For my second entry I tried using Acculips to get my mouth movement whereas previously I used the Live Face app.

I generated the text from the audio file and was really happy with the results! All the words were lined up perfectly except only for the first 3 words which I realigned to the audio file.

The "raw" results from Acculips can be seen in the linked video.

I tried a few different talking styles and I quite like the "smooth" talking style which I think I would use as a base to work off from.

Some things I will be focusing on would be pronunciations on phrases like "knows no".

It can be seen from the initial video that that the -ee sounds are currently too wide.

In this video, I improved on some things from previously with the initial Acculips alignment.

1. I've moved some of the timings of the visemes, to help ease the transitions to different syllables.
2. I've duplicated some visemes where I want to hold that shape for a bit longer.
3. Checked and corrected some of the visemes. For example, an Ah viseme may have been used whereas it should have been an Oh
4. Changed the expressiveness level in certain areas, sometimes bringing it down or up to even 200% expressiveness level
5. I've also adjusted the timing so the mouth closes at the end

I got the tongue sticking out for the T- sounds!



These are the settings I used for keyframe 402




As you can see from the gif, the frames do not interpolate properly by themselves (I wasn't sure whether to do this with a video or gif, in the end I went with gif)


The result at frame 400 looked good, so I added another keyframe there and for the others I adjusted the values (in hindsight would be good to show the out, up, down values instead of tip down but this was actually my second time taking these screenshots and I really didn't want to do it a third).


At frame 419 his mouth is shut so I added a keyframe in 418 to how it automatically interpolated and reset the tongue values in 419.


Afterthoughts
As said in the video, the tongue results appear much stronger in iClone than on the MetaHuman so that's why the values used on the tongue sliders were quite high.

Got my scene ready (roughly)! Probably would play around with secondary fog data to get some fog on the lower levels.



Added some post processing effects (god rays and lens flare)



Read on if you want to read the boring stuff/me rambling on:
Actually it was difficult finding a good shot.
I have other ones where the tree is more central, but the lens flare isn't at as much of an angle.
Why don't I just change the position of the sun to get it at an angle then? Because then the face of the character (who I haven't included here) won't be lit.
Also Have to rotate the tree to find gaps where the light can seep in through the leaves.

@Just a dude
Nice 😀
Greetings rosuckmedia

Added some head movements and blinks. Would fine-tune the position of the hand and book later. Now to find some interesting shots! and refine the mouth movements more.

So I went through the mouth movements again and corrected the words where its over pronouncing the visemes. For example, when he says "we seek", Acculips narrows the mouth when going to the S- viseme but in reality, we just maintain the wide position after "we". Another example was like in the word "goddess", where it closes the mouth from Ah- to D- and then opens the mouth to Ah- again, but in reality we say the word in one motion. So it was me saying these lines myself to see how we really say it and going along correcting things like these.

ermm...not sure why my video post posted twice (probably cuz I changed subscription settings and clicked post really quickly thinking it hasn't posted yet), and I can't delete one of them, so filler post here I guess :/

So I wanted to post this much earlier (end of August) but got tied up with the Epic MegaJam, which Reallusion were sponsors of this year! Then after that I came across a lot of issues with rendering. Anyways, here it is entry #2!