10' e-ink readers? Kindle Scribe, Kobo Elipsa 2E...

[Quoth]

Originally Posted by jackm8

It is 2480 x 1860 black and white. That makes it the same as Scribe. The again it's also lower contrast compared to pure black and white eink screens, so yeah, it's off the radar because of that.

No, there is colour filter on top of the 2480 x 1860. It physically can't look like a Scribe.
It might appear better than 223 approx to some people, but it's by using coloured edges to enhance sharpness.
Unless you take it apart and scrape the coloured filter off it can never do 2480 x 1860 black and white / 300 dpi mono. That's the underlying panel resolution.
You can't call a 1920 x 1080 Colour LCD panel a 5760 x 1080 black and white (stripe filter) or a 3840 x 2160 (4k) black and white (Bayer filter), though that's the resolutions of the mono LCDs used to make them. Cleartype or similar is used to make high contrast text seem sharper by using coloured edges (sub pixel addressing).
That's what the colour eink is doing to get better than 150 dpi (Bayer pattern) or on older striped panels, better than 100 x 300 dpi.

[cellaris]

Originally Posted by Quoth

It's not. That's rather misleading. It's a 300 dpi mono screen with a patterned colour filter on top. You can do tricks to improve the resolution that mono high contrast stuff has compare to the resolution a colour image has, but it won't be as sharp for mono text and images as even a 223 dpi screen.
It's 150 dpi on colour and can be made to be a bit sharper with mono text and subpixel addressing.

I think you are wrong. The differences between Kaleido 3 and a 300 ppi monochrome e-reader are two: the greater darkness of the screen without illumination on and a grainy background that is barely noticeable, but if you look closely you can see it. But when it comes to contrast and sharpness there is no difference. An e-reader with Kaleido 3 and the illuminated display is virtually identical to any other e-reader with the same resolution. You would need a microscope to find significant differences. And, of course, the difference with a lower resolution e-reader is noticeable. Much better an e-reader with Kaleido 3 than one with 227 ppi. As @ottischwenk says, I also have devices with these screen resolutions and my experience tells me that Kaleido 3 (with illumination) has a better display than a 227 ppi monochrome e-reader. And personally I think that the differences with another 300 ppi monochrome e-reader (darkness of the screen without illumination and grainy backgroun) are more than compensated by the colour contribution. I use an e-reader with Kaleido 3 for regular reading and I don't miss a monochrome e-reader (and I have monochrome e-readers with the best display you can find on the market today).

And I have to add that this came as a surprise to me, because I really thought like you. I thought that a colour e-reader would look much worse than a monochrome e-reader. And now I can't say that. I think that the Kaleido 3 is a breakthrough and that the differences are not objective, but a matter of personal assessment as a whole. Personally I have no objection now to recommending a colour e-reader. I don't think it is better or worse than a monochrome e-reader, but different. And it is up to each user to personally assess whether the difference is worth it.

[kandwo]

Originally Posted by jackm8

So, I've been looking at Note Air 3 C as well lately. As high resolution BW screen as Scribe. But what I really like, is that they've a very thin glass cover over the screen. On other hand it's 150€ more compared to the rest. Granted it's got coloured display, but I really don't see that as a bonus.
I wonder if there's any resource out there that mentions glass thickness of the screens, either in reviews or in any other way?
Only way that I can get a feel for this is hands on or by watching some youtoube reviews from people that do detailed footage like My Deep Guide. Problem with the later is inconsistent video quality, problem with the former is that my local stores don't have all the models on hands to try.

My Deep Guide is good, but his main use case is note taking and screen sharpness and contrast is a secondary priority for him.

The Note Air 3 C will be released in black and white later this year according to Boox. The problem with it is that it has an additional layer on the glass screen to make the writing experience better. This means that the contrast and sharpness will be slightly reduced compared to something like the Tab Ultra C.

I think that I will personally wait until they release a 300 PPI Tab Ultra reader in black and white.

[montessori]

Originally Posted by ottischwenk

I have both, can compare directly and the Kaleido 3 300ppi screen is much sharper for text than the 227 ppi of the Tab Ultra and Note Air 1.
Even images appear sharper

I have Kaleido 3 300ppi screen and is exactly what you need.

[ottischwenk]

Originally Posted by cellaris

I think you are wrong.

He certainly is - black must be 300 ppi; and when reading in black on white this is important

[Quoth]

A one pixel black line will be a real 300 dpi, but the white background will have pale coloured fringes.
A one pixel white line line isn't white. You need a two pixel line for white and certain white diagonals at 2 pixel wide will not be white.

There is no longer a mono 300 dpi panel because of the bayer filter pattern (or similar as it's 150dpi in colour).
https://en.wikipedia.org/wiki/Bayer_filter

Do learn some physics.

Make sure there is no colour and a range of text in same font and line art on two ereader being compared and the front light is perfectly adjusted to same white. Then via cut out window have test subjects compare. Repeat several times with sometimes swapping positions.

Then do same with enough room lighting for front light off.

For reading mono text, and especially in dark mode or ambient light, the Kaliedo 3 is not 300 dpi mono. It's impossible.

It's pointless unless you want comics. It's pretty rubbish for photos being only a 16 level panel (Up to 4095 combinations of shades / colours and black) and desaturated filer so it's not too dark like Triton.

I'm sometimes wrong, but not on this, it's a marketing lie to claim a 300 dpi panel with a filter for colour can still do 300 dpi monochrome. A one pixel black line with coloured fringes can be 300 dpi, nothing else.

I was writing colour manipulation SW 30 years ago for CGA, EGA and 256 colour VGA.

[ottischwenk]

Originally Posted by Quoth

A one pixel black line will be a real 300 dpi, but the white background will have pale coloured fringes.

Colorful white at the edges can only be seen with a magnifying glass, not without it - I certainly don't use one to read.

And what is theoretically possible is of very little interest to me when it comes to applying it.

Edit:

A one pixel black line with coloured fringes can be 300 dpi, nothing else.

The correct answer would be: colored fringe only visible with a magnifying glass; when I read, I don't care at all whether they exist theoretically - I don't use a magnifying glass when reading, neither on color nor on grayscale devices.

I have no idea what the trick is here, but when I read it the text seems clearer than in Tolino Epos 3

[Quoth]

The 2 x 2 Bayer filter or similar will absolutely work better than a 3 x1 stripe filter for text.
But neither works so well for text unless the font is black on a white background and there is sub (colour) pixel addressing that especially "understands" what to do with vertical and diagonal edges.

We are a bit more sensitive to text resolution in the horizontal direction anyway.

[jackm8]

Originally Posted by Quoth

A one pixel black line will be a real 300 dpi, but the white background will have pale coloured fringes.
A one pixel white line line isn't white. You need a two pixel line for white and certain white diagonals at 2 pixel wide will not be white.

There is no longer a mono 300 dpi panel because of the bayer filter pattern (or similar as it's 150dpi in colour).
https://en.wikipedia.org/wiki/Bayer_filter

Do learn some physics.

Make sure there is no colour and a range of text in same font and line art on two ereader being compared and the front light is perfectly adjusted to same white. Then via cut out window have test subjects compare. Repeat several times with sometimes swapping positions.

Then do same with enough room lighting for front light off.

For reading mono text, and especially in dark mode or ambient light, the Kaliedo 3 is not 300 dpi mono. It's impossible.

It's pointless unless you want comics. It's pretty rubbish for photos being only a 16 level panel (Up to 4095 combinations of shades / colours and black) and desaturated filer so it's not too dark like Triton.

I'm sometimes wrong, but not on this, it's a marketing lie to claim a 300 dpi panel with a filter for colour can still do 300 dpi monochrome. A one pixel black line with coloured fringes can be 300 dpi, nothing else.

I was writing colour manipulation SW 30 years ago for CGA, EGA and 256 colour VGA.

That link is completely off. It talks of Bayer filter for digital imaging sensors. It's not even applicable for lcd monitors, even less for e-ink devices. It's system used for capturing images on digital cameras, not for displaying them. Eink screens are different, and coloured eink screens are another thing altogether. From all that I've seen of Kaleido screen tech, it's BW screen underneath then coloured filter directly on top of it. That way they can both be off different resolution.

[Quoth]

Cameras use monochrome sensors with a colour filter on top, unless using a splitter and three sensors. Used to be stripe, but now variations of Bayer. So the camera sensor is 3 x1 pixels or 2 x 2 pixels per usable image pixel.

LCDs are monchrome, like eink, except illuminated from the back instead of the front. Some still use RGB stripe filters, but many now use Bayer pattern filters. QLED are actually mono LCD with quantum dots to frequency change blue light to red or green. They can use a Bayer pattern or stripe, but blue bits are clear and the backlight is pure blue LED source instead of white*.

The physical and mathematical principles of colour filters on an image detector and image generator are the same. Early colour using mono photographic film had the filter on it. After development using image reversal (so it's not a negative) the film was projected using the original integral colour filter (often random R G B particles).

[* White LEDs don't exist. They are either really multiple R, G & B LEDs, which is rubbish colour rendition, or a single blue or violet LED with a phosphor to make it white by adding colours energised by blue or violet.]

[ottischwenk]

Originally Posted by Quoth

Bayer.

Bayer (RGGB) is certainly not used.
The filter used here is RGB, whereby one of the 3 colors is always present in 2 other and different filter strengths in the 2*2 matrix.
It is clearly visible with green, less so with red and blue.

[Quoth]

Any pattern in a 2 x 2 turns mono dpi into 1/2 colour dpi, i.e. 300 dpi underlying panel giving 150 dpi colour. Those are all variations of the original Bayer idea. The alternative is stripes (3 x 1). That gives same resolution in one direction for colour as underlying display and 1/3 resolution in the other direction.

To do colour on ANY monochrome panel (eink or LCD) using a coloured pattern filter reduces the effective resolution even for mono images. This is basic optics, physic and mathematics, no matter how clever the filter is or what name you give it.
There are more complex schemes using unfiltered pixels that allow brighter monochrome. They further reduce resolution of saturated images but allow slightly higher apparent resolution of pastel images.

It absolutely doesn't matter what pattern or layout or system is used. The only way to get colour on a monochrome panel / display of any kind is a coloured filter pattern or stripes aligned to the pixels. Then the underlying native panel resolution is gone. Certainly some kinds ot mono text or images can still be apparently sharper than saturated coloured ones, but they can't display as high as resolution as the unfiltered panel. It's just total fantasy to claim true 300 dpi mono on a panel that's 300 dpi with a coloured pattern on top. You'd only get the real 300 dpi mono if your scraped the coloured pattern (or striped on the older panels) off.

Bayer is
RG
GB
in a 2 x 2 square of native pixels for each pixel the viewer sees. Or a variation of it, rather than simple stripes as was originally used on sensors and flat panels (3 x1 layout of R G B). Though CRTs used triangles of R G B before using stripes and the cameras used either 3 sensors or stripes.

[ottischwenk]

Originally Posted by Quoth

It's just total fantasy to claim true 300 dpi mono on a panel that's 300 dpi with a coloured pattern on top.

This is correct for black and white - but incorrect for black/something; Black is always 300 ppi and as long as you think you recognize the "something" as white, it doesn't matter to the viewer.

And that is exactly where your mistake lies.

And the most important thing is overlooked - BSR and Carta 1200 with 7.8 inches and bigger are only available with CFA

[jackm8]

Originally Posted by ottischwenk

This is correct for black and white - but incorrect for black/something; Black is always 300 ppi and as long as you think you recognize the "something" as white, it doesn't matter to the viewer

True, e-ink isn't Bayer array at all. Kaleido screens are basically two individual screens in one. Just like Foveon screens mentioned in that wikipedia article. They are stacked.

Spoiler:

Really unique tech, used on a few Sigma cameras. They're low resolution sensors that trade resolution for colour rendering. Plenty of shortcomings as well, from battery life to low light performance, so they're sidelined at the moment. A few links explaining them with some samples: https://petapixel.com/foveon-x3-image-sensor-explained/
https://www.mpb.com/en-us/content/ki...-foveon-sensor

First layer is 300p layer at the bottom, then over it, it's lower resolution transparent rgb layer. When colour isn't utilised that makes screen 300p, albeit white tones will be of more grayish look, as coloured filter can not be completely invisible, even when not in use.

Looking at this screen from mydeepguide's review of boox nova 3 color, it's plainly seen how this works. Coloured pixels are much larger, but are transparent. White spot pixels of bw array can still be seen under them. The same should go for black pixels as well, but they are not plainly visible in this sample. This possibly means that coloured filter has a very small, practically invisible effect on black content. It could also be that they're not seen from image quality uploaded.

It's a complex tech, and little is known how it actually works. I'm almost leaning towards it right now so I could just see how it functions in practice. How does the device decide whether to utilise 300p or lower resolution? Vectored fonts in pdf could go high resolution, but what if pdf text isn't rasterized, and includes coloured content, but is still in 300p resolution or higher. Would that downscale everything and display text at coloured resolution?

[DNSB]

Originally Posted by ottischwenk

And the most important thing is overlooked - BSR and Carta 1200 with 7.8 inches and bigger are only available with CFA

I'll have to break the news to one acquaintance of mine who owns a Boox Tab 8, that his device does not have BSR despite the claims made in the advertising on Boox's web page. See attached images.