My last job was shot with a slightly high shutter. Sometimes we ask for this, for keying, rotoscoping… but only the vfx shots. Then we try to regenerate the motion blur.
But on this occasion, everything has been recorded like this. So I have to apply some subtly motion blur to avoid that look so shuttered… but I have to do it massively to a whole timeline.
I wonder if anyone has a quick and effective method. I’ve sometimes used this one but it tends to generate motion blur with artifacts. In addition, working with motion vectors would be almost impossible, as they are many 4K long versions.
There is an old-skool trick of doing this with a “mo-blur” axis. Create a new axis and parent the axis of a surface to it. Whatever transform you add to that axis will seed the motion blur. In the case of the bike in the example you gave, track the bike and add the tracking to the “mo-blur” axis. Of course, you will need to mask the bike. It’s hardly accurate to the real world, but it’s a good fake, especially if you keep it subtle.
Why not just use a timewarp node. I did this on a sequence that was shot with a narrow shutter angle, which looked strobey. I used TW, with motion selected and motion blur, it worked well for just adding that little bit of mo blur to take the curse off it.
We have to do this quite a bit, when stuff has been shot off-speed (high fps) but client wants to pretend it was shot at native fps. I was never very happy with the native options in Flame or even Sapphire S_BlurMotion.
But a while ago we got RE:Vision ReelSmart Motion Blur. It’s cheap, effective and quite often a drag-drop-render with default settings does the job nicely.
Great tip. I didn´t know that a 100% tw can enable motion interpolation and mblur options. It works great, It’s the best method so far, but motion interpolation sometimes make some artifacts, even with low shutter. So I can’t use it globally in a timeline.
Reelsmart is not an option for me now. I knew the tip of inverted parent axis , from pleistocene age, but , I don’t know, it doesn’t work for me now.
The 100% TW trick can create really dodgy results when compared to simple Motion Analysis/Motion Blur. If you use motion analysis, make sure you select local rather than global. It seems to give much better results.