Adobe Releases Flash Player 10.1, But Hardware Acceleration on OS X Not Supported

[KnightWRX]

Yes, and?

You've conveniently ignored the fact that I have never faulted Adobe for not implementing H.264 acceleration, nor have I ever faulted them for choosing not to use system APIs. The only problem is claiming they don't exist.

QTKit might as well not exist is the point people keeping pounding into you and you keep conveniently ignoring. What part of "too high level" and "needing access to decoded frames, not a QTMovieView" do you not get ?

Last I will say on this subject.

 

[matticus008]

QTKit might as well not exist

Which is still not the same as not existing, even if it were true.

is the point people keeping pounding into you and you keep conveniently ignoring.

It hasn't been ignored at all. It's the key that you're not seeing. Building on top of Quicktime (the framework) isn't impossible. It's not how Adobe chose to go, because it would require some extra work, and that's a perfectly valid choice. But saying they couldn't do it because Apple wouldn't let them is a fabrication.

Until the 10.1 rewrite, Adobe couldn't use any API, regardless of its limitations.

Further, saying that Flash's historical poor performance has anything to do with H.264 decoding is an absolute lie, since the specialized hardware decoding is brand new for Windows in 10.1.

What part of "too high level" and "needing access to decoded frames, not a QTMovieView" do you not get ?

Quicktime, the framework, and its associated APIs are not just "high level" drag and drop functions. It's an integral part of the system graphics stack at a low level, a peer to the Core technologies, which Adobe does now use. Using it has some limitations that are undesirable for some projects, just like any other API. But again, claiming they don't exist when they do is clearly unsupportable.

Last I will say on this subject.

We'll see.

 

[KnightWRX]

Quicktime, the framework, and its associated APIs are not just "high level" drag and drop functions.

Gah... ok, real last thing :

QTKit, the framework, is not what you think it is then. QTKit is very small and high level :

http://developer.apple.com/mac/libr...Kit/index.html#//apple_ref/doc/uid/TP40001164

 

[matticus008]

QTKit, the framework, is not what you think it is then.

There is nothing in the QTKit prepackaged classes that isn't available throughout the Quicktime stack from the lowest levels to the highest (a higher-level function can't be built without the necessary components at lower levels).

The QuickTime API includes over 2500 functions, divided into tool sets for particular tasks, with special functions and data types for virtually any task [...] You can interact with the QuickTime API at many different levels:
* You can simply open and play movies, letting QuickTime handle all the file and format conversion, synchronization, data buffering, component loading and unloading, memory management, and even the user interface. Prebuilt controls are available for play/pause, volume control, time scrubbing, and cut-and-paste editing.
* You can control the playback or editing yourself, setting the play rate, scaling the duration of movie or track segments, rearranging the playback order, and so on, creating your own user interface and controls.
* You can work with individual components, loading particular importers or image decompressors, applying them to groups of files or blocks of memory, and disposing of them when you are done.
* You can work with individual data samples—synthesizing graphics and overlaying them on video frames as a movie plays, for example—or performing pattern recognition on groups of samples, or even generating whole movies programatically.
* You can write new QuickTime components to support features such as new compression algorithms, new media types, new media capture devices, output devices, or data sources.

All the necessary ingredients are there. Adobe could have passed off video streams to Quicktime, but they chose not to for the Mac (which, again, is perfectly valid). They could have done the same for Windows and left Flash in DirectShow (which doesn't include access to the DXVA APIs, required for H.264 acceleration). But they rewrote for Media Foundation and DXVA; they didn't falsely claim Microsoft withheld the APIs.

 

[Hal Itosis]

finally, why do all you flash deniers insist on using Click2Flash? do you not know how to uninstall a plug-in? honestly.

Because... some lame-o websites still insist on clinging onto the past, so every two weeks or so i find myself "needing" to run Flash. honestly.


--


EDIT: Btw, a huge thanks to matticus008 for holding down the fort here so well... a task made more difficult since these Flash devs are (apparently) also such "experts" in so many aspects of OS design.

 

[Max(IT)]

EDIT: Btw, a huge thanks to matticus008 for holding down the fort here so well... a task made more difficult since these Flash devs are (apparently) also such "experts" in so many aspects of OS design.

I second that.
Huge thanks to matticus008, keep speaking to someone who clearly doesn't listen ...

 

[Amdahl]

I've given the GPU accelerated decoding a go on my parents PC with the latest drivers and the total reduction of CPU utilisation is something like from 20% down to 15%, so there is no 'massive' drop as some will try to make out. Right now on my Mac I can view a YouTube video with CPU utilisation fluctuating between 18-28%, so 10.01 is still a marked improvement over 10.0. The question is whether we'll see further work being done by Adobe or whether they'll stand back on their laurels.

The performance improvement does vary by codec and encoding, so some sites see bigger differences in CPU load before/after. I personally don't expect to see much improvement past 10.1.(excepting the hardware acceleration feature)
[/QUOTE]

Intel realised :

Clock. Speed. Doesn't. Matter.

Clock speed matters plenty.

There is no data point passed the Pentium 4 which was actually slower clock for clock than the Pentium 3 architecture

You are confusing pipeline length for ILP. The P4 had a very long pipeline. Core reversed that. But pipeline length has no (direct) effect on ILP. ILP is affected by things like out of order execution, register renaming, etc. These things have already been done, and can't be done any better than already done. What you're observing is that a 3.4Ghz Pentium 4 isn't much different than a ~2Ghz Core. True. But Not the point. The point is that you can't get a single thread going much faster than that 3.4Ghz Pentium 4, which is pretty obviously true as compared to the performance growth curve we used to be have. In order to get to 3.4Ghz, you have to lengthen the pipeline. But now it turns out that is a dead end, Intel shortened the pipeline and is staying in the 2 to 3Ghz zone that is less hot and power hungry. This doesn't change the raw performance. The majority of improvements in the past 5 years are not CPU improvements, rather finally fixing their memory bus, and piling on huge amounts of cache with the extra transistors that they continue to see growth in.

I appreciate the point you're trying to make (Mhz Myth), but you're trying to stretch it to say that Mhz doesn't matter at all, which is not true. What Intel realized is that Clock speed is played out, because it ends in a brick wall in the vicinity of 4Ghz due to power and heat dissipation. Herb Sutter also says that Instructions per clock (in a single thread of execution) has maxed out. The conclusion is that multi-threading is the only way programmers can get increased compute capacity.

I thought this was all accepted history at this point, and the reason why you have things like 12 core Mac Pros and Grand Central.

 

[Darkroom]

Btw, a huge thanks to matticus008 for holding down the fort here so well... a task made more difficult since these Flash devs are (apparently) also such "experts" in so many aspects of OS design.

I second that.
Huge thanks to matticus008, keep speaking to someone who clearly doesn't listen ...

you know, thanking people for agreeing with you doesn't help strengthen your arguments. rather, it shows your insecurity.

 

[Hal Itosis]

you know, thanking people for agreeing with you doesn't help strengthen your arguments. rather, it shows your insecurity.

Actually... that very statement only served to reveal your own insecurity.

[and questions such as "don't you guys know how to uninstall a plugin" shows real despair as well.]

 

[aliensporebomb]

Fyi

FYI:

I got an e-mail from Adobe a few days ago confirming that PPC Mac 1080P playback thru youtube (and presumably other flash sites like Vimeo etc) have problems with frames dropping etc.

Audio works fine, frames drop precipitously and generally bog down and hang up.

Same exact video in its native format and in html 5 play correctly.

They filed this as bug #2631223 internally.