It evolved into a hybrid and extended cut. All caretaker deleted scenes and appearances are back because they’re awesome.
Sorry for the double post, but I did find a minor error at 1:20:20. As Rey reaches out to heal Ben, you can also hear her pick up her lightsaber and start walking away. This is likely because of the switcheroo that you did to this scene.
On a further note, if you do feel like fixing that, there was the Leia line of her saying “Come home” to Ben when she’s holding Han’s medal. I don’t remember if you ever shot down that idea, tbh.
Planning on watching TLJ Rekindled and v3 back to back this weekend. Excited for this since I haven’t seen this movie in its entirety since v1 of this project. I went ahead and skimmed through some of the audio changes, and you did a fantastic job integrating them! I’m confident if I were a first time viewer (who I’ll be watching it with this weekend) I wouldn’t be able to tell the changes.
EDIT: Nevermind
Here is another thing I noticed which ChatGPT has happily explained the distinction between:
“…which has arisen from a dark Imperial fortress…”
“Arisen” is the past participle form of “arise.” The verb “arise” generally means to begin, originate, or to come up. Using “arisen” emphasizes the emergence or origination of something new or something that wasn’t there before.“…which has risen from a dark Imperial fortress…”
“Risen” is the past participle form of “rise.” The verb “rise” typically means to move from a lower position to a higher one. Using “risen” emphasizes the act of elevation or increasing in power or status.Both can technically work in the context you provided, but they give slightly different shades of meaning:
If you use “arisen,” it emphasizes the FIRST ORDER originating or emerging as a new force from the fortress.
If you use “risen,” it emphasizes the FIRST ORDER ascending or increasing in power from the fortress.
Given the narrative nature of the passage and the notion of the FIRST ORDER’s growing power, “risen” might be slightly more fitting. However, “arisen” could also work if you want to emphasize the emergence of the FIRST ORDER as a new or distinct force. Choose based on the nuance you want to convey.
Me personally, I think “arisen” is the proper word here.
Yup, that sounds pretty darn good. I’d just make it quieter and more subtle.
It sounds like JEJ - doing an interview.
We ran into this same issue with my Mark Hamill clone. Just the way the actor voices the character is completely incompatible with their normal speaking voice (which is what ElevenLabs always emulates).
It just sounds like a Darth Vader voice performance from a video game. Voice clones can be a perfect 1:1 and still not sound right because it depends on the intonation and manner of speaking so much. Just the little ways that somebody shifts their voice and pronunciation in a performance.
Kind of a link in from the Rogue One thread discussion - if we ever choose to commission Ryan Golden I’d be willing to pitch in a bit just to have him say that line for this scene.
I don’t think AI Vader sounds fully there quite yet. Which is surprising because AI Kylo Ren (masked) sounded pretty darn good when it got it right.
Or just “Finish what I started…” for a callback to TFA.
Waiting for Andor’s next season is a good call. Unless I am grossly mistaken, the final season has a timeline of a few years to lead into Rogue One. I just hope it is not significantly delayed by Disney’s current issues and the Hollywood strike drama.
It only had a week of filming left and then the actors strike happened.
Woah, that’s unlucky.
They don’t shoot things chronologically so it’s probably some scenes with a similar location in the middle of the season that need to be done. But since they’re so close, it makes me wonder if they might just work around that for now in post-production (depending on how long the strikes go for).
I mostly just describe what needs to happen on a scene-by-scene basis. ChatGPT recently introduced a new feature where you can have it remember how to respond to you for every conversation and response. I gave it a portion of TFA script so it’s responses are always in the form of a script.
For dialog, I’m usually a little more direct. You can tell it what needs to be said by a character and it will fill in the blanks on that, or you can tell it exactly what they should say in quotes.
It’s the most helpful for action scenes. Most of the time I just say something stupid like: “They fight on the wreckage, Ben aggressively and Rey defensively” and it fills in all the detail.
Hello everybody!
Recently, a small YouTube channel by the name of “PRISM” released a couple of YouTube videos detailing a potential Star Wars sequel trilogy which incorporates more prequel elements, plenty of scenes/action between the OT trio, as well as fascinating alternative takes of the characters that already exist in the sequel trilogy. What I liked most about the video was its use of AI imagery and narration.
Here is a description from the second video in the series: “Our goal with this series is to build a fan continuity that hopefully feels more authentic to the Star Wars experience fans have come to love, building off of 1-6, while still incorporating the new characters and over arching plot established by Disney. We are NOT reviving legends or George Lucas’ treatment in this cannon but instead reorganizing the Disney canon.”
Inspired by this novel use of technology, I decided to run a lot of these ideas through GPT-4 to create a functional script in my freetime. So far, I’ve made it about a quarter of the way through episode seven. I have modified a few of the ideas in the video slightly in order to be better for a movie / more believable story-wise. For example, I’m not pitting Jedi Master Luke Skywalker up against Captain Phasma or evil apprentice Rey. Those two would be defeated almost immediately. Yet, I also expanded on plenty of the ideas in there. For example, we get to witness Captain Phasma be a Jedi-killing badass.
I’ll link to the original videos and the rough draft of the script that I have so far below:
TFA Video: https://youtu.be/CyxQyNQ00qo
TLJ Video: https://youtu.be/xWaAQVU2IwQ
Script (ROUGH DRAFT, NOT EDITED YET): https://drive.google.com/file/d/1m-Fdu0N-fXfqmZBReAobplCuQXBslmqI/view?usp=sharing
I would like to hear from some people on the forum here: What would you like to title the first film in this series? PRISM never actually titled the first one except referring to it as “The Force Awakens”, but I personally don’t think that title is appropriate anymore.
Is it just me or are the forums getting hit with more bot accounts than usual lately?
Could you try applying the same effect to the AI line? Because now the AI line is very clear while the other line has more artifacts present.
I do find, though, that putting the lines you want in quotation marks and adding something like “I said slowly and menacingly” after that helps (if you use “he said” the AI tries makes the voice you want then narrator and they deliver the actually line in a slightly different tone so it doesn’t quite work)
Hey, this is a pretty good idea, especially seeing as how ElevenLabs probably intended for their service to be used to narrate books this way.
Would you say this works most of the time?
Ah… I’m not sure what feature might do that in audacity?
I’m not skilled at applying reverb but here it is with quieter volume and a very slight vocal reverb: https://youtu.be/8NFKKmt56yU
Here is the file with only pitch and eq changes so somebody else can tinker with the reverb and volume: https://drive.google.com/file/d/1SA_mjRJS5yGCpETffTrWumPEf5P34MW1/view?usp=sharing
Not unless you want it to no longer have the gap space between “you” and “give me”. It would be a very quick reading more than likely.
I can fix that. The original line was already a bit deep, so I didn’t apply as significant of a pitch change to it before putting it into the video. I still have the original so I’m going to experiment with EQ and a more impactful pitch change.
EDIT: Here is a -5 pitch change instead of -3 (Audacity) plus some EQ: https://youtu.be/IPEDsdQrjlc
Here are two test I had time to put together: https://drive.google.com/file/d/1skwWcHS3STJdEGOoNNSTji1qBlDo2wwW/view?usp=sharing
Whatever you did here sounds nearly perfect and I can’t seem to replicate it on my end. Your two other recent tests sound robotic so if you’re doing the same thing for everything I’m not sure what happened.
EDIT: FINALLY replicated it with the new line - https://youtu.be/4OoYAj4iFcM
Use the following prompt to get it to speak slower krlozdac:
“You’re a ghost… In a rotting clone…”
The periods help a lot.
I think you pitched it a little too harshly. It’s starting to sound like he’s got his mask on.
I’ll let krlozdac handle the new lines we want then since he seems to be getting better outputs than me.
Ideally, we could throw in some mask less Kylo audio from Lego TFA to help balance out the results, but it seems Octorox’s drive no longer has any of the audio from that game.
That’s moreso for music covers I thought?
The truth of the matter concerning our current voice clone is that Adam Driver is reading an audio book with a higher pitched, cheerier voice. Meanwhile, Kylo speaks with a very dark authority in these scenes. Krlozdac and I tried to lower the pitch on the voice outputs but there’s only so much that can do.