That is a very interesting observation - and the opposite of what you might expect. To try and force this into the frame of a dynamic memory issue, I would be thinking along the lines that the memory gets allocated for a frame and deallocated when it plays, but for 30fps the memory still gets allocated at 60fps - hence the frames which are not required are not deallocated. It also vaguely correlates with our little calculation about the time it takes to get the problem being roughly twice the time that the physical memory indicated.
A prediction - when you run top, you will find that the 60fps video will remain fairly steady on memory use but the 30fps will rack up. I suppose the next thing is for you to check that the same codec is used and then to raise a bug with the codec developer.
A prediction - when you run top, you will find that the 60fps video will remain fairly steady on memory use but the 30fps will rack up. I suppose the next thing is for you to check that the same codec is used and then to raise a bug with the codec developer.
Statistics: Posted by VinceL — Tue Apr 02, 2024 12:56 pm