New Flash Player released

yup, Flash Player 9 Update 3 (about which several people here talked as the nice new beta player for a while) is now downloadable in release player version here:
http://www.adobe.com/products/flashplayer/

It has lots of new features and bug fixes, probably most noteworthy:
it brings support for various new media formats (video/audio) and makes flash content play at pretty much the same performance as in standalone player inside the browser.

This really makes quite a big difference, so yeah, sweet :)

[EDIT] Hold on a minute, after some quick tests i´m not sure anymore the things i marked in italics hold true for the release player. I just tried a few movies and all run WAY worse than in the last beta player before this release.
I have to do some more tests but so far it doesn´t look good.

Sweet, sweet, sweet, sweet.

Sweet.

Squize.

Hmmm, maybe not that sweet actually.

Just chatting with Tom and he voiced concerns that it may not be as fast as the beta, and that seems to be holding true on the few tests I've run.

The plasma effect in my sig was always running at a rock solid 30fps using the beta player in FF, now I'm getting around 22fps.
The tunnel effect is much slower too.

Oh adobe, what have you done ?

Have a test everyone, let's us know what you've found.

Squize.

yeah, i just updated my first post in the thread, too.
So my setup is like this: On firefox i had the latest beta player installed, on IE i now installed the just released official release player.
Now my content is running at way better performance in ff (beta player) than in IE (release player).
Again, will do some more tests before getting mad, but wow, just wow (for now).

It feels like the swf starts at full speed, and then drops off after a second or two.

Cosmic.

Squize.

i have done some more tests now and so far it boils down to this for me:
The release version ( 9.0.115.0 ) achieves pretty much the same
performance in browser playback as in standalone playback.
So on that end there is a (partially very big) improvement when comparing the old release player and the new one. That is great.
On the other side the latest beta player ( 9,0,64,0 ) still achieved way better
performance than the release player now does for me.

To verify this i tested a performance intensive game featuring an fps counter (and the game is set to a high target fps)
On firefox with the latest beta player it achieved around 100 fps.

Then i installed the release player in firefox and now it only achieves
around 63 fps, also with more drops way below.

Its still a great step when the player achieves pretty much the
same fps in browser playback as in standalone playback but yeah, i
wonder why there´s quite a loss (in my test case over 35 fps, may vary of course depending on what you do and to which target fps you had it set) from beta player to release player and i hope this can be adressed.

I'm getting less than desktop performance, which is weird.

You've got a dual-core machine haven't you mate ? May have something to do with my results being worse than yours.

Really gutted about this, the beta 3 ( 2 Also ) was just sex, flew along no matter what you threw at it, now release 3 doesn't.
I can't see there being an update in the near future which gives us back that speed, so that's it until F10. At best.

I just can't understand why a final release can run slower than the previous 2 beta versions.
It's better to have loved and lost than never to have loved at all. Unless it's performance in Flash. Adobe you big tease.

On the bright side, I was going to add a cool raindrop effect to the current game and only run it if the player was >= beta 2, glad I didn't waste a couple of hours of my life on that now.

Squize.

Sorry, i wasn´t clearer about that: With the release player i´m getting less performance in browser playback than in standalone playback. When i said
"The release version achieves pretty much the same
performance in browser playback as in standalone playback."
i meant the performance with the release player in browser playback is quite close the one it achieves in standalone playback for me (and a good bit higher than the old release player achieved in browser playback (as far as i can say by now reviewing the content i tried so far).
And yes,i have a dual core so performance probably is a good bit worse for single core users.
In either case, i tested several performance intensive things now also setting their target fps high and there´s a huge performance difference between beta and release player, the release player partially loosing way over 40 fps. Such a huge difference can´t be easily explained with "oh,yeah, we fixed some bugs and that made it a tad slower".

Unlike you i´ve actually already worked quite a while on performance intensive content that ran great with the beta (but was quite a bit above and beyond what the previous release player could handle nicely) and now i´m pretty much f*****.

So you can imagine, i´m getting way more than just annoyed about this.

This goes way beyond just saying "oh,see how much faster the player runs now" while showing content on a very special mac os+certain safari version setup at a conference where it just can achieve such performance on this special setup; this also goes way beyond showing "vector shapes" moving around at unbelievably high performance and it later turns out it was made using cacheAsBitmap which in real life scenarios is a feature one should avoid at any cost.


I really hope this is a last second bug related issue and will get addressed, and not some crazy marketing stunt where in beta phase ram/cpu caps get maxed to limits of which one knows: yeah, we´ll just drive those down again on release, otherwise it wouldn´t be stable, nice to get the performance PR now though"
.

I´ll see what the news are on this topic in the next few days and if we can get any further info or word on whether this will be addressed, right now this flash player for me turned from the one i was most looking forward to ever to the one i´m most disappointed of.
BUMMER.

So, should I bother to update?

I can install it on Firefox but it wont install on IE. After installation it still says 9.0.60.120. I think I have to uninstall the beta first but I dont know how to...

wait, there's a flash player 9

I gotta gets with the times!

I had the same problem as Cortana, although I wasn't aware that I had the beta player installed (nothing seemed to run at standalone speed for me), but I could be wrong about that...

Ali

Download and run the Flash Player uninstaller. You can find it at the bottom of this page:

http://www.adobe.com/support/flashplayer/downloads.html

Then reinstall the plugin in both browsers.

I don't know if this is a stupid question, but the standalone player is a different program than the browser-player isn't it?
And that's why the standalone player doesn't get updated when you update flashplayer in a browser? Where do you find the download for the standalone then? Or is this what the automatic updates are usually about when I open up flash?

Thanks!

My test peformed exactly the same in the new player as it did in the previous.

(edit, by that I mean: it plays at the set framerate in the player, drops about 7 frames in the browser.)

I guess I'll be shaving things off here and there for a while longer.

to tonypa: even if its annoying to see content running at way worse performance than in the beta player, yeah, as long as you create content for the web you should test with the release player the people will have, not keep your web beta player with its way better performance in hope adobe will next time really actually release the thing running at that performance as release player.

to Ali: you´d be the first who doesn´t have a huge performance difference between beta and release player. I guess either you had a very old beta version then or no beta at all.

(the issue of having to uninstall the old player before being able to install a new one can also happen if you have a release player installed btw)


to walnoot: yeah,the browser plugin is something different than the standalone projector, you can download those here: http://www.adobe.com/support/flashpl...loads.html#fp9


to Charlie: For most content i tried the performance with the current release player was better in the browser than with the previous release player.
As i said above though that is still a huge loss in performance compared to what the beta player achieved (with which Adobe teased us for more than 2 months).

Sniff I want to go back ot the beta player... that screamed, this new one... sluggish... disappointing, saddening.

exactly.

I did post a comment here:

http://www.kaourantin.net/2007/12/ad...-released.html

Asking what was going on, but the comments are pre-approved and looking at it, mine wasn't.
Perhaps if more people ask there we may get an answer.

As adobe's customers we don't have a god given right to anything, but I think we're entitled to have an explanation somewhere about the difference in performance.
We've been encouraged to join in the beta programme. Part of the beta2 changes was a huge performance increase ( Which it was ). Why release ( well two ) a beta with what turned out to be an artificial performance increase ?

Squize.

It seems the audio trick to create sounds from a ByteArray is not working anymore. Can somebody confirm ?

http://lab.andre-michelle.com/fl-909

Quote:

---

Originally Posted by Squize

I did post a comment here:

http://www.kaourantin.net/2007/12/ad...-released.html

Asking what was going on, but the comments are pre-approved and looking at it, mine wasn't.

---

same with me, posted there, too.
I think everyone who is annoyed about the issue should write a bug report here:
http://www.adobe.com/cfusion/mmform/...?name=wishform
stating that the beta player achieved way better performance than the release player and ask what is up there.
If you have a blog a blog post there might help, too.

Quote:

---

Originally Posted by Squize

Perhaps if more people ask there we may get an answer.

---

Hope the same.

Quote:

---

Originally Posted by Squize

As adobe's customers we don't have a god given right to anything, but I think we're entitled to have an explanation somewhere about the difference in performance.
We've been encouraged to join in the beta programme. Part of the beta2 changes was a huge performance increase ( Which it was ). Why release ( well two ) a beta with what turned out to be an artificial performance increase ?

Squize.

---

yeah, totally agree on all. Of course we have no assured right to expect that a beta or final release of this or that of their apps/technologies has this or that functionality. On the other side hyping the performance of the beta player, getting it hyped for free by all early adopter longtime flash platform developers in beta phase and then releasing something which is massively worse in performance is borderline to fraud in my eyes if it was intentionally done.
If it wasn´t done intentionally done but as result of some technical last minute issue i hope they react on it (fix it) when realizing lots of people see such performance differences and it is an issue.

Quote:

---

Originally Posted by Sietjp

It seems the audio trick to create sounds from a ByteArray is not working anymore. Can somebody confirm ?

http://lab.andre-michelle.com/fl-909

---

Andre Michelle posted the first comment in the post on http://www.kaourantin.net/ about the new player stating there that his audio hack doesn´t work with the release player and is very glitchy with it.

Yes I saw the comment. But I've just tested and it works perfectly for me (Win XP IE v7). Can anyone test the fl-909 on other confs ?

the issue with that one for me is that it works but runs worse than with the beta player, i´m running dual core on windows cp though. Vista itself is more ressource hungry and therefore content running on it should be more optimized to achieve the same performance as on xp.
As André said it doesn´t run well for him on Vista anymore and did before with the beta player.

Plausible answer (but im just guessing here): The beta had no frame or CPU consumption cap, (which is unrealistic for public consumer use).

Well this is avery bad news because I was working on the bitboy player ( http://board.flashkit.com/board/showthread.php?t=712120 ) and it uses the same sound hack...

yeah, i´m very disappointed, too, as i said i was working on stuff making use of that additional performance, too.

To Ray:
yeah,i´d guess so, too (if the performance drop is not just caused by some bug(s)) but then i don´t understand why that is done and also why the previous beta ran so nice and then the release player comes out with so much weaker performance (without several more beta versions in between to maybe adjust the performance or something else some more).

got some crashes with opera (not in firefox ) with some FLV loading stuff- hope they or opera will fix that at some point

Render, could u give links to site that crashes Opera with new FP? I would like to see if they crash my Opera too (PM the link if u dont wish to post it in public).

well not illegal (yet :D)
http://songza.com/
(loads youtube movies and playbacks only the sound)
if I click on a song it crashes as soon as it starts to load- didn´t had that issue before because adobe recently forced me with the beta player because of some labs stuff I wanted to see.

hm, it does not crash my Opera. It doesnt work either since it stays on Loading and no sounds can be heard for any song but no crash.

Quote:

---

Originally Posted by tomsamson

...this also goes way beyond showing "vector shapes" moving around at unbelievably high performance and it later turns out it was made using cacheAsBitmap which in real life scenarios is a feature one should avoid at any cost. ...

---

What do you not like about cacheAsBitmap? It makes using complex vector art possible for games. I've never had any problems with it.

CacheAsBitmap is waaay too slow compared to using BitmapData rendering with copyPixels, you can also use the vector art if you loop in each frame and draw() it to the array of bitmapData that the pixels will be copied from

to turtlesoup:
normally vector art is more performance intensive than using raster art as the shape/fill etc has to be calculated at runtime, each frame refresh.
Now the idea behind cacheAsBitmap is to store the vector art in ram like if it was a raster image once its look with its current props is calculated once (so it doesn´t have to be recalculated each frame refresh).
So far so good (in theory).
Downsides: the vector graphic still has to be recalculated any time you do more than just changing x/y pos of the graphic (for example changing rotation, scale etc). Then that caching has to be redone (which costs ressources each time, too)
which in effect means it runs with worse performance if you have cacheasBitmap turned on for a vector graphic and constantly change is scale,rotation etc.
Ok, then let´s just say we don´t change the scale,rotation etc of the vector graphic but really just the position (in which case one could think right from the start why use vector art if i don´t intend to change the scale, rotation etc at runtime anyway, but yeah..).
Is using cacheAsBitmap on the bitmap then a better choice than just having turned the graphic into a compressed raster image at creation time and having used it like that?
No, because cacheAsBitmap doesn´t store the image compressed as if you for example had that graphic imported as png 8 or jpg. Therefore you use up more ram at runtime.
In my eyes using cacheAsBitmap therefore only makes sense for code generated runtime only existing vector art (using the drawing api) which doesn´t change anymore during runtime (for example used in component graphics).
So yeah, all in all, there is a very specific isolated case where its a good choice but its not at all a feature one should constantly use for all creation time made vector art.

Hey tomsamson & Cimerian,

thanks for the replies. My current game has avatars that move (and animate) around occasionally. When they move they do scale occasionally (in addition to the animation) but they are just standing around a lot too. Everything is vector art and I've been using cacheAsBitmap. Mainly for the standing around images but also for clothing they wear that doesn't scale.

I have thought about drawing the animations to bitmapData and then copying the bitmapData to the screen when the avatars move. Is there a way to save them as compressed images at runtime? It has to be at runtime because the avatars can swap clothes around in game. There would be way too many combinations to save off outside of runtime.

Sorry to derail the thread. But with regards to the new update, it fixed a crash bug I was having so I'm pretty happy with that. I haven't really noticed a change in performance but I never installed the beta update.

if you only use cacheAsBitmap for cases where the graphic doesn´t change in scale/appearance than you can do that (although, yeah, even then having a compressed raster image in first place is a better choice).
If you use cacheAsBitmap on graphics that change in scale a lot that is no good choice (also performancewise).

If you have a setup where you have many images in keyframes in a timeline maybe you should just convert the individual assets into compressed raster images.

Regarding runtime compressing bitmapdata: In theory you could work with bytearray and use some compression algo on that of course, but in reality if you have many art assets going that will probably end up beeing too ressource intensive for runtime game usage.

I´m glad the release player fixed a crash that annoyed you (though i´m not sure if the last beta player didn´t already have that improvement, too ;) )
and yeah, if you didn´t try the beta player with performance intensive stuff to compare (or haven´t tried it at all) then you can´t know what we moan about of course :)