Re: [Flashcoders] A fast FFT in Flash?
Just a thought - try moving the variable declarations out of the loops: http://www.rozengain.com/blog/2007/05/01/some-actionscript-30-optimizations/ http://osflash.org/as3_speed_optimizations http://gamedevjuice.wordpress.com/2008/02/15/seven-tips-about-performance-optimization-in-actionscript-3/ Might help a little... do you have some test code lying about so we can compare / optimise? Best - karim On 3 Aug 2010, at 04:40, Gerry Beauregard wrote: On 2010-08-03 , at 06:07 , Patrick Matte wrote: Maybe check this http://blog.inspirit.ru/?p=405 Hey Patrick, thanks for the link (to Eugene Zatepyakin's Fast Fourier Transform for Flash). It looks like an interesting implementation, but it uses Alchemy. I've been reading up more on Alchemy, and according to Adobe it's pre-release, and they advise that it not be used to generate code for use in production. In any case, the Alchemy APIs make passing arbitrary data to/from the FFT a bit clunky, e.g. passing a Vector of Numbers requires conversion to a ByteArray first. For what it's worth, I've posted my pure-AS3 FFT implementation here: http://gerrybeauregard.wordpress.com/ ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
Re: [Flashcoders] A fast FFT in Flash?
Hi, You should maybe look using twiddle factors and creating these in the init too - so you are not doing cosine / sin math in the butterfly loops? There are some FFT algorithm optimisations about which init these and just do multiplication / addition in the main loop - you might have to google around for fft twiddle and fft bitrev to find them.. HTH Glen On 03/08/2010 15:10, Karim Beyrouti wrote: Just a thought - try moving the variable declarations out of the loops: http://www.rozengain.com/blog/2007/05/01/some-actionscript-30-optimizations/ http://osflash.org/as3_speed_optimizations http://gamedevjuice.wordpress.com/2008/02/15/seven-tips-about-performance-optimization-in-actionscript-3/ Might help a little... do you have some test code lying about so we can compare / optimise? Best - karim On 3 Aug 2010, at 04:40, Gerry Beauregard wrote: On 2010-08-03 , at 06:07 , Patrick Matte wrote: Maybe check this http://blog.inspirit.ru/?p=405 Hey Patrick, thanks for the link (to Eugene Zatepyakin's Fast Fourier Transform for Flash). It looks like an interesting implementation, but it uses Alchemy. I've been reading up more on Alchemy, and according to Adobe it's pre-release, and they advise that it not be used to generate code for use in production. In any case, the Alchemy APIs make passing arbitrary data to/from the FFT a bit clunky, e.g. passing a Vector of Numbers requires conversion to a ByteArray first. For what it's worth, I've posted my pure-AS3 FFT implementation here: http://gerrybeauregard.wordpress.com/ ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
[Flashcoders] Semi-OT: Alchemy
With Gerry's post last night on his Fast Fourier Transforms question and that Alchemy won't cut it, it makes me curious why Alchemy is still in Adobe Labs. I saw the first preview back at Max 2007 and it's still in beta three years later... wonder if it's some technical hurdles they haven't been able to overcome to make it complete, if Adobe's just not very committed to it, if they are legal issues, a combination of these, or something else. Too bad, it would open up a lot of code libraries to us. If Alchemy is really becoming a dead end for Adobe, it's kind of ironic that alchemy in a historical sense was actually a failed attempt to create precious metals from raw metals (like turning lead into gold). Anyone know anything more about the status of the behind the scenes work on Alchemy? Jason Merrill Instructional Technology Architect Bank of America Global Learning Join the Bank of America Flash Platform Community and visit our Instructional Technology Design Blog (Note: these resources are only available for Bank of America associates) ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
Re: [Flashcoders] A fast FFT in Flash?
On 2010-08-03 , at 22:21 , Glen Pike wrote: You should maybe look using twiddle factors and creating these in the init too - so you are not doing cosine / sin math in the butterfly loops? Thanks for the suggestion. I've tried both ways, and it's actually a tiny bit faster to compute the twiddle factors on the fly, at least compared to pre-computing all of them and storing them in a Vector. (In C++, it's definitely faster to pre-compute the twiddle factors. Slowness is language dependent). In any case, the total number of calls to sin and cos is actually quite small - only one each per stage, the number of stages being log2 of the FFT length (e.g. 10 for a 1024 pt FFT). Successive twiddle factors are computed by complex multiply of a rotated unit vector, which is pretty quick (4 multiplies 2 adds). Turns out the really speed killer is accessing elements of Vectors. In an email a few hours ago, Andre Michelle (author of core stuff for the amazing audiotool.com) gave me an excellent suggestion: use linked lists instead of Vectors. For accessing elements in sequence, it's much faster to follow a link to the next element than to access using a Vector of elements and an incrementing index into that Vector. Using that tip, plus a few other optimizations, I've now got a new FFT that runs at *double* the speed of my previous one. It's nearly 30x as fast as the FFT in as3mathlib. It's still 3.5x slower than my C++ code, but I guess that's the price one pays for using a virtual machine. I'll post the new code in my blog soon... -Gerry On 2010-08-03 , at 22:21 , Glen Pike wrote: Hi, You should maybe look using twiddle factors and creating these in the init too - so you are not doing cosine / sin math in the butterfly loops? There are some FFT algorithm optimisations about which init these and just do multiplication / addition in the main loop - you might have to google around for fft twiddle and fft bitrev to find them.. HTH Glen On 03/08/2010 15:10, Karim Beyrouti wrote: Just a thought - try moving the variable declarations out of the loops: http://www.rozengain.com/blog/2007/05/01/some-actionscript-30-optimizations/ http://osflash.org/as3_speed_optimizations http://gamedevjuice.wordpress.com/2008/02/15/seven-tips-about-performance-optimization-in-actionscript-3/ Might help a little... do you have some test code lying about so we can compare / optimise? Best - karim On 3 Aug 2010, at 04:40, Gerry Beauregard wrote: On 2010-08-03 , at 06:07 , Patrick Matte wrote: Maybe check this http://blog.inspirit.ru/?p=405 Hey Patrick, thanks for the link (to Eugene Zatepyakin's Fast Fourier Transform for Flash). It looks like an interesting implementation, but it uses Alchemy. I've been reading up more on Alchemy, and according to Adobe it's pre-release, and they advise that it not be used to generate code for use in production. In any case, the Alchemy APIs make passing arbitrary data to/from the FFT a bit clunky, e.g. passing a Vector of Numbers requires conversion to a ByteArray first. For what it's worth, I've posted my pure-AS3 FFT implementation here: http://gerrybeauregard.wordpress.com/ ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
[Flashcoders] Using Flash Clients for capacity testing
Hi, Does anyone have any experience of using AS3 apps (be it standalone swfs / AIR apps, or even via browser plugin) to generate load for distributed capacity tests? The advantage of this would is the option to exercise production client code in load-testing / dimensioning. My suspicion is that the flash runtime and network stack is too inefficient to really utilise available hardware compared to a lower level language or framework.. Any ideas greatly appreciated, Tom ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
Re: [Flashcoders] Using Flash Clients for capacity testing
My suspicion is that the flash runtime and network stack is too inefficient to really utilise available hardware compared to a lower level language or framework.. I wouldn't necessarily say inefficient, but there are far easier approaches. The typical load testing approach is to use a single client, and record that client's interaction with the server. Once that's done, the recording becomes the basis of a script that you'd run from within the load testing tool. I believe there's a load testing tool that supports AMF, but if you're not using AMF than you don't even have to worry about that - just use a standard load testing tool that can record all interaction between the client and the server. Dave Watts, CTO, Fig Leaf Software http://www.figleaf.com/ http://training.figleaf.com/ Fig Leaf Software is a Veteran-Owned Small Business (VOSB) on GSA Schedule, and provides the highest caliber vendor-authorized instruction at our training centers, online, or onsite. ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
Re: [Flashcoders] Semi-OT: Alchemy
Hello there, I must say that I was curious too so last week I tried it and I'm really pleased by the results. I have managed to compile Doom in Flash by following the tutorial by Bruce Jawn: http://bruce-lab.blogspot.com/2009/09/build-your-doom-in-flash-step-by-step.html The tutorial is based on the amazing work made by Mike Welsh that made Doom working in the Flash Player. It took me a while to correctly compile the whole stuff, but when it finally worked, I thought it might be interesting to release it as a Facebook app... I modified it a bit more and that's what I did (in only one week in my spare time, so it was not so difficult) *little spam in topic on* http://www.facebook.com/apps/application.php?id=113368622047674 (Application page) http://apps.facebook.com/doominflash/ (jump directly to the game) *little spam in topic off* IlTimido Il 03/08/2010 16:41, Merrill, Jason ha scritto: With Gerry's post last night on his Fast Fourier Transforms question and that Alchemy won't cut it, it makes me curious why Alchemy is still in Adobe Labs. I saw the first preview back at Max 2007 and it's still in beta three years later... wonder if it's some technical hurdles they haven't been able to overcome to make it complete, if Adobe's just not very committed to it, if they are legal issues, a combination of these, or something else. Too bad, it would open up a lot of code libraries to us. If Alchemy is really becoming a dead end for Adobe, it's kind of ironic that alchemy in a historical sense was actually a failed attempt to create precious metals from raw metals (like turning lead into gold). Anyone know anything more about the status of the behind the scenes work on Alchemy? Jason Merrill Instructional Technology Architect Bank of America Global Learning Join the Bank of America Flash Platform Community and visit our Instructional Technology Design Blog (Note: these resources are only available for Bank of America associates) ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
[Flashcoders] flvplayback fullscreen button
hi folks i'm having a really wierd error when i hit the fullscreen button for my flvplayback implementation TypeError: Error #1009: Cannot access a property or method of a null object reference. at fl.video::UIManager/ http://www.adobe.com/2007/flash/flvplayback/internal::enterFullScreenTakeOver() at fl.video::UIManager/ http://www.adobe.com/2007/flash/flvplayback/internal::handleFullScreenEvent() at flash.display::Stage/set displayState() at fl.video::UIManager/enterFullScreenDisplayState() at fl.video::UIManager/ http://www.adobe.com/2007/flash/flvplayback/internal::dispatchMessage() at fl.video::UIManager/ http://www.adobe.com/2007/flash/flvplayback/internal::handleMouseUp() has anyone seen anything like this before? i've tried recreating the conditions that created this error (i have an older version of the application running the fullscreen button just fine - this was before i had added a caption button and reskinned the fullscreen button). I've tried removing the captioning button and reverting to a stock fullscreen button with no success. i add the fullscreen button by waiting for a frame so all of the assets can initialise themselves and then using the following line: _video.fullScreenButton = _mcControls.getChildByName(fullscreenBtn) as MovieClip; i have several other flvplayback objects in the application and if i remove them, the fullscreen button works properly again (although the other instances of flvplayback don't implement any controls of any sort, including a fullscreen button) thanks in advance for your help best a ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
[Flashcoders] Create your own t-shirt
Does anyone know where I might find an AS2 (preferably) or AS3 create your own t-shirt tutorial or any advice on one that is made and for sale? Where the user can upload an image onto a picture of what ever T- shirt they pic and it saves it and sends that pic along with their order. Like on zazzel.com Any help, pointers, links, greatly appreciated. I had no luck on google, but I may not be searching for the right key phrase. Best, Karl DeSaulniers Design Drumm http://designdrumm.com ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
Re: [Flashcoders] Semi-OT: Alchemy
I really wish I could provide a link for this info but... I remember it being said that no further Alchemy development would happen. In other words, it is a dead project. If I can remember where I got the info I will pass it on. On Tue, Aug 3, 2010 at 7:41 AM, Merrill, Jason jason.merr...@bankofamerica.com wrote: With Gerry's post last night on his Fast Fourier Transforms question and that Alchemy won't cut it, it makes me curious why Alchemy is still in Adobe Labs. I saw the first preview back at Max 2007 and it's still in beta three years later... wonder if it's some technical hurdles they haven't been able to overcome to make it complete, if Adobe's just not very committed to it, if they are legal issues, a combination of these, or something else. Too bad, it would open up a lot of code libraries to us. If Alchemy is really becoming a dead end for Adobe, it's kind of ironic that alchemy in a historical sense was actually a failed attempt to create precious metals from raw metals (like turning lead into gold). Anyone know anything more about the status of the behind the scenes work on Alchemy? Jason Merrill Instructional Technology Architect Bank of America Global Learning Join the Bank of America Flash Platform Community and visit our Instructional Technology Design Blog (Note: these resources are only available for Bank of America associates) ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
[Flashcoders] test
test sorry ___ Flashcoders mailing list Flashcoders@chattyfig.figleaf.com http://chattyfig.figleaf.com/mailman/listinfo/flashcoders
