Tweet
I am not sure exactly how to formulate my vision. Perhaps it fits the DLNA conceptual framework; then perhaps it does not...
I want the simplicity of maintaining on my music server only one lossless version of each track (probably ALAC). More importantly, I want to derive on-demand streams at various bit rates from those lossless tracks. Various factors can influence my choice of bit rate:
1) raw transport bandwidth: lossless on our home LAN; fairly high quality when streaming over a hardwired connection to my office, my wife's office or any of my kid's dorm rooms; lower quality when streaming to any of our iPhones.
2) cost per megabyte streamed: at moderate rates I can ignore the cost of my broadband connection; phone data plans are quite a different matter.
I observe that the cost of the computes to carry out such on-the-fly conversions is quite reasonable. A few months back I bought an HP Ellite 510t with an i7-2600 Sandybridge for a tad over $700. This processor consistently transcodes a single track from ALAC to AAC at over 40x. (If I am not mistaken such transcoding of a single track is a single threaded activity. Meaning that my 4-core hyper-threaded Sandybridge hasn't begun to crack a sweat.) So with just a modicum of pipelining, after encoding the first track, it should be possible for an on-the-fly encoder to stay at least a track ahead on every active playlist.
Can anyone suggest how I might go about assembling such a scheme? (I do understand that pipelined operation may be difficult to achieve.)
I have studied the very helpful "Renderer specific transcoding" thread. But that only seems to address a piece of the puzzle, namely how to stream losslessly over my local LAN.
I have also been reading about Asset Control, though as far as I can tell that is not much help.
When streaming out to the internet I am willing to use different in-bound ports to indicate whether I have fully hardwired versus wireless connectivity. Then using port forwarding I can direct requests to distinct servers, each transcoding to a different bit rate. Here it is not clear whether Asset UPnP is the appropriate server (though I would love it if it were).
One last wish... When my kids use iTunes to copy tracks from the music server to their iPods/iPhones i would like to convert the lossless tracks to AAC at some reasonable bit rate.
TIA for all suggestions, clues, puzzle pieces, etc.
I want the simplicity of maintaining on my music server only one lossless version of each track (probably ALAC). More importantly, I want to derive on-demand streams at various bit rates from those lossless tracks. Various factors can influence my choice of bit rate:
1) raw transport bandwidth: lossless on our home LAN; fairly high quality when streaming over a hardwired connection to my office, my wife's office or any of my kid's dorm rooms; lower quality when streaming to any of our iPhones.
2) cost per megabyte streamed: at moderate rates I can ignore the cost of my broadband connection; phone data plans are quite a different matter.
I observe that the cost of the computes to carry out such on-the-fly conversions is quite reasonable. A few months back I bought an HP Ellite 510t with an i7-2600 Sandybridge for a tad over $700. This processor consistently transcodes a single track from ALAC to AAC at over 40x. (If I am not mistaken such transcoding of a single track is a single threaded activity. Meaning that my 4-core hyper-threaded Sandybridge hasn't begun to crack a sweat.) So with just a modicum of pipelining, after encoding the first track, it should be possible for an on-the-fly encoder to stay at least a track ahead on every active playlist.
Can anyone suggest how I might go about assembling such a scheme? (I do understand that pipelined operation may be difficult to achieve.)
I have studied the very helpful "Renderer specific transcoding" thread. But that only seems to address a piece of the puzzle, namely how to stream losslessly over my local LAN.
I have also been reading about Asset Control, though as far as I can tell that is not much help.
When streaming out to the internet I am willing to use different in-bound ports to indicate whether I have fully hardwired versus wireless connectivity. Then using port forwarding I can direct requests to distinct servers, each transcoding to a different bit rate. Here it is not clear whether Asset UPnP is the appropriate server (though I would love it if it were).
One last wish... When my kids use iTunes to copy tracks from the music server to their iPods/iPhones i would like to convert the lossless tracks to AAC at some reasonable bit rate.
TIA for all suggestions, clues, puzzle pieces, etc.