Guest Daedric1383 Posted March 24, 2012 Report Share Posted March 24, 2012 Can't we force the ppp daemon? A few years ago, we could connect data on a HTC Kaiser using a app called modem. After a while the problem with data connection was fixed and 'modem' became obsolete... i'll try and find the apk... Link to comment Share on other sites More sharing options...
Guest Scanno Posted March 24, 2012 Report Share Posted March 24, 2012 Eduardo, I am using a huawei usb dongle, that worked on vegacomb. Seems it is on being detected. Should i reconfigure the us b port? Verstuurd van mijn HTC Vision met Tapatalk Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 25, 2012 Report Share Posted March 25, 2012 (edited) I have that same device in my PoV Mobii and GPS worked great in VegaComb after adding a few patches by ejtagle. I guess he will be able to make it work in VegaICS also if he has the time...I think the GPS already is activated, it only needs to be configured correctly. Care to point me a link? :D OFFTOPIC, why so heavy banding ? It must be kernel related... Edited March 25, 2012 by Daedric1383 Link to comment Share on other sites More sharing options...
Guest pischla Posted March 25, 2012 Report Share Posted March 25, 2012 Yes, but it is not happening. The Android framework , apparently properly configures the APN, and then, it should call REQUEST_SETUP_DATA_CALL in the RIL library, which in turn inits the data call and starts the ppp daemon that finally brings up the ppp0 network interface. But the REQUEST_SETUP_DATA_CALL is never done. I think the problem is that the Android framework compilation done by Cass does not have Mobile Data roaming as a valid network access source... So, even if the 3G modem is detected, the framework does not use it. The most suspicious thing is that the signal 3G/cell strenght indicator is turned off when you turn the Wifi off... And i hve seen that the AOSP has some config files that must be modified before compiling to let the connectionManager use 3G as a valid connection .... If you look at Honeycomb logs, you will see the REQUEST_SETUP_DATA_CALL being made... Perhaps, there is a way to change the connectionManager configuration without recompiling the whole framework... I see there are lot of posts on the internet regarding roaming and one in particular that's called "National Data Roaming". private static final String BUTTON_DATA_ENABLED_KEY = "button_data_enabled_key"; private static final String BUTTON_DATA_USAGE_KEY = "button_data_usage_key"; private static final String BUTTON_PREFERED_NETWORK_MODE = "preferred_network_mode_key"; private static final String BUTTON_ROAMING_KEY = "button_roaming_key"; private static final String BUTTON_NATIONAL_ROAMING_KEY = "button_national_roaming_key"; private static final String BUTTON_CDMA_LTE_DATA_SERVICE_KEY = "cdma_lte_data_service_key"; from this file: https://github.com/CyanogenMod/android_packages_apps_Phone/blob/ics/src/com/android/phone/Settings.java Care to point me a link? :D OFFTOPIC, why so heavy banding ? It must be kernel related... Link to what? Where do you see banding? Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 25, 2012 Report Share Posted March 25, 2012 (edited) Link to what? Where do you see banding? Link to this. I have that same device in my PoV Mobii and GPS worked great in VegaComb after adding a few patches by ejtagle. I guess he will be able to make it work in VegaICS also if he has the time...I think the GPS already is activated, it only needs to be configured correctly. Banding ? Everywhere... let me get a screenshot. UPDATE: screenshot is pointless, the banding and heavy dithering show only on screen. I tried taking a picture, but although i see it, it can be considered moire, or jpeg artifacts. To try and see, on the lockscreen, with brightness on maximum, and the screen paralel to your eyes and portrait, look at the top. You'll see the dithering :) Edited March 26, 2012 by Daedric1383 Link to comment Share on other sites More sharing options...
Guest pischla Posted March 26, 2012 Report Share Posted March 26, 2012 Link to this. Ok, It should be in this little thread somewhere: :unsure: http://www.tabletrom...update-3-a.html Banding ? Everywhere... let me get a screenshot. UPDATE: screenshot is pointless, the banding and heavy dithering show only on screen. I tried taking a picture, but although i see it, it can be considered moire, or jpeg artifacts. To try and see, on the lockscreen, with brightness on maximum, and the screen paralel to your eyes and portrait, look at the top. You'll see the dithering :) I can not see any banding on my lockscreen! Did you see the banding right after installing VegaICS beta 1? Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 26, 2012 Report Share Posted March 26, 2012 (edited) Ok, It should be in this little thread somewhere: :unsure: http://www.tabletrom...update-3-a.html I can not see any banding on my lockscreen! Did you see the banding right after installing VegaICS beta 1? Yes, right on first boot. Banding is hard to find, the dithering though, is everywhere (the banding i'm refering might be caused by bad dithering) Just for the reference: Left is as it should be, and how we see it in a distance. But up close, you see the patters in the middle dithering. Oh... better: Dithering Algorithms , look at the Bayer method. Edited March 26, 2012 by Daedric1383 Link to comment Share on other sites More sharing options...
Guest pischla Posted March 26, 2012 Report Share Posted March 26, 2012 (edited) This is what my screen looks like when I look at it: (looks the same on tablet as on screenshot) EDIT: I see what you mean now after I changed my wallpaper! Edited March 26, 2012 by pischla Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 26, 2012 Report Share Posted March 26, 2012 This is what my screen looks like when I look at it: (looks the same on tablet as on screenshot) EDIT: I see what you mean now after I changed my wallpaper! Is it my eyes ? Is it because of the huge screen / relativley low dpi ? I just feel the screen is cheating us, and it's between the framebuffer and the screen, as all screenshots show no dithering at all. ( i belive screenshots grab data from the fb, correct me if i'm wrong :D) Link to comment Share on other sites More sharing options...
Guest ejtagle Posted March 26, 2012 Report Share Posted March 26, 2012 Is it my eyes ? Is it because of the huge screen / relativley low dpi ? I just feel the screen is cheating us, and it's between the framebuffer and the screen, as all screenshots show no dithering at all. ( i belive screenshots grab data from the fb, correct me if i'm wrong :D) Framebuffer is in 24bit colour... But the LCD only accepts 18bit colour. The tegra2 hw does the dithering... It can be chosen between error diffusion and ordered dithering... Sorry, it is a limitation of the LCD screen used... Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 26, 2012 Report Share Posted March 26, 2012 (edited) Framebuffer is in 24bit colour... But the LCD only accepts 18bit colour. The tegra2 hw does the dithering... It can be chosen between error diffusion and ordered dithering... Sorry, it is a limitation of the LCD screen used... And we're using ordered dithering. Can it be changed to error diffusion ? the lack of regular pattern will probably make it less noticable, i think. Just 'food for thought', on the reference image, left is perfect, middle is ordered, right is diffusion. On the wikipedia article, check the Bayer vs the several error diffusion, Floyd–Steinberg,Jarvis, Judice & Ninke,Stucki orBurkes Edited March 26, 2012 by Daedric1383 Link to comment Share on other sites More sharing options...
Guest ejtagle Posted March 26, 2012 Report Share Posted March 26, 2012 And we're using ordered dithering. Can it be changed to error diffusion ? the lack of regular pattern will probably make it less noticable, i think. Just 'food for thought', on the reference image, left is perfect, middle is ordered, right is diffusion. On the wikipedia article, check the Bayer vs the several error diffusion, Floyd–Steinberg,Jarvis, Judice & Ninke,Stucki orBurkes The latest patchset i posted against the rel14r7 linux kernel is using error diffusion, unless someone has changed it. You can check that in board-shuttle-gpu,c, .dither = TEGRA_DC_ERRDIFF_DITHER ... unless there is a bug somewhere in either the tegra hw or the tegra driver... Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 27, 2012 Report Share Posted March 27, 2012 The latest patchset i posted against the rel14r7 linux kernel is using error diffusion, unless someone has changed it. You can check that in board-shuttle-gpu,c, .dither = TEGRA_DC_ERRDIFF_DITHER ... unless there is a bug somewhere in either the tegra hw or the tegra driver... This may sound extremly dumb: where are the sources ?? the github on the OP is 404'ing... Link to comment Share on other sites More sharing options...
Guest brucelee666 Posted March 27, 2012 Report Share Posted March 27, 2012 (edited) This may sound extremly dumb: where are the sources ?? the github on the OP is 404'ing... Posted only a few pages back, try this post from me (page 77) which points to nvidia kernel (rel-14r7) and Eduardos patch required or goto wooshy1 git here which I think he updated a few days ago with latest files. For info my boot image posted earlier also had error diffusion set in kernel. Edited March 27, 2012 by brucelee666 Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 27, 2012 Report Share Posted March 27, 2012 (edited) The latest patchset i posted against the rel14r7 linux kernel is using error diffusion, unless someone has changed it. You can check that in board-shuttle-gpu,c, .dither = TEGRA_DC_ERRDIFF_DITHER ... unless there is a bug somewhere in either the tegra hw or the tegra driver... This? static struct tegra_dc_out shuttle_disp1_out = { .type = TEGRA_DC_OUT_RGB, .align = TEGRA_DC_ALIGN_MSB, .order = TEGRA_DC_ORDER_RED_BLUE, .depth = 18, /* Enable dithering. Tegra supports error diffusion when the active region is less than 1280 pixels wide - Otherwise, ordered dither is all we will have. */ /*.dither = TEGRA_DC_ORDERED_DITHER,*/ .dither = TEGRA_DC_ERRDIFF_DITHER, If so, either the kernel in VegaICS B1 is not compiled against this source, or this .dither is being ignored/bypassed. Still, thank you ejtagle and brucelee666. I know i'm being pesky, and that there's alot to solve before this eyecandy issue, but my amateur photographer eye saw it and claimed "HIGH ISO NOISE!!! Wait.. what?" Edited March 27, 2012 by Daedric1383 Link to comment Share on other sites More sharing options...
Guest brucelee666 Posted March 27, 2012 Report Share Posted March 27, 2012 (edited) Daedric1383 Attached a boot image with kernel compiled with dither set to ORDERED to see if it makes any difference to your amateur photographer eye. Saw the adam devs experimented with changing back to ordered from errdiff in an attempt to fix banding (don't know result) and it may be that libs default to ordered if unable to do errdiff depending on active region pixel width so will not make any difference. Also not sure if dithering gives different results based on density for example 160=1024x600actual but 120=1365x752virtual so try at different density values to see if you get different results. edit 28/3 - uploaded new boot image forgot to add Eduardos 3g mods to ramdisk on first uploadboot.img Edited March 28, 2012 by brucelee666 Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 27, 2012 Report Share Posted March 27, 2012 (edited) Daedric1383 Attached a boot image with kernel compiled with dither set to ORDERED to see if it makes any difference to your amateur photographer eye. Saw the adam devs experimented with changing back to ordered from errdiff in an attempt to fix banding (don't know result) and it may be that libs default to ordered if unable to do errdiff depending on active region pixel width so will not make any difference. Also not sure if dithering gives different results based on density for example 160=1024x600actual but 120=1365x752virtual so try at different density values to see if you get different results. It should not make a diference. Dither should be applied to the final imagem when unable to render the requested color (by the screen). Meaning, dithering should only happen has the last thing before the image is "downsampled" from 24bits to 18bits before being sent to the screen. What bothers me, is that my HD2 has a 16bit screen (it has a 24bit screen but the cable connecting the screen only supports 16bits). So the Adreno200 drivers do a dithering themselves, but it's... almost perfect. I cannot see the dither, i cannot see the banding... marc1706 / cm-on-wince-htc Can it be the RGB format ? The HD2 is using RGB565 (16bits).. I'll give it a shot as soon is i get home :) Thank you brucelee666 Edited March 27, 2012 by Daedric1383 Link to comment Share on other sites More sharing options...
Guest pischla Posted March 27, 2012 Report Share Posted March 27, 2012 Somehow I now get weirdness after flashing VegaComb back onto my tablet regarding Mobile data... http://pastebin.com/ZVAXfzh5 Anybody have an idea of what's going on? Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 27, 2012 Report Share Posted March 27, 2012 Somehow I now get weirdness after flashing VegaComb back onto my tablet regarding Mobile data... http://pastebin.com/ZVAXfzh5 Anybody have an idea of what's going on? I had this before, a power off/on cycle solved it, as long as the sim card was already on the slot when powering on. Link to comment Share on other sites More sharing options...
Guest pischla Posted March 27, 2012 Report Share Posted March 27, 2012 Does not help powering of and on. Also tried to flash again to no avail...wonder if I have messed up some setting trying to get the modem to work in VegaICS? Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 27, 2012 Report Share Posted March 27, 2012 Does not help powering off and on. Also tried to flash again to no avail...wonder if I have messed up some setting trying to get the modem to work in VegaICS? EDIT: Oops double post... Would you be willing to try Honey Ice Chimera ? Just to debug.... Link to comment Share on other sites More sharing options...
Guest pischla Posted March 28, 2012 Report Share Posted March 28, 2012 I re-flashed VegaICS again and now it seems to respond to AT commands again, phew! Wonder what I did differently this time when flashing VegaCOMB? I'll try some more now that I know the modem is not broken atleast! Link to comment Share on other sites More sharing options...
Guest MikhailM Posted March 28, 2012 Report Share Posted March 28, 2012 Would you be willing to try Honey Ice Chimera ? Just to debug.... I might not have understood your meaning/intention, but I can confirm that 3G is working well under HoneyIce Chimera Edition Updated. Link to comment Share on other sites More sharing options...
Guest Daedric1383 Posted March 28, 2012 Report Share Posted March 28, 2012 I might not have understood your meaning/intention, but I can confirm that 3G is working well under HoneyIce Chimera Edition Updated. 3G works well in Chimera, so maybe RIL/libraries or similar are sending better AT commands to the 3G module forcing it to reset/restart. Just my idea :D Link to comment Share on other sites More sharing options...
Guest ejtagle Posted March 29, 2012 Report Share Posted March 29, 2012 3G works well in Chimera, so maybe RIL/libraries or similar are sending better AT commands to the 3G module forcing it to reset/restart. Just my idea :D You could replace Chimera RIL with the one we are using ;) ... As far as i know, the problem right now is the framework... And to explain it more cleanly, the problem is not a bug in the framework, it is a misconfiguration of the framework when i wrote the initial support for a Shuttle target for ICS AOSP: When doing so, you must tell the AOSP build system the limitations of the available connections. I don't exactly know why, but seems the 3G is being disabled and enabled at the same time as wifi, making impossible to use 3G while wifi is disabled, and when wifi is enabled, 3G is also enabled, but the framework always prefers wifi over 3G. The fix is to recompile AOSP with the right configuration. I am waiting for more RAM for my computer to be able to successfully recompile it. Hope in a few days i will get it recompiled and hopefully working... Eduardo Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now