Certainly with older generation products, most hardware acceleration seemed
to be based around (inverse) discreet cosine transforms, so really it's
working at a lower level than individual frames.

I'm less familiar with the latest kit, but from what I understand it lets
you feed a compressed stream (mpeg, h.264, etc.) direct to the GPU and have
it appear as rendered video in the frame buffer.


On 30 September 2010 15:03, Chris Adamson <[email protected]> wrote:

> I'm not sure about "most" H.26x decoding being done in hardware. On
> the Mac, we only got this with QuickTime X in Snow Leopard.  Prior to
> last year, H.264 decoding was all in software.  Of course, it is
> genuinely a big deal for mobile to use hardware decoding because of
> the performance and energy wins.
>
> I'll also cop to being ignorant of just how much help you get from the
> hardware decode: does the GPU decode individual frames on demand
> (which is what I would expect) or does it do more than that? I don't
> imagine the GPU depacketizes the stream and renders it all by itself
> (for one thing, it wouldn't know what to do with the muxed audio).
>
> --Chris
>
> On Sep 30, 9:47 am, Kevin Wright <[email protected]> wrote:
> > I think the real issue here is that most h.26x decoding nowadays is done
> > with hardware support, via native APIs
> > That's where C/C++ has a definite advantage over Java, and not in the
> > inherent performance qualities of VM vs statically compiled code.
> >
> > On 30 September 2010 14:36, Carl Jokl <[email protected]> wrote:
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > > It is at lest somewhat comforting to know that even an experience C/C+
> > > + developer makes plenty of pointer errors.
> >
> > > The way I see it is that even with the Computer Science background and
> > > understanding O/S memory allocation models and how C/C++ allocates
> > > memory as well as Stack and Heap memory theory the errors still happen
> > > because I am human and miss things sometimes. It is all good and well
> > > when the program flow is simple enough for the allocations and
> > > deallocations to be easy to see and in a straightforward logic. The
> > > more complex the application gets the more complicated it becomes to
> > > manage the memory correctly. Both feeing memory which something is
> > > still using or not freeing it leads to problems.
> >
> > > I think even someone at Sun implied or said that in an ideal world
> > > everyone would write perfect C/C++ code but things like Java exist
> > > because in practice it turns out that this is really really hard to
> > > do.
> >
> > > Maybe I should try Assembler just for kicks....I like pain.
> >
> > > --
> > > You received this message because you are subscribed to the Google
> Groups
> > > "The Java Posse" group.
> > > To post to this group, send email to [email protected].
> > > To unsubscribe from this group, send email to
> > > [email protected]<javaposse%[email protected]>
> <javaposse%2bunsubscr...@googlegroups .com>
> > > .
> > > For more options, visit this group at
> > >http://groups.google.com/group/javaposse?hl=en.
> >
> > --
> > Kevin Wright
> >
> > mail / gtalk / msn : [email protected]
> > pulse / skype: kev.lee.wright
> > twitter: @thecoda
>
> --
> You received this message because you are subscribed to the Google Groups
> "The Java Posse" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected]<javaposse%[email protected]>
> .
> For more options, visit this group at
> http://groups.google.com/group/javaposse?hl=en.
>
>


-- 
Kevin Wright

mail / gtalk / msn : [email protected]
pulse / skype: kev.lee.wright
twitter: @thecoda

-- 
You received this message because you are subscribed to the Google Groups "The 
Java Posse" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/javaposse?hl=en.

Reply via email to