This code works fine on my host Linux machine (18.04LTS) with gstreamer 1.14.1.
But does not work on TX2 (1.8.3) and Xavier (1.14.1).
The result image should be smpte75 image which have snow pattern in the middle.
However, Jetson gives me an image with black region instead of snow pattern.
Hi
I could not solve this.
For my application, I did like this instead.
src->mix.
src->mix.
compositor name=mix (multiple image into one image, just merging without alpha channel)
→ nvivafilter (takes composited image and alpha blending)
nvivafilter only takes 1 input so I use compositor before it.
Hi rary,
thank you for the tip with nvivafilter. I followed your suggestions, but I get an error message with and without alpha channel: erroneous pipeline: could not link mix to nvivafilter0.
Please can you show me a part of your code?
Thank you in advance,
Wilhelm
I found no solution.
With this code the right blue bar of the testpattern is black (should be transparent) and the whole pattern is transparent, see screenshot.
Hi,
thank you for your suggestions.
The example in #2 does not work.
It would be very nice, if someone would inform the developers that chroma keying (alpha method=green) is not working.
Best regards,
Wilhelm
Hi DaneLLL,
your code in #2 produces the video/picture shown in #3 and #5.
There is transparency, but that is not chroma keying. Finally, only the black (former green) bar must be fully transparent, see the attached picture.
Because it is working for you, I guess, we are talking about different things. Therefore a screenshot would be very helpfull.
I’ m using jetson nano with jetson-nano-sd-r32.2.1.zip
Nevertheless, thank you for your support.
Best regards,
Wilhelm
Hi DaneLLL,
I think, I misunderstood your post #16. Sorry.
Did you inform the developers?
Chroma keying will me give new features for searches with a thermal camera and a ususal vis-camera. Nvivafilter is too complex form me.
What is the normal procedure? Will they release a patch?
Best regards,
Wilhelm
Hi WiSi-Testpilot,
Please make a new post for your issue and share steps in detail.
This post is about nvcompositor and we have confirmed it working well.