Android Tablets Forum banner
1 - 11 of 11 Posts

·
Registered
Joined
·
338 Posts
Discussion Starter · #1 ·
Tom3q delayed the release of his highly anticipated 3D driver because of the discovery of framebuffer driver being another buggy nutjob of Samsung for their Spica. He also writes on the matter:
It contains a new framebuffer driver indeed, but it's even more limited in functionality than the old one. I've found a way to enable 32-bit RGBA framebuffer in Android with dithering for our 18-bit screen (in fact it's 24-bit, but as far as I know it has only 6 bits per color connected). This should make color gradients look way better. Unfortunately in current driver dithering is disabled by hardcoding the register value and much of the code is written for RGB565 in specific.EDIT:I may use the new driver as a base. It looks pretty good in terms of code quality.
Now, as I can remember from my short stint with FB text/images rendering back in 2002...2004, FB paradigm is kinda sensitive when going from stock 32 bit RGBA to 18 bit color rendering with simple dithering algorithms.Many PDN users (especially readers) complained about pearly grayish, and especially shimmering qualities of white background on our screens. This peculiarity of low-bit dithering becomes quite irritating with prolonged reading sessions.Then, to add injury to insult, the quality of so called "auto-hint" Android feature with its non-existent vertical hinting (try to read "i" from "l" at smaller pointsizes) can be only called as dirty hack compared to what a simplest Linux distro can offer.On top of that there are giveaway DroidSans, DroidSerif Ascender fonts with questionable hinting qualities and overall bland and blind faces. Well, well-hinted Segoe WP for WinPhone 7 corrects much of hinting problems, but unskilled dithering stays. Everyone with a jeweller's 10x, or better 20x loupe can see quite a bit of parasitic color fringing/shifts of a grid (with hinting set to "medium" from the default "light").Has anybody opened his PDN deep enough to know our screen model, to search the color bit specs of it? Is it really 18-bit? Then, sure it's bad, and together with Tomasz's new 3D driver we'd need to bring in his whole new framebuffer driver. Oh boy....Anyone was doing some framebuffer work lately? RGB transforms, namely ditherig schemes?Thank you
 

·
Premium Member
Joined
·
2,216 Posts
I LIVE on this site, and I have seen none of these complaints you're talking about... "Pearly grayish, and especially shimmering qualities of white background"??
 

·
Registered
Joined
·
338 Posts
Discussion Starter · #3 ·
sure I might have mistaken and this response might have appeared in different forum's review. Like Mobleread, or some other where they are mainly concerned about readability of their ebooks. But even if there were no complaints at all, everyone with a PDN might try to put it beside his laptop with fine resolution of 120+ dpi, dim either to equalize white levels and see for himself a smooth white of that laptop and a fine-grained texture on white of PDN with bigger dpi count. Good laptops have good dithered 24 bit, or 32 bit. With 18 bits -- if it's true, and not so good dithering, we have a problem.
 

·
Premium Member
Joined
·
2,338 Posts
I use three dollar reading glasses from the drug store to read the phone book. I know this is your thing aludal, but for someone like me, everything looks fine. I don't want to discourage you from trying to get the most from this device, quite the contrary. But I've got no complaints on readability. I saw a Sony reader at Target yesterday and thought ours blew it away despite its claims to be able to read in direct sunlight. I mean, unless you're immune to skin cancer, how much time should you really spend reading in the sunshine? Personally I like leaning back against a nice tree - or a hammock lying under said tree.
 

·
Registered
Joined
·
1,351 Posts
For what it is worth, this is the PixelFormat I used for the custom logo to make it work properly:
Format32bppRgb Specifies that the format is 32 bits per pixel; 8 bits each are used for the red, green, and blue components. The remaining 8 bits are not used.
So for the logo, it appears that it uses 24-bit color (8 each, as stated above)I suppose you could flash a logo with data only in the upper bits of the pixel and see what it looks like?
 

·
Registered
Joined
·
361 Posts
I notice & have read reviewers describe the almost sparkly white of the PDN screen. I don't know if it's an artifact of the touch screen or what, but I see a barely perceptible sparkle in a solid white screen. Brightness setting may have an effect on this, not sure.Cheers,kev
 

·
Registered
Joined
·
338 Posts
Discussion Starter · #7 ·
Retina it isn't with retina's having 250+ dpi on AMOLED, for sure. However, with its 142 dpi the PDN screen might be milky white without any sparkle/shimmering in white backgrounds (=around text) if only it had proportionally thinner grid. But it's thick for cheapo resistive action, then it's squares instead of stripes. Better anti-aliasing with deeper color depth might improve the situation, but then again, if, as I suspect, we deal with 18-bit screen, this means heavy dithering, hence non-clean colors out of limited palette, plus whites with sparkly artifacts. Grayish-yellowed old paper imitation (as in Laputa's backgrounds) may actually help to eliminate some eye strain when reading ebooks.But, as I said before, I would love to see our dev's trying to play with color conversions, dithering, anti-aliasing, and hinting when building a better FB driver
 

·
Registered
Joined
·
361 Posts
[quote name='Mark Adams;49492]snipI saw a Sony reader at Target yesterday and thought ours blew it away despite its claims to be able to read in direct sunlight. I mean' date=' unless you're immune to skin cancer, how much time should you really spend reading in the sunshine? Personally I like leaning back against a nice tree - or a hammock lying under said tree.[/QUOTE'] Just the other day I was seated in a well lit room with my back to a large window. In this setting, the PDN was just barely readable, only with careful angling of the screen. I pulled out my Sony Touch & read my book. eInk readers are way more legible in bright light & I'm not talking about sitting outside in the sun. No surprise there.
 

·
Registered
Joined
·
23 Posts
If you look at the screen from an angle past the actual viewing area to the black sides, you can see that the sparkles continue past the screen edge. Therefore, I suspect that the inside of touch screen itself is not smooth and creates the sparkles itself. I have the same effect on my Nokia N800, although it is more difficult to tell where the effect is coming from. If the PDN touch screen is in fact the cause, I can't imagine why it is designed that way except perhaps to diffuse the light going in or out???
 

·
Registered
Joined
·
338 Posts
Discussion Starter · #10 ·
Our resistive screen might be built by quite old (if well proven) 4-wire technology:4
-Wire Resistive Touchscreens 4-Wire Resistive touch technology consists of a glass or acrylic panel that is coated with electrically condictive and resistive layers. The thin layers are separated by invisible separator dots. When operating, an electrical current moves through the screen. When pressure is applied to the screen the layers are pressed together, causing a change in the electrical current and a touch event to be registered. 4-Wire Resistive type touch screens are generally the most affordable. Although clarity is less than with other touch screen types, resistive screens are very durable and can be used in a variety of environments. This type of screen is recommended for individual, home, school, or office use, or less demanding point-of-sale systems, restaurant systems, etc.Advantages DisadvantagesHigh touch resolutionPressure sensitive, works with any stylusNot affected by dirt, dust, water, or lightAffordable touchscreen technology75 % clarityResistive layers can be damaged by a sharp objectLess durable then 5-Wire Resistive technology
For our particular implementation, the key words here are "75 % clarity". For whatever reason, nobody could produce much more transparent resistive and conductive wires for the grid in question here. And more durable and precise 5-wire resistive technology doesn't add any of the said "clarity".Sooo, with Brightness at max comparatively much darker non-transparent and relatively thick grid produces unpleasant and eye straining shimmering and sparkling effect.To an extent, it can be diminished by using 50...60 % of max Brightness on our device.But again, it's all OFF-TOPIC: I was asking a simple thing: what colour depth we have. The reason of my curiosity is the forthcoming Tomasz Figa 3D/2D hardware accelerated driver which if taken alone won't change anything in terms of colourspace. However, as it occured, it won't work with dirty hack we all have instead of correct framebuffer driver. Now the implementation of this one depends of color depth of a given screen.Now I myself will go OT a tid bit: 4-wire resistive screens have less than 10 msec reaction, why do we have around 500 msec delay between say, "click" and a orange/amber frame around, say, link. I guess it's how they wrote that haptic driver for Android 1.5, then never updated it.Generally speaking, I yet to see any of our present PDN code components that deserve to be used unchanged in 2.2, without complete re-write.
 
1 - 11 of 11 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top