GPU acceleration for Live View

[gfunkernator]

This is probably not going to happen because of the extra layer of code that has to be added to support either OpenGL, DirectX, or CUDA but I'm not a coder so.  Would be nice to see some help from the GPU during Planetary Mode capture.  When I zoom to 5x my framerate gets cut from 55fps to 35fps, then 15fps when going to 10x zoom.  Is it possible to use the GPU to accelerate the live view?  Or is this a limitation of the SDK?  I've seen the live view on the camera's LCD even at 30x and the framerate is smooth, not choppy like Live View.  So yeah, that is my request/suggestion.  Thanks.

[astroman133]

I think that the speed and the choppiness is due to the delays inherent in pushing the data over the wire from the camera to the PC.

[admin]

As mentioned above, the limitation is the physical USB data bus and camera's ability to push the live view frame. There is nothing to gain in using GPU as a result.

[s3igell]

10x LiveView is Digital Zoom, and gains you Nothing in regard to Planetary Video Imaging.  5x LiveView is 1:1 Pixel ratio on Canon DSLRs, and as close as possible on Nikons.  This 1:1 Pixel Resolution is the best you can have as a base for Planetary Stacking (unless you move to a purpose-built Planetary Camera with Smaller Pixels).

The only help you can give for Frame Rate is to make sure that there is no competition on the USB Bus for your Video Stream.  Use USB3 if your Camera supports it.  Use a Quality Cable connected directly from Camera to Laptop.  Don't connect other USB Gear to Ports on same side of Laptop (most Laptop designs are parsimonious and have only one USB root device per side of the laptop motherboard - or in some cases only one in total).  And don't have all sorts of other Software running in background to compete with BYE/BYN.

[astroman133]

For lucky planetary imaging the methodology is to collect lots of frames over a relatively short period of time. BYE/BYN can save the individual JPG images/frames or combine them into an .AVI file. Whichever file type is used, the follow-on processing analyzes and grades the individual frames to find those where the seeing is momentarily improved. These are selected and combined. So at this point you have reduced the hundreds of individual frames into a single frame. So even if GPU processing were possible to improve the quality of the video, the video is no longer needed, except as a source of the raw frames.