Debug nvidia video converter

My program uses 2 video converters, which I’m using the V4l api to access. the program looks like this

Argus camera --YUV video–> Video converter --ABGR Video–> Opengl Shader --ABGR Video–> Video converter --YUV Video–> Video Encoder --H264–>

The problem is this: in developing this, I created software implementations of all of the blocks. Everything worked fine until i added the second video converter (right before compression…)

When i added the second video converter, frames from the argus camera seem to be out of order. This only affects the pre-opengl video, because the extra UI elements that get rendered by Opengl don’t seem to be affected.

The two video encoder objects don’t share any data. Do i need to do some locking on the hardware when i’m using it twice like this? How can I start debugging this issue?


There are samples which create two converters:


Please refer to these and check your implementation.

This turned into a race condition in my code, the problem had nothing to do with the extra converter. Thanks for the pointer.