Reply
Thread Tools
Posts: 74 | Thanked: 15 times | Joined on Dec 2009
#1011
Originally Posted by noipv4 View Post
proof of hw acceleration on n900 vidz.
the core libraries for h264 acceleration are located in /lib/dsp

h264vdec_sn.dll64P mp4vdec_sn.dll64P
baseimage.dof h264venc_sn.dll64P ringio.dll64P
cexec.out ipp_sn.dll64P usn.dll64P
dfgm.dll64P jpegenc_sn.dll64P wmv9dec_sn.dll64P
dummy.dll64P m4venc_sn.dll64P
So it seems and it's turned of by default. Switching it on improved performance quite a bit next I'm going to get in to dsp overclocking and see if I can then enjoy any flash video on the internet..
 

The Following 3 Users Say Thank You to keljuk For This Useful Post:
Posts: 122 | Thanked: 121 times | Joined on Mar 2010
#1012
Originally Posted by keljuk View Post
So it seems and it's turned of by default. Switching it on improved performance quite a bit next I'm going to get in to dsp overclocking and see if I can then enjoy any flash video on the internet..
I went to the settings window and hw acceleration was checked already. Are you on PR 1.2?
 
Posts: 515 | Thanked: 266 times | Joined on Nov 2009 @ Oelsted, Denmark
#1013
Same here. HW acceleration on by default (PR1.2).
__________________
Christian Wilken - tux-POWER.dk!
... May the Source be with You ...
 
Posts: 49 | Thanked: 8 times | Joined on May 2010 @ Sweden
#1014
Hardware acceleration that you see in the window with flash 9 has _nothing_ to do with the kind of flash gpu acceleration that flash 10.1 offers.
 

The Following User Says Thank You to poka64 For This Useful Post:
Posts: 3,319 | Thanked: 5,610 times | Joined on Aug 2008 @ Finland
#1015
Also, Flash 10.1 does not automatically mean hardware acceleration, it is just something that can *optionally* be present on certain platforms (like for example the Flash 10.1 beta had disabled h264 video accel on Android).
__________________
Blogging about mobile linux - The Penguin Moves!
Maintainer of PyQt (see introduction and docs), AppWatch, QuickBrownFox, etc
 
devu's Avatar
Posts: 431 | Thanked: 239 times | Joined on Apr 2010 @ London
#1016
Originally Posted by kevloral View Post
The microb .....
Regarding to your post and solution I have to add my 2 cents because made me scary what are you doing here.

This showing me lack of understanding of how flash player works.
And this solution is useless. You can fool the flash player version number this way but it doesn't mean you can play real flash 10.0 or 10.1 content at all. Do you really think the only difference between those players are the version no.??

First of all. I am working on my flaemo project and have some internal code to recognize witch version you running to be able to prevent situation that entire project will use some non compatible code.

With this solution you are for sure crash flash player. in any other project that using this method.

Secondly if you developing some project and compile it in Flash 10.1 and you are NOT using anything flash 10.1 or flash 10.0 specific you will be able to see whole content as is. It doesn't mean your original flash 9 will run faster. (It must be some placebo effect guys )

I will give you example in pseudo code.

if you running flash 9 use
private var somevariable:Array = [];
if you running flash 10.0 use
private var somevariable:Vector<*>= new Vector<*>
if you running flash 10.1 use
private var somevariable:Vector<*>= new Vector<*>
Multitouch.inputMode = MultitouchInputMode.TOUCH_POINT;

if you fool your flash 9 and pretend you running 10.* guess what will happen. Yes, dead flash player. Otherwise Every project that is not running flash player version specific code will handle this task itself.
You don't have to cheat this way.

There is another one situation.

If in html code somebody added minimum version required is 10.0.0 and in fact swf is not using anything from flash 10 can happily run. The only blocker is html tag. There is simple way around this issue just call swf file directly!. [Or you can use Attilla77's Flash Launcher and you can REALLY see performance improvements.]

So once again I DO NOT RECOMMEND this solution because is useless, will not speed up anything and make even more content unable to run if some project like flaemo has encapsulation of methods that specific player can or cannot run.

Check this out.
go there with this player installed
http://www.flaemo.com/ctest/

Pseudo Flash player 10.* will run the code up to meet some flash player version specific code and will stuck on manufacturer, because next feature required real flash 10.0. witch is obviously not present in your fake player. There is at least tens of features for each player version added. speeding up a lot of things, there is big difference between flash drawing API methods between flash 9 and 10 and so on.
Flash developers are happy to use it since we where asking adobe waiting for all this stuff. Don't fool yourself!
__________________
My Toys: N900, Samsung Nexus S
Flaemo - Web based OS for Flash-enabled devices, more info here

Last edited by devu; 2010-08-01 at 14:13.
 

The Following 7 Users Say Thank You to devu For This Useful Post:
Posts: 3,319 | Thanked: 5,610 times | Joined on Aug 2008 @ Finland
#1017
Just a small addition - what is broken is essentially the logic of version determination. Web/flash devs make two mistakes that break the N900 Flash use-case horribly:

a) they abuse the version determination as an upgrade mechanism. The assumption is that if someone does not have the latest version is because they simply have never gotten around upgrading it, so they just need a nudge. These people think they are making users a favor since they will be upgrading to a version which is faster, more safe, etc.

b) a more generic problem, making compatibility choices based on gut feeling, and not actual technical requirements. Often this is also a side effect of html-copy, when people copy over html sections from other sites on on-line tutorials, without matching version info. "It works on my machine, so it’s OK, right ?"

The problem is IMHO actually deeper - I disagree with the general Adobe practice that the *server* decides the content version the *client* is getting. The server should *advertise* available content versions, or the client be able to *request* a particular version, but the final choice should NOT be on the server - it’s the client that knows what it can or cannot do.
__________________
Blogging about mobile linux - The Penguin Moves!
Maintainer of PyQt (see introduction and docs), AppWatch, QuickBrownFox, etc
 

The Following 8 Users Say Thank You to attila77 For This Useful Post:
PradaBrada's Avatar
Posts: 294 | Thanked: 240 times | Joined on May 2010
#1018
Flash 10.1 support is great!

Posted from my HTC Desire
 
Guest | Posts: n/a | Thanked: 0 times | Joined on
#1019
@devu what's really going on is the fact that most of the requirements of these sites to require FP 10,0,0,0 or higher is a false claim in most parts. Not unless you start to use specific parts of publishing to FP 10.x - such as the newer 3D routines and perhaps the newer FLV encoding do you truly require a FP 10.1 plugin.

I remember a site I had done in FP 8 about two years ago and the people redid the website and via SWFObject made the site require FP10 when they hadn't touched the *.swf nor *.flv's at all.

That is what you're seeing. The moment that things come in such as signed/certificate signed material that requires FP 10 via Flash Media Server or via Flex 4 SDK or Flash CS5; that's when this trick will start to fully fail.

But YouTube, Facebook and even Vimeo don't really require FP 10. The adoption rate of FP 10 allows them to exclude prior iterations of the flash player because most folks upgrade due to security concerns or because it's automatically updated for them.

That supports my rant against other Flash developers. They need to deploy to the level that the features they require are needed. And if it's VP6 encoding, then that's just Flash Player 8. If it's the new 3D stuff, then it's FP 10. If it's h.264 encoded video, then FP 9. If it's AS3 using the newer Spark or fx.namespace, then it's 10.1 (via Flex).
 

The Following 5 Users Say Thank You to For This Useful Post:
devu's Avatar
Posts: 431 | Thanked: 239 times | Joined on Apr 2010 @ London
#1020
Attila77, gerbick you both are obviously right, my concern was that fake flash 9 will not pretend 10 or 10.1 and that is just pointless. But regarding to what you said I can only add, all devices that keeping-up version incrementation are ok and is not big pain really from customer point of view if you can see real benefits like performance improvements.

(I was ignorant to security improvements some while ago, but some day I have seen real danger what Flash can do using binary data, on flash 9.0.115 that was fixed or just patched. I am talking here about direct access from swf to any MySQL data base.)

But from developer point of view, if you are lazy enough or just think vector is so cool that you can use it occasionally to store 3 strings in it... instead of use array you shouldn't be called developer. I will join to your rants here gerbick

But that situation only showing us how looks the platform or technology abandoned without support and continuation and not allowance to maintain this by ourselves.

Anyway. From couple of days I'm in deep investigation about cross-compiler possibilities AS3 ->ABC -> SWF - ABC 2 any other stuff even python ->swf, C# -> ABC ->SWF via haxe, LLVM even Action Script based compiler. In other to be able to find solution to write AS3 on n900 on the tube I founded something interesting but I don't want to make off topic here.

I founded something even more interesting during this journey.

Lightspark 0.4.2 released on July 20, 2010.
Open Source Swf Player for linux. maybe instead of moaning about lack of Flash 10.1, Adobe, Nokia and all the this crap, linux guys should invest their free time and power to support that project.

This is first time I see good effort and developers behind this, did this from ground-up. Actually I am surprised (since Tamarin - ActionScript Virtual Machine 2 is open sourced from 3 years and swf specification is also transparent.) that none of Linux devs didn't notice that potential yet.

Lightspark looks very promising... is anybody here able to port this solution to Maemo? Than we can stop all this annoying threads, get rid of corp dependencies and be happy
__________________
My Toys: N900, Samsung Nexus S
Flaemo - Web based OS for Flash-enabled devices, more info here

Last edited by devu; 2010-08-01 at 23:03.
 
Reply

Tags
adobe, adobe flash, blahblah, flash, flash 10, flash 10.1, fremantle, future, idiotic thread, maemo, maemo 5, nokia, nokia fails, update


 
Forum Jump


All times are GMT. The time now is 02:32.