sometimes not at all and sometimes it creates strange artifacts. I avoid using it, when I can. When there is text near the edge of the page, it is imossible to place the plastic tingy in a way, that it gets detected.
Thanks! Now I'd like to see what books you have on the shelf about electronics. ;-) Perhaps also good idea to post hi-res photos of device guts on your site. Would be great when you do reviews, especially for newer and more expensive devices, often there is lack of information about them online, and sometimes it's irreplaceable for somebody who is just looking for visual photo, for example to check what is missing/cabling.
The contrast is high on the video section, but it seems like there's pretty heavy sharpness adjustment on it too, if you can get the edges to be less harsh and the background a little less bright, it will probably be more comfortable to look at. I'd second (third, however many) the want for a landscape orientation. Not entirely necessary for the size of the text you're using, but works better non-fullscreen viewing and for having the subject of the video actually taking up most of the frame. Even sort of background/filler would be preferable in my mind to just the black mat.
@@Thesignalpath Have to agree the live contrast edge sharpening is really off putting. All of the live video through the thing is terrible. In the the LT doc it washed out out the LT logo it turned white in the center. But most of the pages look pretty good for what it is designed to do.
I think that for a specific situation/project where you have a load of books etc to scan then you'd make this work and be happy with the results, otherwise I'd think it wasn't worth the hassle or bench space for it.
Did you notice that it removed the shading on the chip diagram in the datasheet? I found it's impossible to get good results with shading, making many diagrams impossible to scan. Overall the software is buggy and they don't seem to be able to fix major flaws like that.
The type of content would be really fun and amazing to hear from you! Could look at that all day - well guess that's whats already happening :D The camera setup seems not perfectly ideal, feel a screen capture of a tablet or so could lead to smoother images
Thanks Shahriar for another great video! Only not to confuse younger engineers watching the video: In the analysis of the cadcode there was a minor oversight (the PMOS transistor Q3 was assumed to be a bipolar transistor). The gain assuming a PMOS active load is obviously again independent of the DC current: gain~-gm1 ro3=-1/(Vt x lambda). ro3 =1/(lambda x ID), where lambda is modeling the channel length modulation (level 1 cmos model) -> gain~-gm1 ro3=-1/(Vt x lambda). 😊
I am not sure what you are referring to. I did not assume that Q3 is a bipolar. I assumed it is a PMOS as it was drawn. The input transistor is a bipolar.
@@Thesignalpath the small signal drain source resistance of Q3 is 1/(lambda ID) and not VEarly/ID. Early effect /early voltage is usually defined for bipolar transistors. In a MOSFET the parameter which models the dependency of the drain current on Vds is lambda (assuming simple level 1 model). Both effects have of course similar physical roots: extension of space charge layer of a reverse biased pn junction. In bjt (CB junction) this will result in a „shorter“ base width, therefore less recombination of minority carriers in base and therefore larger current. In cmos (DB-junction) resulting in smaller channel length and larger current: I proportional to W/(L-deltaL).
Although the early voltage VA is a parameter which originated from BJTs, due to the similarity of the effect of (1/lambda) on the output resistance, the term is also used for MOSFETs. This helps maintain the similarity of the effect on various calculations and we usually just say for MOSFETS 1/lambda = VA. I should have written lambda in the derivation for completeness too, which I always do with my students. You can also see it mentioned here: en.wikipedia.org/wiki/Early_effect I did however specifically mention that the VA of the MOSFET scales with its length. If I had thought it was a bipolar, that would have made no sense.
When I read "vertical scanner" for some reason I thought it creates a 3D Model of the object underneath with the help of lasers and so on... the price tag would justify that, but that's just a camera on a stand ..with an integrated light source. ^^ One of the first things I build from a 50 bucks action cam and a magnetic machinist mount and some 3d printing when I started smd soldering. I replaced the lens on the camera though, it 's so nice to be able to zoom in on your work and read the fine print on ICs on a big screen.. occasionally scan a qr tag etc. It will probably also scan books, that's purely a software thing I guess. it cost me around 80 bucks all in all.
Looks very, very unpleasant to my eye. :o/ The contrast is too high, color temperature is too cold, shadows are so sharp that even the hairs becomes distracting. For no shadows and pleasant illumination maybe try diffuse light, e.g. a white fabric tent to diffuse the light, or other light-body with big area and far away, or reflected light from white walls, etc. Preferably use incandescent light bulbs.
Colors from the video camera are completely fucked up, and it seems there is some weird denoise combined with extreme detail enhancement. These settings may be good for scanning a document, but using it as a generic camera looking at a table? Better use your cell phone camera than this, will likely produce much better image. Not sure if you can tweak any settings of this contraption. In case not, back to square one, using a normal camera that has a decent image.
Well, when I received the unit there were not any reviews that I knew about. It takes me some time to make videos as I have many other things to do. Of course you can always skips a video which is not of interest to you.
@@Thesignalpath you can stick whatever you want on your channel. It's your channel not ours! I think we're all terribly lucky that you share you immense skills for free with the world. Many people, myself included, have a far better understanding of highly specialized RF engineering entirely because of your generosity. If you post it, I'll always watch it
The document digitisation looks excellent. The live video less so. Big Clive's set-up with a relatively inexpensive smartphone and LED lights clamped to a bookshelf seems to be noticeably better lit, with less harsh contrast, higher resolution and better clarity in general. He seems to have just as good a control over focus, brightness, zoom, framing etc. You may want to consider using the paper in landscape mode when recording.
The processing is terribly offputting, I'd say it outweighs the positives. There's a lot of encoding 'dancing' when its interpolating an already oversharpened image. The old scottish Clive has the best balance. I really don't need a perfect top down over a more pleasing rendering... Cheers!
I have one of this scanners.
The hardware seems to be quite good, but the software leafs room for improvement.
The trick with the searchable PDFs is, to keep the picture and use the OCR detected text only for the search. As far as I remember this option isn't the default.
The page flattening function is very usable in most cases, unless the page is really glossy.
My biggest issue with this feature is, that it only works if you scan both pages of a book at once. I have a lot of stuff, that has some pages, where you can fold out a drawing, which makes each page double wide. I wasn't able to scan these properly as a single page yet.
The automatic finger removal is also soso. Sometimes it works, sometimes not at all and sometimes it creates strange artifacts. I avoid using it, when I can.
When there is text near the edge of the page, it is imossible to place the plastic tingy in a way, that it gets detected.
I suggest to use a slightly finer marker maybe. The edge enhancement looks to be optimized for thinner lines. Other than that it looks nice!
Thanks!
Thanks! Now I'd like to see what books you have on the shelf about electronics. ;-) Perhaps also good idea to post hi-res photos of device guts on your site. Would be great when you do reviews, especially for newer and more expensive devices, often there is lack of information about them online, and sometimes it's irreplaceable for somebody who is just looking for visual photo, for example to check what is missing/cabling.
The contrast is high on the video section, but it seems like there's pretty heavy sharpness adjustment on it too, if you can get the edges to be less harsh and the background a little less bright, it will probably be more comfortable to look at. I'd second (third, however many) the want for a landscape orientation. Not entirely necessary for the size of the text you're using, but works better non-fullscreen viewing and for having the subject of the video actually taking up most of the frame. Even sort of background/filler would be preferable in my mind to just the black mat.
Thanks for the feedback.
Is there a way to turn down the contrast / edge sharpening in the live HDMI viewer?
I could not immediately find a way to do it.
@@Thesignalpath Have to agree the live contrast edge sharpening is really off putting. All of the live video through the thing is terrible. In the the LT doc it washed out out the LT logo it turned white in the center. But most of the pages look pretty good for what it is designed to do.
@CuriousMarc is very fond of this one for scanning and preserving old computer and HP test gear manuals 🙂
Do you know of an example of how the results look like (eg. a HP-typical huge schematic)?
I think that for a specific situation/project where you have a load of books etc to scan then you'd make this work and be happy with the results, otherwise I'd think it wasn't worth the hassle or bench space for it.
The scanner is nice --- but I think I will be looking for the book ;)
Thanks again
Did you notice that it removed the shading on the chip diagram in the datasheet? I found it's impossible to get good results with shading, making many diagrams impossible to scan.
Overall the software is buggy and they don't seem to be able to fix major flaws like that.
The type of content would be really fun and amazing to hear from you! Could look at that all day - well guess that's whats already happening :D The camera setup seems not perfectly ideal, feel a screen capture of a tablet or so could lead to smoother images
Thanks Shahriar for another great video!
Only not to confuse younger engineers watching the video:
In the analysis of the cadcode there was a minor oversight (the PMOS transistor Q3 was assumed to be a bipolar transistor).
The gain assuming a PMOS active load is obviously again independent of the DC current:
gain~-gm1 ro3=-1/(Vt x lambda).
ro3 =1/(lambda x ID), where lambda is modeling the channel length modulation (level 1 cmos model) -> gain~-gm1 ro3=-1/(Vt x lambda).
😊
I am not sure what you are referring to. I did not assume that Q3 is a bipolar. I assumed it is a PMOS as it was drawn. The input transistor is a bipolar.
@@Thesignalpath the small signal drain source resistance of Q3 is 1/(lambda ID) and not VEarly/ID.
Early effect /early voltage is usually defined for bipolar transistors.
In a MOSFET the parameter which models the dependency of the drain current on Vds is lambda (assuming simple level 1 model).
Both effects have of course similar physical roots: extension of space charge layer of a reverse biased pn junction.
In bjt (CB junction) this will result in a „shorter“ base width, therefore less recombination of minority carriers in base and therefore larger current.
In cmos (DB-junction) resulting in smaller channel length and larger current:
I proportional to W/(L-deltaL).
Although the early voltage VA is a parameter which originated from BJTs, due to the similarity of the effect of (1/lambda) on the output resistance, the term is also used for MOSFETs. This helps maintain the similarity of the effect on various calculations and we usually just say for MOSFETS 1/lambda = VA. I should have written lambda in the derivation for completeness too, which I always do with my students. You can also see it mentioned here: en.wikipedia.org/wiki/Early_effect
I did however specifically mention that the VA of the MOSFET scales with its length. If I had thought it was a bipolar, that would have made no sense.
Thanks for explanation!
I appreciate that you brought it up. Now anyone reading this will also benefit from your comments.
Looks like its contrast is set to 100%
It does. It helps with scanning, but does make a somewhat jarring image through video.
When I read "vertical scanner" for some reason I thought it creates a 3D Model of the object underneath with the help of lasers and so on... the price tag would justify that, but that's just a camera on a stand ..with an integrated light source. ^^
One of the first things I build from a 50 bucks action cam and a magnetic machinist mount and some 3d printing when I started smd soldering. I replaced the lens on the camera though, it 's so nice to be able to zoom in on your work and read the fine print on ICs on a big screen.. occasionally scan a qr tag etc. It will probably also scan books, that's purely a software thing I guess. it cost me around 80 bucks all in all.
The heavy post processing on the HDMI output looks awful. On the other hand it scans really nice. I wish I had one for a couple old manuals I have.
Looks very, very unpleasant to my eye. :o/
The contrast is too high, color temperature is too cold, shadows are so sharp that even the hairs becomes distracting.
For no shadows and pleasant illumination maybe try diffuse light, e.g. a white fabric tent to diffuse the light, or other light-body with big area and far away, or reflected light from white walls, etc. Preferably use incandescent light bulbs.
Thanks for letting me know.
Looks like the "zoom" is software cropping, not optical. No details on their web site.
Yes, it is digital zoom.
@@Thesignalpath Thanks for the reply
Colors from the video camera are completely fucked up, and it seems there is some weird denoise combined with extreme detail enhancement. These settings may be good for scanning a document, but using it as a generic camera looking at a table? Better use your cell phone camera than this, will likely produce much better image.
Not sure if you can tweak any settings of this contraption. In case not, back to square one, using a normal camera that has a decent image.
It does seem highly tweaked for their particular use-case.
👍👍
The insane oversharpening makes everything look horrible. books included. Looks like this is only good for feeding into OCR.
Prefer your handwriting on the tablet PC from previous videos.
In case anyone is interested: $629.
This scanner is $629.
This has been flogged on multiple channels recently.
Third review of this exact product I'm getting pushed by my favorite channels. It getting boring.
Well, when I received the unit there were not any reviews that I knew about. It takes me some time to make videos as I have many other things to do. Of course you can always skips a video which is not of interest to you.
@@Thesignalpath you can stick whatever you want on your channel. It's your channel not ours! I think we're all terribly lucky that you share you immense skills for free with the world. Many people, myself included, have a far better understanding of highly specialized RF engineering entirely because of your generosity. If you post it, I'll always watch it