Daz 3D is part of 
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.Daz 3D is part of
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Added some actual light and am very nearly done with all the dynamics, now I just need to tweak some materials, fix a couple animations and resolve the pathfinding, figure out what's going on with transparency in Karma, and finish texturing the scene.
Nice job & a lot of work.
The second part of my lip sync test:
-- Walt Sterdan
Great stuff, Walt!
I've a OT question for. you, since you are on a Mac:
So quicktime player has a limit of 1 min. when opening and saving image series? It does not seem related to file size, am I right?
Can you import image sequenses straight into something like daVinciResolve or Final Cut, without limits? Need to improve my workflow, thanks:)
Thanks!
I wasn't aware of a one-minute limit on Quicktime image import, as all my clips are shorter (I think the longest is about 48 seconds). I occasionally did have an issue where Quicktime couldn't load the full sequence and kept stopping at the same place, but by either restarting Quicktime and sometimes closing all other apps it then imported the full sequence. I know you have a lot more memory than I do, but you could try that just to check.
daVinci Resolve (https://jayaretv.com/tips/import-image-sequence-in-davinci-resolve/) and Final Cut Pro can both import sequances of images. I've never used either, but there are lots of tutorials that show how.
A neat little free app for working with videos is Adapter (https://macroplant.com/adapter), which will also import image sequances. I use it a lot when I want to take larger files and squish them down for messaging and such.
Let me know what you wind up using.
-- Walt Sterdan
Looking into it, tks!
Resolve is great. It's what I use to stitch together render sequences, it's got very powerful postprocessing and especially color tools, and can even work with 3D models (though I haven't quite figured out how to do that without crashing the program).
A preview of a couple animations I'm rendering out:
I figured out how to create dynamic dreadlocks, so I'm rendering out a simple walk cycle of this character (without the jacket):
The crowd scene is in the final stretches (even though I'll need to render it all over again with the above character once I've got her fully set up), but here is what it looks like now:
Tks @Gordig, will check out! Great looking Houdini-renders!
My daughter uses the free version of daVinci Resolve to stick my clips together and add a few special effects as well to make the finished episodes. Highly recommended.
-- Walt Sterdan
Looking forward to episode 3. This is perfect for Saturday morning TV or the kid's specialty channels.
Thanks very much, that's very kind. This was originally just meant to be a two- or three-minute teaser as a proof-of-concept for DAZ's lip sync and a learning project as it's my first serious attempt at animation. I just gave too much script to the voice actors as I wasn't sure how much I'd need for a sampler and wanted to use it up (I hadn't actually written any story past Episode Two). I'm currently testing computer-generated voices to see if I can get use those with minimal effort to do another style of cartoon. I'll make an Episode 3 to just finish that little story off afterwards.
-- Walt Sterdan
It's a challenge to get actual performance out of the computer voices. I found these to atleast deliver the lines with a bit of emotion. https://replicastudios.com/product/studio
I've looked at them and they're in my finalists; I was wondering, though, when you have "4 hours of voice generation" is that for exported audio, or every time you generated it whether it's a final version or not?
From their description, it looks like they deduct from your time for each generation; so, if you generate a minute of speech, add a comma or change a word and regenerated, that's two minutes, and so on.
-- Walt Sterdan
Got dynamic dreadlocks figured out. There's some wonkiness in her animation, I don't fully have a handle on the hair shader, and I need to dial in her skin, but it's basically on to the next steps.
Looks like you've almost got it nailed! Great character.
--Walt Sterdan
You get the option to "PLAY" the dialogue as often as you need to (they do charge something for this), to test out the delivery, the different voices, adding commas and such, and then get charged the seconds of time for the actual "SAVE TAKE" which you can then download as wav or mp3 or ogg. ( so a 10 second audio output uses say 30 second of time from your available pool) Outputing the same dialogue with a different voice is considered another take, and would be applied to the deduct time.
I have found keeping the takes short, just a sentence or two allows for more focused output. 15-20 seconds still reduces those 4 hours, but as you learn the functions, the Svae takes get better.
There is a try for free option on the site, and well worth trying out.
I'll give it a try this weekend. Thanks very much for the info, it's an excellent process.
I had planned on using another system, like Speechelo (which I have free access to) to fine-tune the text before sending it to Replica, but I won't bother for the trial.
Thanks again,
Walt Sterdan
150-ish hours is probably too long to spend rendering something that is basically just a test. I thought I fixed all the foot sliding, but I guess switching from "locomotive" to "in-place" animation to stop them from walking through each other destroyed that work. Solaris outputs EXR, so I guess I also need to finally learn how to color correct well.
Hooooooooooo boy. This took me way too long to crack, so at least now I know a whole bunch of ways NOT to do it. I tried all kinds of way to prevent the hood from sliding off her head, from morphing it down and pulling it forward, to adding wind behind her, to giving the hood entirely separate and much looser simulation settings. Finally, I just pinned some polygons on the hood to the top of her head. All the while, I had a full head of hair that I needed to stuff under the hood, which involved sculpting, scaling and using what is effectively a dFormer, but after all the work of stuffing the hair under the hood was done, I still had problems. Including the hair along with the figure and the already-simulated pants as colliders for the jacket, the collision went all weird, sometimes falling through the hood, sometimes falling through everything, and sometimes simply exploding. I then had the idea to cheat the hair by restricting it to the front of her scalp, but that didn't solve the collision problem. Finally I had the idea to fill the gaps in the hood to create a volume, then using a boolean operation to dynamically delete the hair points that fall outside of that volume. AFTER ALL OF THAT, I discovered that parts of the jacket were just wobbling around for no reason I could figure out, so I spent way too much time adjusting simulation settings until finally deciding that it was good enough. Also, the mesh eyelashes were causing some weird surface glitching on her face because opacity in Karma is a little weird, so I gave her some new lashes using Houdini's hair system that involved painting attribute weights and doing some math to set normals for semi-realistic hair growth, and you can't even see them in this shot.
Long story short, here's a thing I made:
Still some wonkiness, but I don't think there's anything wrong with the underlying animation, so maybe I just didn't reset all the solvers before re-caching the simulations. Now I need to give her some proper animation so that I can fit her into the crowd scene, and then months of work for what will probably be a 3-second shot will be complete.
Fixed all the foot sliding, although there are some small tweaks I could still make to certain animations. I also tried a different color grade, and am on the way to figuring out how to keep my render times more reasonable without sacrificing quality.
A new short animation. A lot of audio mixing at the end to create echo effects for various sized rooms, halls and "open spaces" depicted on board the spaceship.(Project time Jun 15,2023 - July 01,2023)
Could watch it on youtube, but the forum embedded video said error.
It was interesting how there was no sound (the silence of Space) outside the vessel and the music started once the camera entered the cockpit of the ship. You might want to consider using dissolves for your transitions between shots to make it all flow better in editting. It all looks good by showing the big, expansive, and functioning spaceship. What happens next?
Very cool, it really shows off the ship well. I see they've arrived in times for New Year's, interesting. Great job with the audio of the RF scan. Just waiting for the wake-up now.
-- Walt Sterdan
I'm not sure what to tell you about the embedding, it's working for me from Chrome. Thanks for popping over to YouTube to check it out.
Although I wanted to silence of space, I think some viewer may give up before the video enters the ship.
I usually do dissolves, honestly I got caught up in the audio edits. The video is the product of "How much animation can I put together in two weeks". I'm going to try it again with a little more time, 30 days.
And then next :)
Thank you. December 31st was picked to cover the music used. All of the music used was from the same year and being broadcast. I picked the date so that all of them had been released by Dec 31 of that year. I picked the date by adding the broadcast year of the music to the lightyear distance to Antares and an another 300 years until an interstellar ship like that might exist.
I'm making cameras and sets for the next video. I'm not sure how to accomplish dialog in a convincing way on a zero budget.
I'm stepping away frmo 3D animaton for a bit while I wait to see what DAZ Studio 5 will give us.
In the meantime, I'm converting some of my 3D assets to 2D ones and using Reallusion's Cartoon Animator 4 to see how hard it would be to animate a cartoon. As before, the goal is to do something comparable to a 60s cartoon, not Disney feature film.
I've created and had rigged custom male and female prototype bodies and I'm currently working on the heads (I'm using the heads from "Campus Heroes" for now), but I wanted to see how things would work with their lip syncing.
More importantly, I was testing play.ht's "Ultra Realistic" voices for text-to-speech, which I will be using that for the forseeable future. The story and animation might not be your thing, but if you're interested in text-to-speech voices, you should give it a listen.
To see play.ht voices in action, you can check out the first episode, part 1, of Starship Hrimfaxi ("The Hrimfaxi Adventures" will be 3D, "Starship Hrifmaxi" will be 2D).
I set a goal of seeing how much 2D animation I could do in 20 hours, max, if all the assets needed were ready to go..
This is it. It probably needs an hour or three of tweakng, but overall I'm happy. Moreover, I haven't had this much fun in a very, very, very long time. I'll start on Part 2 on Monday.
-- Walt Sterdan
For decent text-to-speech voices for free, I think Microsoft's Azure has a good selection of voice and a very generous amount of words/characters per month (as well as free transcription to give you text from a video):
http://ttps://azure.microsoft.com/en-us/products/cognitive-services/text-to-speech/
For $9 amonth, play.ht has a decent number of words per month as well as their "Ultra Realistic" voices, used in my 2D video linked above.
I use DAZ's 32-bit version of DAZ Studio and the built-in Mimc lip sync, but other, better animators than I should be able to advice you of better methods.
-- Walt STerdan
The built-in Lipsynch in the best system with artificial voices while Papagayo is the best when dealing with real voices and songs (you need a free plugin to work with DAZ Studio)
I tried many other options but no one gave me better results than those two.