Augmented Reality in Flash
Link. This is definatly the coolest thing I've seen on flash for a very long time. 3D feedback from a camera's input. Hard to explain so just watch the first minute of the video. I'm just about to test it out.
Shit, i didnt even notice the date!
skins a fattie
At 4/20/09 03:42 PM, Glaiel-Gamer wrote: murder is bad
Explain this further.
At 4/20/09 04:43 PM, Depredation wrote:At 4/20/09 03:42 PM, Glaiel-Gamer wrote: murder is badExplain this further.
i don't get it...
At 4/20/09 04:13 PM, trig1 wrote: Augmented Reality in Flash
Link. This is definatly the coolest thing I've seen on flash for a very long time. 3D feedback from a camera's input. Hard to explain so just watch the first minute of the video. I'm just about to test it out.
That is like amazingly awesome :P, im trying to figure a way on how to put that into a game somehow XD
.
4/20
It's 4/20 everyday in Holland. I don't see anything special or cool about it... I find the date rather macabre as well...
Let's all celebrate a pointless illegal "holiday" for no real event changing reason (besides the fact that you're all smoking an extra 2 grams today), the Columbine anniversary, and Hitlers birthday.
what a great day.
At 4/20/09 07:05 PM, zrb wrote: Finally !
Go play my game.
Its not amazing due to laziness but meh.
Instead can you just play it for me?
At 4/20/09 09:25 PM, MichaelHurst wrote: Haha happy 420 guys!
My german teacher acted like Hitler today... It wasn't intentional, but it was it hilarious.
Hey yo
At 4/20/09 04:13 PM, trig1 wrote: Augmented Reality in Flash
Horray for a legal topic conversation. That looks really awesome. I only watched the first two minutes, still, it looks awesome!
hey guys im sorta struggling to figure out which style i should use.
the one on the left is a lot more tedious and takes up more time but looks nice (zoom into 600% then draw with size three brush), while the one on the right doesn't look as nice but is a lot (and i mean a lot) quicker then the first style.
Sig made by me
Once again i'm falling down a mountain like a metaphor
Here ends another post by the grand master of all things: fluffkomix
At 4/20/09 10:07 PM, fluffkomix wrote: the one on the left is a lot more tedious and takes up more time but looks nice (zoom into 600% then draw with size three brush), while the one on the right doesn't look as nice but is a lot (and i mean a lot) quicker then the first style.
From the way you've animated it, it looks pretty much the same. I'd recommend using the more refined technique for scenes when the characters aren't moving a whole lot and use rougher lines when theres more action.
:U
At 4/20/09 10:33 PM, Josh-B wrote:At 4/20/09 10:07 PM, fluffkomix wrote: the one on the left is a lot more tedious and takes up more time but looks nice (zoom into 600% then draw with size three brush), while the one on the right doesn't look as nice but is a lot (and i mean a lot) quicker then the first style.From the way you've animated it, it looks pretty much the same. I'd recommend using the more refined technique for scenes when the characters aren't moving a whole lot and use rougher lines when theres more action.
thats exactly what i was thinking!
and thats exactly how eddsworld does it but whatever. i dont care, its easier.
Sig made by me
Once again i'm falling down a mountain like a metaphor
Here ends another post by the grand master of all things: fluffkomix
Happy 4/20 one and all.
From three dudes from Duder Entertainment.
Twas an awesome day and an amazing 4/20 rally at the art gallery in Vancouver. :D
screenshots from stuff most of you will never get to use
http://spamtheweb.com/ul/upload/200409/8 0120_Picture_20.png
At 4/20/09 03:39 PM, KaynSlamdyke wrote: On the next few pages can we have something discussed that isn't an illegal activity?
It's annoying we are asked to change the subject of conversation every time we hit on a subject that's even less than slightly controversial. I think that discussions between civilised and intelligent people can only make improvement. The same thing happens when we discuss religion or politics.
It's be censorships that the true problems are caused.
I know piracy is supposedly illegal in some countries (although no one really gets caught, do they?), but i bet almost everyone in the reg lounge pirated at least once or twice in their life, whether it be by downloading music, flash, tv shows, or games - so it's not like stopping the piracy discussions will make people stop the piracy. On the countrary. If you are against piracy, join the discussion and explain why you think it's wrong. THEN you can stop people from pirating.
I was more reffering to 420 than Piracy. The discussion regarding the theft of digital media was interesting and most of the points raised were thought out and intelligent from all sides. But while I don't have a problem with recreational drug use, I do have a problem with thinking counterculture needs to designate a day for it.
...
At 4/21/09 05:29 AM, KaynSlamdyke wrote: I was more reffering to 420 than Piracy. The discussion regarding the theft of digital media was interesting and most of the points raised were thought out and intelligent from all sides. But while I don't have a problem with recreational drug use, I do have a problem with thinking counterculture needs to designate a day for it.
Although since the topic's been raised, I've been addicted to Ben Goldacre recently and quite a while ago he brought up some pretty good points about how bollocks recent cannabis statistics are, which I've only just dug up :p I like the one about how if all these scare stories are true then the stuff should be around 350 times stronger than it was in 1970, providing more THC in a single plant than the volume of the plant itself.
This is more of an effort to bring up Ben Goldacre then cannabis. He rocks!
Ben is a great writer, but thanks to him and the other Guardian columnist I read, I have trust issues with the entire media in general.
The former appeared on the latter's TV show last week, talking about his specialist subject - the MMR vaccine media shitstorm. Well worth a watch if you're in the UK
Charlier Brooker's Newswipe Episode 4
For you Americans, all I can say is "this is what you get for us not being able to watch the Colbert Report online"...
The eternally amusing antics of Kayn trying to build that sodding perspective tile engine
Fourth Incarnation
Click the tiles to make them vanish and reappear. Got to redo the scrolling...
...
OMFG
OMFG
OMFG
OMFG
WTF AS?!
Using Flex 3 Actionscript Profiler, the following results are evident (And fucking stupid!)
function f1(a:int):int
{
return a*10;
}
function f2(a:*):*
{
return a*10;
}
function f3(a)
{
return a*10;
}
over 98,830,000 function calls to each:
f1 took 35581ms
f2 took 17989ms
f3 took 32858ms
------
In otherwords, in this instance at least (i'm doing more tests now)
Strict data typing is the slowest thing you can do:
No data typing 'at all' is faster
and using the wildcard data type is about twice as fast as strict data typing
WTF FLASH!?
At 4/21/09 11:38 AM, dELtaluca wrote: Strict data typing is the slowest thing you can do:
No data typing 'at all' is faster
and using the wildcard data type is about twice as fast as strict data typing
Why in hell would that be right? Something seems fishy. Did casting as different data types speed up that function? (Number/int/uint..ect...ect)
At 4/21/09 11:38 AM, dELtaluca wrote: Strict data typing is the slowest thing you can do:
No data typing 'at all' is faster
Ha! Finally a clear proof that my AS: Variables tutorial was truly the best. In your face, smart people!
At 4/21/09 12:10 PM, Toast wrote:At 4/21/09 11:38 AM, dELtaluca wrote: Strict data typing is the slowest thing you can do:Ha! Finally a clear proof that my AS: Variables tutorial was truly the best. In your face, smart people!
No data typing 'at all' is faster
In a lot of instances yes, but in things like simple 'for' iteration, uint still triumphs over int, which triumphs vastly over the wildcard, and no data type
At 4/21/09 11:38 AM, dELtaluca wrote: function f1(a:int):int
I've got a slightly more specific set of tests running now. I'll let you all in in a bit.
...
Ok here's my test in Flash CS3.
I ran tests on four data types. Number, int, * and nothing defined. In each case I was just changing the type of the parameter in the following function:
function b(a:#){
return #;
};
So the return type remained undefined in each trial.
In addition, I went through these types with five different functions to be returned:
return a;
return (a + 1);
return (a * 2);
return (a + a);
return (a * a * a * a * a * a * a * a * a * a);
For each combination of function and input data type, three trials were run, for a(1), a(2), and a(3) (as we will see, the actual number fed in doesn't seem to have any notable effect). In each case the function was run 10,000,000 times and timed.
The results were as follows (I can't be arsed making them more presentable - they're fairly self-explainitory anyway):
A- : Trials for return a;
B- : Trials for return (a + 1);
C- : Trials for return (a * 2);
D- : Trials for return (a + a);
E- : Trials for return (a * a * a * a * a * a * a * a * a * a);
-1 : Number data type
-2 : int data type
-3 : * data type
-4 : type not defined
All times in ms
BACKGROUND (EMPTY LOOP):
66
68
69
A1: 1403
1407
1404
A2: 1502
1500
1502
A3: 1310
1309
1308
A4: 1309
1306
1306
----
B1: 1602
1603
1603
B2: 1562
1558
1560
B3: 1831
1829
1829
B4: 1834
1832
1829
----
C1: 1605
1605
1610
C2: 1608
1605
1605
C3: 1692
1695
1699
C4: 1693
1695
1692
----
D1: 1538
1538
1537
D2: 1601
1598
1599
D3: 1837
1834
1839
D4: 1831
1830
1833
----
E1: 1761
1759
1759
E2: 1800
1800
1804
E3: 1894
1893
1897
E4: 1894
1894
1896
What can we see from this data? Well, unsurprisingly having no data types and the wildcard come back exactly the same, so let's not bother discussing that any more.
Having an undefined data type proved faster in the first case when no mathematical operations were actually performed. However, in cases where addition and multiplication were required, defining a Number or int data type provided a substantial improvement. This was most notable in the case of addition, and less so in the case of multiplication.
The int data type outperforms number in simple addition (a + 1), but falls behind in all other trials.
Most strikingly, in the two trials where the parameter, a, was used several times (a + a) and (a * a... * a), the relative efficiency of defining a data type wasn't significantly higher than in cases where the variable was only used once. This is most noticable in the last set of results, where despite using the variable ten times in a calculation, the relative benefit of setting a data type is lower than in simple addition. This could be due to Flash automatically sticking with whatever data type was relevant for the last operation.
As far as the benefits of data typing goes, I'd conclude that unless you're going to make empty functions which just take in one value and shit it back out again, setting a data type is probably beneficial. In addition, we could probably extend this to setting a return type; unless you have a function where no operation needs to be performed on the return, it is probably beneficial to set its type.
This post is much, much longer than I intended :/
At 4/21/09 12:27 PM, Paranoia wrote:At 4/21/09 11:38 AM, dELtaluca wrote: function f1(a:int):intI've got a slightly more specific set of tests running now. I'll let you all in in a bit.
Ok here's my test in Flash CS3.
Send me the .fla please, i want to test it in the profiler, testing withing flash cs3 is no where near as accurate
To be honest, I'd say that the precision of the results is a lot more than I was expecting, expecially considering how both the trials and sets 3 and 4 were all within a percent of each other.
That said, I noticed a pretty substantial offset in my results. When doing the background checks I just looked at an empty loop - I just did it again properly with an empty function (function b(){};), and the actual background is about twenty times higher:
1065
1065
1063
What this effectively means is that all of the points highlighted are actually about three times as significant :P
The profiler lets you see exactly how many milliseconds over the span of the performance profile that a function occupied, all of these tests were done with over 10 million calls to each function.
Paranoia's speed tests : Profiler
function (a:#) { return a; }
Number 0.017 735 ns
int 0.038 024 ns
* 0.036 789 ns
0.035 753 ns
function a(:#) { return a+1; }
Number 0.017 175 ns
int 0.037 775 ns
* 0.038 900 ns
0.042 638 ns
function a(:#) { return a*2; }
Number 0.019 486 ns
int 0.038 092 ns
* 0.038 768 ns
0.037 092 ns
function a(:#) { return a+a; }
Number 0.020 162 ns
int 0.038 185 ns
* 0.038 962 ns
0.042 131 ns
function a(:#) { return a*a*a*a*a*a*a*a*a*a; }
Number 0.038 496 ns
int 0.045 727 ns
* 0.047 295 ns
0.047 590 ns
which pretty much support your results exactly:
---------------
There are times when strict data typing is slower than wild card, or even not typing at all.
I am not saying however that you shouldn't strict data type, because remember that if you are using objects, then accessing members of those objects is going to be (using profiler again) 4x faster than not typing, or using wildcard, so any of these 2-3 times diferences (the max i've managed to find) are going to be outweighed by the benefits of accessing members with strict data typing :P
At 4/21/09 12:54 PM, dELtaluca wrote: which pretty much support your results exactly:
ignoring the fact that Number was about 2x faster in most of these tests than int and the others :P
