Thinking of taking the plunge
From the specs it has a GPU ( Graphics Processing Unit ).
1) Is this only used for the display ?
2) If I ran for example ps3mediaserver, or mediatomb, would the GPU be used for transcoding to serve remote TVs etc.
Thanks
Jim
Re: How does hardware video acceleration work.
The GPU accel. will be only used if told to.
The only apps successfully doing that are XMBC , Quake and omxplayer , AFAIK.
The GPU supports OpenGL ES , but many apps support OpenGL only. Porting seems to be hard work (Why ? Is the naming scheme just a marketing trick ?)
The answer to your last question is therefore no.
Please note that only two or three codecs are properly licensed for the current Pi (out of the few accelerated ones)
The camera module will change that.
To summarise : No ENcoding or recoding legally or technically supported ATM.
ghans
NB : I'm generalizing and omitting , just not to complicate things.
The only apps successfully doing that are XMBC , Quake and omxplayer , AFAIK.
The GPU supports OpenGL ES , but many apps support OpenGL only. Porting seems to be hard work (Why ? Is the naming scheme just a marketing trick ?)
The answer to your last question is therefore no.
Please note that only two or three codecs are properly licensed for the current Pi (out of the few accelerated ones)
The camera module will change that.
To summarise : No ENcoding or recoding legally or technically supported ATM.
ghans
NB : I'm generalizing and omitting , just not to complicate things.
• Don't like the board ? Missing features ? Change to the prosilver theme ! You can find it in your settings.
• Don't like to search the forum BEFORE posting 'cos it's useless ? Try googling : yoursearchtermshere site:raspberrypi.org
• Don't like to search the forum BEFORE posting 'cos it's useless ? Try googling : yoursearchtermshere site:raspberrypi.org
Re: How does hardware video acceleration work.
OpenGL ES is designed for embedded systems, which are systems with small processors and memory. One guiding design principle was that while in OpenGL there are several ways to do things, in Open GLES there is exactly one. That keeps the processor requirements down and the memory requirements small.ghans wrote:Porting seems to be hard work (Why ? Is the naming scheme just a marketing trick ?
Murphy's Law then ensures that whichever OpenGL application you choose, at least some of it will have been coded using the approach that doesn't exist in OpenGLES.