Major Kong:
One topic that I've seen confusing a few people is how Nuke handles caching. At its most basic, Nuke is caching results so that it doesn't have to keep recalculating frames, or specific parts of your tree, in an attempt to keep working in your comp nice and snappy. The details can get rather more complicated, however!
Nothing is more frustrating than thinking you have a cache but you don't.
Frank:
I often find that the easiest deflicker setup works best and fastest for global flicker: use a frame hold to and divide a blurred version of it by the moving footage that is blurred by the same amount. Than multiply the result back onto the original footage:
It's shocking how much this is screwed up on the production side. It's both the Bestboy and the Gaffer. Shame on you.
Jack Binks:
After being involved in the enjoyable project of figuring out the colour correction/grading improvement specs for Bison, aka Nuke 8 (notably the new in panel colour wheels, scopes, maintain luminance, etc as well some of the other bits that didn't make it in time for the release - thanks to Frank R, Jeremy S and other notable luminaries in our world who put aside some time up early on to chat to me), I've been punished with involvement in the project of sorting out colour shifts in our Quicktime handling.
It works pretty darn well.
There is some good disscussion going on, on the Nuke list about Nuke Studio. Just throwing it out there if you need to kill some render time.
Here's is the Foundry video. At the bottom of the page is the NAB event.
CUDA or OpenCL?
The Onion:
“Sometimes a little script tinkering is needed to work in the fact that everyone’s wearing oxygen tanks and 50 pounds of climbing gear, but given the incredible savings it’s nothing you can’t smooth out in post.”
What could go wrong?
Via:John Montgomery
If you are interested.
There have been a lot of searching for J_Ops. I will assume it's to see if there will be a Nuke 8 version soon. I have no official word but it is being worked on.
Remember we are all very busy and sometimes that means other projects take a backburner.
I am also looking forward to a full Nuke 8 version.
Have a good weekend.
John Montgomery writing for Fxguide:
In Pascal, the solution is to move the memory and stack multiple memory chips it on top of the GPU on a silicon substrate (a slice of silicon). They then cut through the memory, directly connecting the memory to the GPU. This will solve having to get memory off the actual GPU chip and onto the surrounding board.
Seems simple enough.