3 ms·

> Have you seen the photo of an ACeP rendering? Where do the "extra" colours come from, full-pixel-level dithering? What do you mean "ACeP rendering"? Like art

> Have you seen the photo of an ACeP rendering? Where do the "extra" colours come from, full-pixel-level dithering? What do you mean "ACeP rendering"? Like artificial photo? I've seen ACeP at SID 2019 in real life and I could see from the zoom that there were only a small number of actual colors, less than 10 for sure. > ACeP has subpixel complexity: every pixel contains many particles of all the four pigments. That is of course perfectly apt, further implementation constraints aside, to represent the full logical range of combinations within a single pixel. Where do you see any statement that they can achieve "full logical range of combinations within a single pixel"?

mdp2021 5y ago
Sorry for the delay. For «photo of an ACeP rendering» I meant some "witnessing" representation of the results of ACeP's workings; you have seen it live: good. I have not, and I am relying on the available material. From the pictures I could see, the apparent colours are many. Traditional B/W E-Ink displays achieve that (i.e. pixels rendering grey shades) through in-pixel dithering - within the cell you get an amount of B-particles and an amount of W-particles. To get similar "detail", or "effect of more colours being shown", with out-pixel dithering, you would need a proportionally high resolution. Some available photographic material taken of ACeP displays "in action" seems to exclude out-pixel dithering, hence work as one would suppose - you cannot get that richness with a "mosaic". This, especially if implicitly compared with some displays of similar technology that use fully discrete colour states (e.g. B/W/R). So, if the ACeP you saw at SID 2019 revealed using discreet states of primary colours (e.g. BWRGBYO), is it possible that you saw an implementation similar to the "ACeP E Ink Gallery Palette" instead of the "full" "ACeP E Ink Gallery"? Because: again: what is the difference between the two, if they both were "discreet BWRGBYO", without any "traditional" in-pixel dithering? Of course, judging from the pictures can be tricky: the Waveshare WS-18295 as seen at https://www.waveshare.com/5.65inch-e-Paper-Module-F.htm https://www.waveshare.com/5.65inch-e-Paper-Module-F.htm seems to show many more colours than the seven full-discreet. The small palette is much more evident in the corresponding E Ink AB1024-EGA ("E Ink Gallery Palette 4000"), as seen at https://shopkits.eink.com/product/5-65%cb%9d-e-ink-gallery-palette-4000-epaper-display-ab1024-ega/ https://shopkits.eink.com/product/5-65%cb%9d-e-ink-gallery-p... . Such gross dithering is of course expected as that product is explicitly an "ACeP-EIG*P*". Edit: and by the way, to achieve that effect of richness in colours that the available "photos" show is increasingly difficult using only "four inks and a base": there is no (e.g.) native yellow and orange in state-of-the-art "E Ink Gallery Plus ACeP"... > Where do you see any statement that they can achieve "full logical range of combinations within a single pixel" I have never expressed that and the idea to "achieve a logical range" makes no sense in the context. I said that subpixel complexity in ACeP is logically sufficient to represent a full range of combination without bringing on the table ideas of colour composition imported from heterogeneous technologies such as LCD.
- robinsoh 5y ago
  > judging from the pictures can be tricky: the Waveshare WS-18295 as seen at https://www.waveshare.com/5.65inch-e-Paper-Module-F.htm https://www.waveshare.com/5.65inch-e-Paper-Module-F.htm seems to show many more colours than the seven full-discreet. I don't agree at all. Zoom in to the pictures of any of that 5.65 panel. It is literally very visibly dithered and shows discrete 7 colors only! There is no "subpixel complexity". You can choose one value of 7 colors for any pixel and that's it. You can't mix. You don't have "range" within each color. https://www.waveshare.com/media/catalog/product/cache/1/image/800x800/9df78eab33525d08d6e5fb8d27136e95/5/_/5.65inch-e-paper-f-2_2.jpg https://www.waveshare.com/media/catalog/product/cache/1/imag...
  - mdp2021 5y ago
    > I don't agree at all. Zoom in... You can choose one value of 7 colors for any pixel and that's it But that was not the point. The point was «judging from the pictures can be tricky»: the Waveshare WS-18295 ones were just examples. That there could be better examples is irrelevant. Of course in that Waveshare WS-18295 «You can choose one value of 7 colors for any pixel and that's it»: it's explicitly declared - it's called "ACeP 7-Color", and its "greyscale" specification is "2"... That does not mean that other displays may not be more complex. But incidentally, again about dithering (internal and external) and the difficulty of judging from pictures of that scale and focus: in that very Waveshare of the cheap range, where ACeP internally remains with fullest probability a CMYW, while functionally a BWRGBYO, pigment composition is invisible in those pictures, and this while the mixes are not trivial - both R and O must come from Y+M, through different amounts (from different voltage switch tricks)... Other material insists that they have more capabilities in the park of achievements: the article https://spectrum.ieee.org/how-e-ink-developed-full-color-epaper https://spectrum.ieee.org/how-e-ink-developed-full-color-epa... , for IEEE Spectrum (22 Jan), insists that «By stopping the particles at different levels, we can use these particles to create up to 50,000 colors». While the theoretical difficulty of moving precisely four classes of particles through the same input of "voltage variation in time" remains, if that statement did not mean that they can actually somehow determine where to move the particles "halfway" (more states than just halfway) through the microcups - well, there would be no reasonable straight interpretation of that statement. This given, while "60000" strongly suggests "2^4*4", it remains a mystery how they could have achieved that - again, because what they seem to have remains "voltage switch tricks over different sized and charged particles in the microcups".

mdp2021 5y ago

Sorry for the delay. For «photo of an ACeP rendering» I meant some "witnessing" representation of the results of ACeP's workings; you have seen it live: good. I have not, and I am relying on the available material. From the pictures I could see, the apparent colours are many. Traditional B/W E-Ink displays achieve that (i.e. pixels rendering grey shades) through in-pixel dithering - within the cell you get an amount of B-particles and an amount of W-particles. To get similar "detail", or "effect of more colours being shown", with out-pixel dithering, you would need a proportionally high resolution. Some available photographic material taken of ACeP displays "in action" seems to exclude out-pixel dithering, hence work as one would suppose - you cannot get that richness with a "mosaic". This, especially if implicitly compared with some displays of similar technology that use fully discrete colour states (e.g. B/W/R). So, if the ACeP you saw at SID 2019 revealed using discreet states of primary colours (e.g. BWRGBYO), is it possible that you saw an implementation similar to the "ACeP E Ink Gallery Palette" instead of the "full" "ACeP E Ink Gallery"? Because: again: what is the difference between the two, if they both were "discreet BWRGBYO", without any "traditional" in-pixel dithering? Of course, judging from the pictures can be tricky: the Waveshare WS-18295 as seen at https://www.waveshare.com/5.65inch-e-Paper-Module-F.htm https://www.waveshare.com/5.65inch-e-Paper-Module-F.htm seems to show many more colours than the seven full-discreet. The small palette is much more evident in the corresponding E Ink AB1024-EGA ("E Ink Gallery Palette 4000"), as seen at https://shopkits.eink.com/product/5-65%cb%9d-e-ink-gallery-palette-4000-epaper-display-ab1024-ega/ https://shopkits.eink.com/product/5-65%cb%9d-e-ink-gallery-p... . Such gross dithering is of course expected as that product is explicitly an "ACeP-EIG*P*". Edit: and by the way, to achieve that effect of richness in colours that the available "photos" show is increasingly difficult using only "four inks and a base": there is no (e.g.) native yellow and orange in state-of-the-art "E Ink Gallery Plus ACeP"... > Where do you see any statement that they can achieve "full logical range of combinations within a single pixel" I have never expressed that and the idea to "achieve a logical range" makes no sense in the context. I said that subpixel complexity in ACeP is logically sufficient to represent a full range of combination without bringing on the table ideas of colour composition imported from heterogeneous technologies such as LCD.

robinsoh 5y ago
> judging from the pictures can be tricky: the Waveshare WS-18295 as seen at https://www.waveshare.com/5.65inch-e-Paper-Module-F.htm https://www.waveshare.com/5.65inch-e-Paper-Module-F.htm seems to show many more colours than the seven full-discreet. I don't agree at all. Zoom in to the pictures of any of that 5.65 panel. It is literally very visibly dithered and shows discrete 7 colors only! There is no "subpixel complexity". You can choose one value of 7 colors for any pixel and that's it. You can't mix. You don't have "range" within each color. https://www.waveshare.com/media/catalog/product/cache/1/image/800x800/9df78eab33525d08d6e5fb8d27136e95/5/_/5.65inch-e-paper-f-2_2.jpg https://www.waveshare.com/media/catalog/product/cache/1/imag...
- mdp2021 5y ago
  > I don't agree at all. Zoom in... You can choose one value of 7 colors for any pixel and that's it But that was not the point. The point was «judging from the pictures can be tricky»: the Waveshare WS-18295 ones were just examples. That there could be better examples is irrelevant. Of course in that Waveshare WS-18295 «You can choose one value of 7 colors for any pixel and that's it»: it's explicitly declared - it's called "ACeP 7-Color", and its "greyscale" specification is "2"... That does not mean that other displays may not be more complex. But incidentally, again about dithering (internal and external) and the difficulty of judging from pictures of that scale and focus: in that very Waveshare of the cheap range, where ACeP internally remains with fullest probability a CMYW, while functionally a BWRGBYO, pigment composition is invisible in those pictures, and this while the mixes are not trivial - both R and O must come from Y+M, through different amounts (from different voltage switch tricks)... Other material insists that they have more capabilities in the park of achievements: the article https://spectrum.ieee.org/how-e-ink-developed-full-color-epaper https://spectrum.ieee.org/how-e-ink-developed-full-color-epa... , for IEEE Spectrum (22 Jan), insists that «By stopping the particles at different levels, we can use these particles to create up to 50,000 colors». While the theoretical difficulty of moving precisely four classes of particles through the same input of "voltage variation in time" remains, if that statement did not mean that they can actually somehow determine where to move the particles "halfway" (more states than just halfway) through the microcups - well, there would be no reasonable straight interpretation of that statement. This given, while "60000" strongly suggests "2^4*4", it remains a mystery how they could have achieved that - again, because what they seem to have remains "voltage switch tricks over different sized and charged particles in the microcups".