erik Hi Gecastro , You can check our encoding max limit example that does exactly that - 2x 720P mono and 4k color H264 video encoding. Thoughts? Thanks, Erik
Gecastro That looks good, but is it possible to do the same with the depth image? I'm receiving the following error: [VideoEncoder(3)] [warning] Arrived frame type (14) is not either NV12 or YUV400p (8-bit Gray)
erik Hi Gecastro , Yes, VideoEncoder can't encode depth - only mono frames (8bit gray) or NV12 (colorCam.video). It's a hardware limitation (of the videoencoder silicon block) so we can't add support for it. Thanks, Erik
Gecastro So, does the disparity compression work? I did see this example. I can send and receive the compressed images, but I'm not sure if I'm preserving the correct disparity values after decoding. Do you have an example for decoding the disparity image?
erik Hi Gecastro , Yes, as long as it's INT8 you can encode it. So if you enable subpixel mode, video encoder won't work. And stereo depth is INT16, so it doesn't work etiher. You can use demos below to decode disparity, depending on which profile you used for encoding: https://github.com/luxonis/depthai-experiments/tree/master/gen2-play-encoded-stream thanks, Erik
Gecastro Would this preserve the disparity values? I'm trying to create the depth image from it, instead of visualize it. Thanks
erik Hi Gecastro , Only lossless MJPEG is lossless, others (h264,h265, "non-lossless" mjpeg) aren't, but if "good enough" is ok for your application you might want to consider it, as lossless mjpeg doesn't provide great compression. Thanks, Erik
Gecastro Thanks, MJPEG works well. I was testing with h264, but the compression seems to alter the image.
