Database Users

Fix your ATI ImageON woes with this patch!

So, after a day with my AKU3.3 equipped 8525, Platform Builder, IDA Pro and HexCmp, I believe I've finally gotten to the bottom of the ATI ImageON acceleration bug. This is the bug that causes the screen to "tear" during accelerated video playback using CorePlayer, TCPMP or other players that take advantage of the ImageON hardware acceleration. Once this bug is occuring, the entire screen is unreadable until you somehow exit from the video player.
Note that even after this patch is applied there will still remain some "pixelation" artifacts. However, there's a great thread on CoreCodec.com that can be found here. The thread explains how to resolve most of these. Consider that thread "extra credit" though since all in all, this patch alone resolved about 90% of the ATI issues with my AKU2.3 test device (running the South African 1.35 TyTN ROM).
It was a buffer problem you see....
I've attached a cab which modifies the ATI DDI setup file to the AKU3.3 parameters. I've experienced great results with CorePlayer when using this on an AKU2.3 hermes. I only had to check all the boxes in CorePlayer's ATI IMAGEON setup page. By Default, 2 aren't checked:
"Green Tint" bug compensation
Keep ATI driver active (just test)
I believe it is the second parameter that corrects an out of memory crash when a clip is played in full screen mode the second time.
The above settings are workarounds however since the full benefits of this fix will only be available if the setup file is used in conjunction with the latest versions of the ATI drivers. Unfortunately, these drivers must be "baked" into a ROM. Do not try to install these drivers! You will brick your device if you do and a hard reset will be necessary to fix. Install the attached cab file instead and wait for the chefs here to build the new drivers into their next ROM release.
ROM Chefs: you can find the files that make up the release here
Delete the old versions and bake all three files in the zip into the /windows directory.
Now to the Technical.
ATI released 3 files as part of their DDI update: ace_ddi.dll, ahioem.dll and a hidden little configuration file named atihwtbl0.txt. It's this little text file that contains the magic since ace_ddi.dll uses it to configure several settings at startup. I haven't fully analyzed the changes yet and am in the process of doing this now but at first glance they appear to either move or increase the size of several video buffers in memory. It almost looks like there was an overlap issue with audio buffers
Anyway, enjoy this little patch!
As always, while I'm pretty sure that this patch will rock your ATI world, I take no responsibility for any "undocumented features" that may crop up. I've only done minimal testing on the TyTN 1.35 ROM so far. It must be right though... there weren't any syntax errors.
Installation instructions:
Download the attached CAB file
tap to install (it will ask to reset your device)
note: you must install this to your device. The patch won't work if you install it to a storage card.
If you experience side effects, removal is simple. From Settings->System->Remove Programs just remove "Sleuth's ATIFix". Be sure to reset your device after the uninstall finishes.
edit:
It's important to read the TyTN thread located on the CoreCodec forum. Here is the link to that thread. There's a lot of good information concerning the scope of this patch (which takes the form of AKU3.3 experiences) and what still remains to be done. In this thread, schriss does a good job benchmarking the ATI playback and also has some good suggestions, some of which I'm hoping will be implemented in future versions of CorePlayer (such as allowing a YV12 option for the ATI decoder). Also, as the thread points out, DivX decoding using the ImageON remains a challenge (like I said, my patch alone solves about 90% of the issues). Hopefully more will now be able to focus on this once the Hermes DDI setting modifications encompassed by my patch become ubiquitous.

i see you been busy to! thanks man.

Thanks Sleuth,
WMXL v0.30 will incorporate this fix.

Just tried, works perfectly! thanks!
(installed on WMXL.1)

Nice to hear Your post reminds me of something else:
Those of you running WMXL .2 already have the driver portion of the fix baked in. Installing this patch will give you full functionality.

Sleuth255 said:
Those of you running WMXL .2 already have the driver portion of the fix baked in. Installing this patch will give you full functionality.
Click to expand...
Click to collapse
since this patch also appears to be working with the older set of ATI DLL's, what's the "full functionality" what you're talking about? what functions are missing now? I didn't really see any differences playing with TCPMP with AKU3.3 and now with this patch on WM6.......

I experienced intermittent crashes on AKU2.3 when running full screen video if the driver wasn't set to remain active. This problem didn't crop up in my WM6 build that had the new drivers.
The problem appeared to be completely resolved in TCPMP/CorePlayer by simply checking all the boxes in the ATI IMAGEON setup screen however.
However, being a purist, I like to see the config file along with its matching driver running.

Lovely Jubbly! and Sleuth, thank you for all your time and hard work with this
Mike

WMXL extras updated sir. Bloody brilliant work!!

Heimiko said:
since this patch also appears to be working with the older set of ATI DLL's, what's the "full functionality" what you're talking about? what functions are missing now? I didn't really see any differences playing with TCPMP with AKU3.3 and now with this patch on WM6.......
Click to expand...
Click to collapse
Same here but as Sleuth said it was more for fixing crashes in a sense, my cingular device running a wm5 rom without the new drivers now don't crash when using this patch, though ofcourse I still get garbled boxes when playing divx files when I enable acceleration, while x.264 plays perfect when set up right with or without acceleration enabled.
My softbank tytn running xdalive .20 runs good as well as before my device froze at times when using ati video with and without acceleration enabled playing fullscreen. I get a good 20% increase now being able to use it compared to DirectDraw. Thanks for the fix

thanks man...
we were all waiting for this to come

I installed the .cab to the device and checked the two default unchecked boxes on the ATI page in TCPMP.
I opened a divx file and the audio and video were off. I performed a soft reset and everything is now perfect. This is a great find.

Advanced encoders: check out the link to corecodec on post 1. There's a lot there regarding how to optimize encoding for the accelerator (plus a few odd quirks too).

Thanks man. I will test ASAP (now my TyTN is in the hands of my wife that is playing "Ladybugs" game. Nobody in the world is dare enought to even think in ask her for a five minutes check)

Admins do you think this is worthy of a sticky?? Im still trying to root around in the coreplayer forums to find a recommended bitrate/codec etc... to encode our vids into.

I just went through the coreplayer thread again. Perhaps you might want to pm schriss for his opinion on the corecodec forum. At any rate, it looks like DivX is still a problem for the ImageON so I would avoid encoding with codec that fttb . I don't know if its Coreplayer or Hardware though (suspect the former however).

I try to install that CAB on my DOPOD U1000, after reset never can boot up again need to hard reset the device, someone can help to fix this problem?
Thanks!

That's an HTC Athena simdao! This patch is Hermes specific. As you have determined, "Undefined results" can occur if you install this patch on a different device.
Slightly OT here, but does your Athena suffer from ATI based DivX rendering issues? What other ATI ImageON issues do you have with it?

Sleuth255 said:
I just went through the coreplayer thread again. Perhaps you might want to pm schriss for his opinion on the corecodec forum. At any rate, it looks like DivX is still a problem for the ImageON so I would avoid encoding with codec that fttb . I don't know if its Coreplayer or Hardware though (suspect the former however).
Click to expand...
Click to collapse
The videos i have encoded using AutoGK, its good because i can just queu up the videos to encode and leave it running. Depending on the program i am encoding i set the filesize accordingly:
60mb for a cartoony show (family guy, futurama, simpsons etc...) approximatly 25 mins (so 2.4mb per min?)
120mb for a filmed program (friends, mythbusters etc...) again approximatly 25 mins (so 4.8mb per min?)
128kbps VBR MP3 for audio (fine for all movies)
fixed width of 320 (the program adjusts for the aspec ratio of the source)
XVid Mpeg4 (2 pass)
I used this encoding method since having success with TCPMP on one of the early iMate roms, since the South Africa HTC rom arrived all the roms since have played back on TCPMP/Coreplayer no problem with the rawframebuffer setting. Since the AKU3.3 test rom i have been using ATi Imageon setting instead and get MINIMAL artifacts onscreen with my files.
Worth a go???

Following mrvanx advice on this and another thread...
I used AutoGK and set parameters as follows:
fixed width 320
Predefined filesize 400mb
128VBR mp3
XviD Mpeg4 (2pass)
This was done on a DVD quality divx file, and produced a 400mb file which played back near flawlessly on the hermes, minimal artifacting, and much better quality than even my ipod! Great stuff
Thanks,
J

PocketCM Imageviewer test opengl build

Hi,
I usually post in the "global dev" and not a device specific forum, but as I recently got a Diamond, I've been experimenting a bit with OpenGL.
For the curious, you can download an opengl enabled build of ImageViewer using HW opengl of our beloved device.
You can find it here:
www.pocketcm.com/daily.php
Of course it require real accelerated driver (which come with the device by default if I'm right), and it's still experimental, it seems to works decently until a certain number of images (texture) are loaded.
In practice, on my device, it's A LOT faster than the software version and it makes it as nice to use as... on my old HTC Touch phone.
Enjoy!

Cool program for an experimental version.... i like the way it rotates the image

just wanna let you know GREAT APP .... im sure ur putting in a lotta features...i was wondering if u can make a coverflo activated on the rotation of the phone...it'll be uber cool...thanx

nice idea..
tried it on my XDA Flame which has Goforce 5500 GPU and it does not run!
log file says "cant initialize opengl"
(of course we do have opengl drivers official on Flame)
please fix !!

works on my TP, but often gets stuck...

Kinoma Play on TG01 (touchflo3D killer)

hello all
kinoma Play looks really impressive in this video. http://www.wmexperts.com/next-release-kinoma-play-features-its-own-home-screens
are there any toshiba tg01 rom dumps floating around?
i hope people here are interested in porting this thing to our ''htc devices

This is just the next update ok Kinoma Play, which can be dowloaded on ur HTC.
Just wait few weeks for the update.
It's a good soft indeed.

wow this is very good!

the thing feels like its running on a 3GHZ machine with a geforce 8800
too damn impressive on the tg01

As this software comes bundle with TG01, Toshiba might shared the API for hardware acceleration to Kinoma Team. Remember this is a entire new processor platform (named Snapdragon). This software would never perform like this in ours old MSX7200 processors.

truly impressive
twice as smooth as the iphone with 2.5x the number of pixels so that makes it 5x as immpressive
its hard to tell where the constant 60FPS is comming from. is it from the cool coding tricks or from the CPU specs of the TG01. looks like both

I just love the Toshiba TG01
And Kinoma doesn't seem to be so bad either
Sascha

greatttttttt

I've tried this program, but there are a few hangups that really keep me from using it.
The first one is that it will only "intelligently" scan your device and storage card. There is no option to setup what folders to scan or leave out of a scan. They're only response as a "fix" to this is to put any directory you don't want scanned in the Program Files folder on your device/card. This means I have to redo my entire storage structure to work around their smart scan.
Also, they don't support .avi video files. They say "It's not used for legitimate content distribution" and that "it was always a poorly-defined format with many undocumented mutations that make it a mess to implement and support".
Well, one can argue the legitimacy of certain files, but that's not up to the video player to decide IMO. I have purchased several legitimate programs that convert my DVDs into avi files for mobile playback.
And as far as a mess to implement and support, hell, my very first Casio Cassiopeia supported avi playback with TCPMP. So that's just a cop out as far as I'm concerned.
Yes, the program looks great and has many awesome features, but it just falls short in trying to think too much for me. If I wanted to be a "dumb" user and have a "smart" media player, I'd buy an iPhone.

the new version was just released maybe these issues were sorted out. how about you upgrade and let us know

zarxic said:
The first one is that it will only "intelligently" scan your device and storage card. There is no option to setup what folders to scan or leave out of a scan. They're only response as a "fix" to this is to put any directory you don't want scanned in the Program Files folder on your device/card. This means I have to redo my entire storage structure to work around their smart scan.
Click to expand...
Click to collapse
New version has an advanced option add on that allows you select folders to be scanned. I haven't messed with it as my battery is dead but it sounds like it would help you.

how do i activate the free version of kinoma im getting invalid email man.. how do i do this guys?

Found this in another thread.
fesar said:
Hi,
Please find the link. I have uploaded the DUMP here. Please tell me how to proceed further and changing this DUMP from French to English.
http://rapidshare.com/files/253526287/Utils.rar
Regards,
Francis.
Click to expand...
Click to collapse

hey guys was just wondering what is the latest version of Kinoma Play thats out right now?

Render-To-Texture with HTC Diamond 2

Hi everyone,
I have a small technical problem to solve and hopefully someone knows an answer for it. Someone gave me a HTC Diamond 2 (very nice toy I think). I don’t know which hardware is exactly inside the phone but I assume it’s an ATI ImageOn 2300 which fully supports OpenGL ES 1.0 + Extension Pack (that’s what GL_EXTENSION is saying). And all of these extensions are doing fine. As everyone knows HTC is not providing a D3D driver for their phones and all other drivers I have seen are just a wrapper around OpenGL ES. So using D3D is currently not an option.
I have to write an application which requires some render-to-texture functionality in realtime and here my problems are starting to grow.
[1] glCopyTexImage2D:
I recognized that glCopyTexImage2D() is very slow. My framerate is dropping dramatically from >100 frames down to ~20 frames. Somehow this is done in software by the driver and not hardware- accelerated. glCopyTexSubImage2D() is even more slow (down to ~10 frames per second), but both functions are working. I tried to move it into a different thread, but the driver is not supporting shared contexts. Also it’s not supporting two bound contexts at the same time with two different windows and threads. So this can’t be improved or I’m doing something wrong.
[2]PBuffers:
PBuffer are working fine. But PBuffers which can be bound to a texture are not supported by OpenGL ES 1.0. And that’s what happens also on this ATI card.
[3]GL_OES_framebuffer_object:
This was my preferred choice. But framebuffers are not officially supported by the installed OpenGL driver. But the libgles_cm.dll on the HTC is exporting these functions so I tried them and recognized that they are not working correctly. Somehow the vertex pipeline is allowing only triangles in the center of the viewport. All others will be discarded. When I turn it off it renders correct. Using glDrawTexiOES() will be ignored while using framebuffers. I assume that either the current implementation of the framebuffers is just waste of some development guys or the functionality is locked in some way.
[4]D3D:
Normally D3DM is supporting to switch between different rendertargets. But we are all know the D3D problem. I tested some other D3D drivers but they are not usable. Also I can’t imagine how render-to-texture will be implemented in these drivers while it’s not working in OpenGL.
I’m wondering what TouchFlow3D is using internally, whether they are using render-to-texture or not. The only thing I know is that they are using OpenGL ES and some extensions. But Manila.exe is querying the functions of libgles_cm.dll during runtime. So I have to write a few proxy dlls first and need to hook into the system to track what they are doing. And I don’t want to spend time on this.
Does someone knows an alternative to do some render-to-texture on the ATI or knows some secrets of the libgles_cm.dll which I don’t know? There are a lot of private functions inside but can’t find some documentation about it. Also ATI and Qualcomm are not very helpful to me.
Thanks.

Maybe, glReadPixels and then glTexSubImage2D (what probably glCopyTexImage2D is doing)?
(I know that this goes two times over the graphics bus, but you never know...)
What kind of scene (number of triangles, lights, textures) are you rendering with > 100fps?

glReadPixels() + glTexImage2D() is even more slow than glCopyTexImage2D...
Also this can't be parallized. I thought using AHI2DATI.dll instead to do the same thing, but I don't know how to get a surface handle from a OpenGL texture id.
>>>What kind of scene?
A very simple scene yet. Only a few depth sorted + material sorted objects (via VBO) with some textures (backed lighting) on it. Textures are compressed. Currently no lighting, no skinning or other things. While the render thread is waiting for glFinish() to return, a second thread prepares the next frame. Also the rendering thread is not redrawing the entire viewport each time.

The OpenGL texture id is the handle. There's nothing more you can do with it. PBuffers or framebuffer objects are the only way I know for doing performant render to texture in OpenGL.

No...I mean the surface handle of the AHI2DATI library. Here you have access to the raw data of the surface. Somehow the libgles_cm.dll uses these surfaces for it own buffers and/or textures or not?!
But this is not really useful unless someone tells us whats going on inside libgles_cm.dll.
See:http://greengalaxy.wordpress.com/2009/04/18/ati-direct-access-to-hardware/

Hi jeansmsixer, I don't think there is no efficient way to do it. Have you tried eglCopyBuffers?
Aren't all textures stored in system memory? - they surely have to be because the device reports no available video mem.
Even if you get a pointer to the color buffer, it's impossible to wrap the memory in a HBITMAP to select onto a HDC and use GDI fonts for instance. (which is what I need to do).
If you do write the proxy dll like you were suggesting, can you please let me know what TF3D does for fonts? Are they textures or have they somehow mixed 3D with GDI?

Hi jeansmsixer, something else you could try is what is described here:
http://brewforums.qualcomm.com/showthread.php?t=10668
Looking on glbenchmark.com it looks like HTC devices support the extension. You should be able to get a pointer directly to the color buffer (which I presume is in normal system memory) so you could copy off pixels fairly efficiently with your own memcpy(). However, you will need to eglWaitGL() etc to ensure 3d stuff is complete before attempting to access it.
I love the irony of the only helpful information for WM opengl being found on a brew site for symbian. If WM7 is as terrible as the current mess, then I'm moving to iPhone.

[Q] Touch HD Android Adobe Flash support? Info on how Android on SD works?

Hi everyone,
Just returning to the threads here to see how/if things have moved on. I've been running XDAndroidFRX05 and have just tried XDAndroidFRX06, I've also tried the FroyoXGinger build too from a while back.
My question is this, is there any build out there which supports Adobe flash which can run on the Blackstone? Or are there any other builds worth experimenting with (I'm currently having a go at CyanogenMod to see if that runs (ok CM7 not much works yet for Blackstone), now trying the XDAndroid GBX0A)?
I'll do my best to search for some more detail about Android on SD as I am well aware I don't have a clue how all the files interact so no doubt it will help to gain an understanding of it. If you know of a good guide then I would welcome a genital push in the right direction. [EDIT - I can't find anywhere which gives an explanation of what each file is and how it all fits together - I can probably guess but after some definite information if possible]
The NAND project is encouraging and hopefully it will allow android to run smoothly on our devices at some point, however for me the SDCard solution is much preferred as a dead/partly-working phone is no good for me.

I don't think Blackstone will ever run Flash because of the hardware limitations.

jeanfrederic said:
I don't think Blackstone will ever run Flash because of the hardware limitations.
Click to expand...
Click to collapse
I know full flash was apparently limited to CortexA8 processors and up, but was unsure of the lite version.
Also SkyFire and uZard browsers add 3rd party support for videos from within the browsers. I guess I expected some version of flash lite would be possible, afterall hardware limits are often ignored on here.
It seems like it could be doable but I don't expect it to work that well and I guess the HD is too old for all that effort now.
Thanks for your answer though.

Dunno about Android, but Flash Lite is working on my Blackstone.
I sometimes use FlashLIGHT (http://www.freewarepocketpc.net/ppc-download-flashlight.html ) application...

Bodisson said:
Dunno about Android, but Flash Lite is working on my Blackstone.
I sometimes use FlashLIGHT (http://www.freewarepocketpc.net/ppc-download-flashlight.html ) application...
Click to expand...
Click to collapse
Of course then you can always resort to your FleshLight afterwards.
As for hardware, I've seen other Arm11 devices claim Flash 10.1 support, but perhaps that is due to extra decoder chips maybe.

i don't know about uzard, but skyfire renders the pages on a central server and just sends you a "screenshot" of the page. that is why skyfire is the full browsing experience including flash, java, html5, ajax and whatever else you can find. actually,when i last tested skyfire on winmo, by going to ip and browser checking sites, i found, that i was running mozilla firefox with an american ip, which also allowed hulu. /OT
as stated before, i wouldn't count on the blackstone being capable of handling the full flash experience, but flash light might work, or, depending on what you actually need flash for, there might be another way around it.