- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I have a linux based backend service application which want to use MediaSDK to do some encode/decode work. the video frame will send out through TCP/IP so we dont need any video display on the server side. I follow the sample_decoder codes given by Intel MediaSDK and create some example but didnt create HWDevice handler and FrameAllocator. When I call pVideoDecode->DecodeHeader( xx,xx) function I get the error code : MFX_ERR_INVALID_HANDLE . Do I really need create both in order to make encoder work.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
One of the main differences for porting Media SDK code between Linux and Windows is the display handle requirements. For Windows, setting the handle is only necessary if you are explicitly handling your own GPU memory allocations. It may not be necessary for software or opaque pipelines. For Linux *ALL* applications must set a handle. Since Media SDK for Linux has no software implementation, in many cases you won't want to consider system memory. If you are transcoding and do not need access to the intermediate surfaces you can use the extra convenience of opaque surfaces -- meaning you don't need your own frame allocator to get optimized peformance. However, in cases where you are integrating other components and do need access to the pipeline between decode and encode you will need a GPU frame allocator.
More info in this thread: http://software.intel.com/en-us/forums/topic/479998
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
One of the main differences for porting Media SDK code between Linux and Windows is the display handle requirements. For Windows, setting the handle is only necessary if you are explicitly handling your own GPU memory allocations. It may not be necessary for software or opaque pipelines. For Linux *ALL* applications must set a handle. Since Media SDK for Linux has no software implementation, in many cases you won't want to consider system memory. If you are transcoding and do not need access to the intermediate surfaces you can use the extra convenience of opaque surfaces -- meaning you don't need your own frame allocator to get optimized peformance. However, in cases where you are integrating other components and do need access to the pipeline between decode and encode you will need a GPU frame allocator.
More info in this thread: http://software.intel.com/en-us/forums/topic/479998
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Greate that exactly what i want.

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page