I’m going to be really honest - I’m struggling super hard with this.
I am trying to simply get my sequence to render what the camera sees. I’ve tried to use ChatGPT for help, that brings me in circles, I’ve looked at other posts and I either don’t have the same setup or I’ve checked their solutions.
No actors are hidden. No LOD Screen size quirks (as far as I can tell). I seriously need some help. I want to punch my screen and I’m really calm person. The element that isn’t rendering is my “sky” element. It shows in PIE, when I select the Camera I’m using, but when I choose that camera for the sequencer, the sequence preview nor render result include my sky material output, nor it’s reflections. It’s like it doesn’t exist according to the sequencer.
I would appreciate any help someone could give. I’d be very happy to jump on a screen share with anyone with ideas.
So is the issue limited to a specific cinecamera? Meaning if you drop a fresh camera in the level sequence can you see what you expect? If so what are the Lumen Settings on that camera? Is it possible those are getting overidden and shut off? Does that sky thing have any sort of sequencer overrides?
If its there in PIE then its mostly likely not a MRQ issue. The only other things MRQ might be doing is changing the scalability which you can disable under Game Overrides.
If you have a small repro project I can take a look if you like.
I ended up being able to fix it. What it ended up being is that I had one item that ran from event tuck while all others ran from a function call. Since I’m using audio analysis to drive real time parameter changes, this was the method that I found worked. My sky material was the first material I created and I have a “time” node inside that does not fire off of the same start clock or whatever that is called in unreal (I come from audio / modular synthesis).
I happened to stumble on to this solution. I really don’t understand why some of these things have to be so complicated. It worked when I hit play, so I had no idea it would be a problem in render. There really should be an option to just get what your camera sees at render, regardless of other settings.
PIE and the sequencer both showed the same things FYI, so this was just dumb luck I found it.
When you go to render out of MRQ, things like “tick” and “time” get a little murky, especially when you are multisampling one particular frame for antialiasing or motion blur. The engine can even slow down a little which slows ticks, depending on what else is happening on your machine or even IO. So people see things based on tick/time move super fast or super slow. Or not at all.
If you’re going to end up rendering out of MRQ, especially where you are multisampling, a frame its best to seed those in Sequencer so MRQ knows a specific number.
Is there another way to render that doesn’t get bogged down by those kind of things? This was such an ordeal and it really didn’t need to be, and weirdly, I didn’t find any answers online, I just happened to accidentally stumble upon the issue.
You can try the legacy cinematic capture system. But again, “tick” is a variable rate depending on what else is happening on your machine at the time, so ymmv.