oversimplified: motion jpeg is just like an old fashioned flipbook of photos that can be scrolled through.
normal codecs have only everyso-often a full frame complete image then the following frames only contain the changes.
So say I frame is frame 1 and 2-9 are just changes..
normal playback shows them in order.
reverse playback needs to go from 1to 9 (and all frames in between). then 1to8, then 1to7, 1to6 etc etc..
aka lots of work and repeated work.
many examples, again over simplifying
VJs or techs who may wish to reverse or bounce playback.
Higher image quality, frames are not interpreted as changes but trully rendered as a full frame image.
People who don't care about disk space usage. Full frame images use up more space.
Playback stability (any errors can last 1 frame only... how many times have you seen a video go green then wierd things drawn, then it clicks back into place? thats bad playback till the next i-frame creates a new starting point to work from again).
"proper" Cinema projectors use a type of "motion jpeg" They create a "chain" of solid images that are linked to a specific timecode value. So playback is pretty much guaranteed, any error could be 1 frame long and recovers with no issues. Conversion takes forever though.
Not fobbing you off, but do recommend some googling and reading.
Some codecs are heavy on CPU, some on GPU and some on disk read speeds etc.
No idea about resolume, never looked into, but i know a different media server, codecs with high compression put load on the GPU, codecs with less compression make a higher image quality, larger file size but a load on the CPU.
So, as you add more and more outputs on huge shows, what do you run out of first, cpu, gpu, quality, disk space?
Each media server and show may be better used by a specific codec for different reasons.