Processor Speeds of PW3 & the New PW4 ???

[heyman]

Originally Posted by DuckieTigger

Yes. The eink controller has a native inverted mode on Oasis 2 vs Oasis 1 (in hardware, not software).

if we mark kindle's inverted mode as different, that would be right. koreader is software invert.

but i read NiLuJe's other post

https://www.mobileread.com/forums/sh...er#post3830808

The controller is certainly the updated one found on the Oasis 2 & PW4

the new 167dpi entry one is also using the Oasis 2 controller
and that one doesn't have inverted mode which means inverted mode doesn't always mean newer

[DuckieTigger]

Originally Posted by heyman

if we mark kindle's inverted mode as different, that would be right. koreader is software invert.

but i read NiLuJe's other post

https://www.mobileread.com/forums/sh...er#post3830808



the new 167dpi entry one is also using the Oasis 2 controller
and that one doesn't have inverted mode which means inverted mode doesn't always mean newer

What does koreader have to do with the controller? And yes, the inverted mode on the Oasis 2 is NOT done in software only, so not the same as what koreader does.

[NiLuJe]

I use the term "Controller", because that's what the C stands for in EPDC (the rest standing for Electronic Paper Display), and that's how the driver is referred to in the kernel .

Now that we're done with nitpicking, a bit more details:

Basically *all* revisions of the MXC EPDC (i.e., since the i.MX5, basically, the Kindle Touch in Kindle-land) have had support for hardware inversion.
It may have been a bit buggy in some instances (the Kobo Aura, apparently), but it was there.

What lab126 did (possibly as early as FW 5.5.x, my old PW2 on FW 5.9.7 certainly behaves differently (i.e., better) than my Kobo H2O in that respect) was start to tweak things so that flashes behaved less inanely when said inverted flag is flipped, and to enable a specific waveform update mode better suited to white on black.

What they did with the Oasis 2 is both refine those waveform tweaks (there's a couple of 'em now), and add a bit of frontlight modulation to the mix to avoid searing eyeballs during flashes.

But all that's potentially completely orthogonal to the EPDC v2 changes. In my eyes, the most user-visible changes that it brought would be a much refined hardware dithering handling. It works wonder on the Forma, but I lack the Kindle hardware to check on the KOA2/PW4, and early feedback when we tried that in KOReader would imply things aren't so rosy there, even if it should technically work as well as on Kobo Mk.7...


----

Also, sidebar: KOReader now does HW inversion, unless it's known to be crashy, or we lack the developers to test it (which right know means it's disabled on the Kobo Aura, and on PocketBook).

If you have access to a Kobo, a Kindle Touch or PW1, and a Kindle PW2+ on FW >= 5.5.x, it's easy enough to check: run a current KOReader version, and enable NightMode.
We'll try to use the optimal waveform mode when there's one available (i.e., the earlier iteration on Kindle FW >= 5.5.x, and the refined one on PW4/KOA2) and we'll do HW inversion.
Your Kobo and your Touch/PW1 will spaz out like idiots during flashes, your PW2+ won't .

----

EDIT: And, yes, the Oasis 1 was using an earlier iteration of the EPDC, unlike the KOA2/PW4/KT4 | Forma/Clara.

[NiLuJe]

Hell, the original einkfb one could also do inversion, it was just a tiny bit more unwieldy (fun fact: it's actually running with an inverted palette by default on Kindle, even on the K4, which is all kinds of fun to handle).

[heyman]

it reall confuses me that newer EPDC doesn't mean better display performace

the Kindle voyage (2014) still is the most sharp in any kindle, the "sharpness" of it is unbeatable

I found this, actual contrast measurements of contrast, voyage is actually really better than any kindle. Naked eyes and science are on the same page now, this is funny

https://allesebook.de/ebook-reader/kindle-oasis-1/ (see Kontrast)

[heyman]

Originally Posted by NiLuJe

What they did with the Oasis 2 is both refine those waveform tweaks (there's a couple of 'em now), and add a bit of frontlight modulation to the mix to avoid searing eyeballs during flashes.

what is anti-eye-searing algorithm? can we also add our own algorithm/waveform to koreader to better suit eink?

[NiLuJe]

Can't tell without a device, as I don't even know if it's actually used anywhere. That wouldn't be the first time the kernel exposes a feature that ends up being unused.

In any case, replicating that sort of thing on devices without hardware support went nowhere, because of the latency involved in playing with the FL on various devices, combined with how our main loop is structured.

[rashkae]

Originally Posted by heyman

what is anti-eye-searing algorithm? can we also add our own algorithm/waveform to koreader to better suit eink?

It's a really neat trick Amazon added to the inverted text. Whenever the Kindle has to do a full page refresh, it first turns off the light, so you the screen doesn't flash to all white with the light on high. The light is then turned back on once the screen is black again.

[heyman]

I think the algorithm controls how the refresh "behavior"

algorithm controls how the refresh display when the texts fade and reappear, I notice that every kindle are slightly different when the texts fade

the text isn't fade all at once, it fades part of the texts first "to anti-eye-searing" and then fade the other part of the texts
on paperwhite 3 it fades left-upper corner of the texts first and then the other part, it's funny to observe these behaviors

there's a pattern i'm sure

[heyman]

btw, I used the word "refresh" but I actually mean partial refresh, not full-refresh

I should say page turning behavior

[NiLuJe]

@rashkae: Thanks for confirming it actually behaves as I thought it would just from reading the kernel sources ;p.