David_Harry
Second Officer
United Kingdom
Offline
|
Hi.
With Rocksteady activated there is a very noticeable delay, or latency, between the Action's procesessor doing its work to sample the incoming image from the sensor then calculate and process the stabilisation before writing the file.
The delay is caused because the image you see on the Pocket's screen is the post buffer version. This latency is the difference in time between the photons hitting the photosites of the sensor, the sensor data being fed into a temporary video buffer, the data being manipulated for EIS in the video buffer and then finally exiting the video buffer with a delay/latency and being written to a file. Obviously this is a very oversimplified explanation of the process, there are many other things going on such as muxing, image processing etc. and stuff I don't understand but this is basically how it works.
So, my question. Before the video buffer and stabilisation processing, is the picture data viewable as a coherent picture? If so, would it be possible to send the pre-buffer image data, unprocessed and without stabilisation, to the screen. While this won't show a stabilised image it would however give a realtime, no latency or a latency shorter than a frame, picture on the screen.
If this is possible, there could be a mode switch and indicator telling the operator that the mode is switched on, this would let the operator know that they're sacrificing monitoring a stabilised image with latency for the convenience of a realtime image that doesn't show the processed Rocksteady version.
I'd imagine such a function would be very useful, if possible, for anyone shooting time and sync critical footage, which I'd imagine would be nearly all action shooting and most drama shooting.
Cheers,
Dave.
|
|