TresCom Forums

TresCom : The Trespasser Fan Gateway...
It is currently Mon Mar 01, 2021 7:23 am

All times are UTC




Post new topic Reply to topic  [ 22 posts ]  Go to page 1, 2  Next
Author Message
PostPosted: Thu Feb 18, 2021 1:01 am 
Offline
-=TresCom Website Manager=-
-=TresCom Website Manager=-

Joined: Sun Dec 01, 2002 6:14 am
Posts: 1385
Location: California, USA
Okay everyone, get ready to freak out.

Forum user Velociraptor Alfa pointed me today to THIS amazing AI using Google's "Collab" system. "John Hammond Synthesis" apparently used the untouched John Hammond recordings (that were available thanks to the "alpha" extraction we did years ago that I used for a couple variations of the Jurassic Time Hammond Memoir) that have been available on my site for some time now. It was created by forum user James_UK, AI Clone Voices on YouTube.

https://colab.research.google.com/drive/1AQaAL8lE9sfFtxeSeM_vhMueQmxlnM04

How To Use:

At the link above, as you scroll down make sure you "run" each cell (at the top left of each one, when you hover over the brackets [ ] and it turns into a "play" style button) to download what is needed to use the AI. Or, to put it more step by step, "run" each cell under each heading below, the first line of each heading where the "play" should be right beside it:

"Install Tacotron2 and Waveglow"
!nvidia-smi -L

"Download Models"
force_download_TT2 = True

"Initialize Tacotron2 and Waveglow"
%matplotlib inline

Then under that: # Load Tacotron2
Then under that: # Load WaveGlow

(Note: You will have to go through the process above every time you close the tab/browser, otherwise you'll only have to do what comes next below.)

Once you reach the "Start playing around with the model" cell, change the text within the quotes where it says "Change the text here to make John Hammond say what you want" to make it be whatever you want. Run that same cell after you have changed the text to what you want, and it will produce two sound files below it; a standard one and a "denoised" one of your choosing. You can play each one and also download each one by clicking on the 3 dots on each file.

Some tips:

-Apparently this works better with shorter sentences at a time, so if you really want him to say something specific well it may be best to split it up then edit it together.
-Sometimes if you simply run the cell again you may get a different/better result. Download each one to see which "take" is best.
-Certain words can sometimes be hard for him to say correctly or how you want, so you may want to try a phonetic approach. For example: if you use the word "juniper" he says it like "june-EYE-per". So to have him say it correctly, I changed the actual word to "jew-nip-er".
-Periods, commas, and hyphens can help in pacing of what is said; even if they aren't technically correct.

I have to say, I think this AI is astounding and shocking all at once. If someone could create an offline version of this, that would be awesome! For now, enjoy it as I have! Hours of fun to be had with this!

And please... try not to make him swear TOO much. ;)

_________________
Image


Top
 Profile  
 
PostPosted: Thu Feb 18, 2021 1:10 am 
Offline
Brachiosaurus
Brachiosaurus
User avatar

Joined: Sun Mar 08, 2015 9:54 pm
Posts: 217
Location: North America
This is so awesome! We can now attempt to recreate the unrecorded Hammond lines from the original script, like the "distant call of an apatosaur" line, or Hammond reading off the components of DNA!


Top
 Profile  
 
PostPosted: Thu Feb 18, 2021 1:15 am 
Offline
-=TresCom Website Manager=-
-=TresCom Website Manager=-

Joined: Sun Dec 01, 2002 6:14 am
Posts: 1385
Location: California, USA
[You got ahead of me, TrespasserGuy, haha!]

As a separate note, as the creator of the Jurassic Time Hammond Memoir... this new tool of course puts me in a VERY interesting position.

Some of you may know that there were a big portion of lines of dialogue that were unrecorded for Trespasser, all of which were revealed in writer Austin Grossman's script. Over the years there had been various ways to weave this additional information into the presentation. First, someone once had the idea to create a new character (Hammond's brother) who would describe most of these additional bits of dialogue and have them interwoven between tracks. However, I later found this not very satisfactory, favoring instead of have it be 100% Attenborough's performance. For me the best way to get this additional information was in the PDF books, which places the unrecorded dialogue in what I believed to be the best places within the recorded dialogue.

But now, with this AI technology... what, now?

Do I redo the ENTIRE audio memoir to have this AI Hammond be intertwined with the real one? Part of me does NOT like this idea. While it is his voice, it isn't his REAL performance. Something akin to seeing Tarkin in Rogue One: A Star Wars Story. It both is and ISN'T him; an uncanny valley, but in the world of audio. You can just tell something is off. That and it isn't 100% genuine and never will be, because it isn't really being performed by Attenborough himself how he would have chosen to have done it.

The other idea is simpler; take all the unrecorded lines and arrange them as bonus tracks on their own. A sort of "additional files" kind of thing, like the program already has. This cleanly separates the real Hammond from the AI one. However, it of course does remove the flow and proper placement of information.

So... which is it? Or do I do both? Even I did the former idea... I would have to REDO the ENTIRE MEMOIR from scratch. And I am not sure I have it in me to do that again, I don't know.

What do YOU guys want?

EDIT: Vote on my Twitter poll!

https://twitter.com/jurassicmemoirs/sta ... 29699?s=20

_________________
Image


Top
 Profile  
 
PostPosted: Thu Feb 18, 2021 1:27 am 
Offline
Brachiosaurus
Brachiosaurus
User avatar

Joined: Sun Mar 08, 2015 9:54 pm
Posts: 217
Location: North America
I think recreating the unrecorded Hammond lines and releasing it as a "bonus" track would be good for now. You could call it something like "Jurassic Time: EXTRACTED" or "Jurassic Time: Up the Well". :P I think once the AI gets better at recreating Attenborough's voice you could attempt to intermingle the actual performance with the reconstructed lines. I guess a third option that does not require the use of AI is finding a good impersonator. There are a surprising amount of talented voice actors who can imitate Richard's brother David Attenborough's voice. But I've come to realize that Richard's performance is so much harder to do because of his cadence.


Top
 Profile  
 
PostPosted: Thu Feb 18, 2021 5:45 am 
Offline
-=TresCom Website Manager=-
-=TresCom Website Manager=-

Joined: Sun Dec 01, 2002 6:14 am
Posts: 1385
Location: California, USA
Yeah... I think a bonus track is going to happen for sure. Even if I do decide to redo the entire thing... so that way people who don't want to have fake AI Hammond as part of the program can have that as just a bonus. But yeah... I still am tempted to also redo the whole thing, which just... makes me tired thinking about it LOL. But I put up a poll on Twitter now to decide its fate.

https://twitter.com/jurassicmemoirs/sta ... 29699?s=20

I am kind of against having someone else do a reading honestly. That was kind of the problem with that "Annotated" version I had once done. Despite the other reader NOT meant to be Hammond himself, it all just felt like imitation Hammond... moreso than this literal imitation Hammond will be haha. At least this AI is literally made up of his voice, and no one else's.

_________________
Image


Top
 Profile  
 
PostPosted: Thu Feb 18, 2021 6:53 am 
Offline
T-Rex Killer
T-Rex Killer
User avatar

Joined: Thu Apr 24, 2003 11:20 pm
Posts: 12094
Location: Lima, Peru
Could this be made to work with Papagayo?

http://lostmarble.com/papagayo/

_________________
Visit The Carnivores Saga - a forum devoted to modding Action Forms' Carnivores, Carnivores 2 and Carnivores: Ice Age games
Tres WIP: updated T-Script Reference and File Formats documents
Sound name listings for the Demo (build 117), Retail (build 116), Beta 103, Beta 99, Beta 97, Beta 96, Build 55, PC Gamer Alpha (build 32) and E3 1998 Alpha (build 22) TPA files


Top
 Profile  
 
PostPosted: Thu Feb 18, 2021 9:10 am 
Offline
-=TresCom Website Manager=-
-=TresCom Website Manager=-
User avatar

Joined: Fri Jun 24, 2005 9:40 pm
Posts: 4379
Location: Sweden
This is pretty crazy. Funny thing is that this is the same guy that did the BeamNG levels ;)

_________________
"...there used to be more benches, but InGen's workers removed them during the evacuation in the name of framerate."

"The main laboratory and administrative buildings. This is where we made our work, where the real magic trick happen. When they are in need of height fixing, they'll come here." - Hammond


Top
 Profile  
 
PostPosted: Thu Feb 18, 2021 10:15 am 
Offline
Dinosaur egg
Dinosaur egg

Joined: Fri Jun 16, 2017 5:42 am
Posts: 7
For the sake of keeping everything together, this is the original training data I used to make this with:
https://drive.google.com/file/d/13Fzfr_ ... sp=sharing

It contains the audio files along with the original transcript txt files that are formatted for TacoTron 2 and a filelist txt file for waveglow training. The general idea of retaining the training data is that Neural Network AI technology is improving all the time. As the years go by, newer iterations of voice cloning technology will be made which will no doubt allow this to sound even better.


Top
 Profile  
 
PostPosted: Thu Feb 18, 2021 12:15 pm 
Offline
Stegosaurus
Stegosaurus

Joined: Sun Jul 15, 2012 6:40 am
Posts: 63
This has been on my mind for the longest time. I'm glad someone actually did it. With a bit of work and tweaking we might just be able to get something that could pass for a real voice recording of John Hammond. This is really great stuff


Top
 Profile  
 
PostPosted: Thu Feb 18, 2021 3:52 pm 
Offline
-=TresCom Website Manager=-
-=TresCom Website Manager=-

Joined: Sun Dec 01, 2002 6:14 am
Posts: 1385
Location: California, USA
james_uk wrote:
As the years go by, newer iterations of voice cloning technology will be made which will no doubt allow this to sound even better.


First of all, let me personally thank you for creating this! Not only did this shock me upon hearing of its existence, but the implications of it are astounding to me. The quote I am highlighting from above is an important one I am glad you have made, given that I am polling people on whether I should do a new version of Jurassic Time or just make a bonus track of Hammond's unrecorded dialogue from Trespasser. Your comment definitely has me leaning now on "bonus" for now just to have something, but should definitely wait on a whole new version for a couple years at least.

Which is fine by me... I have redone Hammond's Memoir so many times now from the ground up, and even just released a enhanced version with new videos last year... so I am not in a rush to go through this all over again LOL. I'm still curious to see how my question continues to poll, as it has been pretty split with people! But for now, I can pretty confidently say with this new info. I will just be doing a bonus track for the time being.

Anyway, James_UK, thank you again so much for this! You helped make an impossible dream come true! (In addition to hours of entertainment of making Hammond say funny things.)

_________________
Image


Top
 Profile  
 
PostPosted: Sat Feb 20, 2021 4:40 pm 
Offline
T-Rex Killer
T-Rex Killer
User avatar

Joined: Mon Dec 06, 2004 5:21 pm
Posts: 12888
Location: Portland, OR
This thread sounds very interesting... Unfortunately I am too mentally stressed to go through the hoops to actually hear the voice synthesis. I'm curious when someone might post some clips? Perhaps also recreating famous lines like "Welcome, to Jurassic Park".

_________________
T-Script Reference / Trespasser Millennium / JP Dark Secrets / Tres File Box / Tres Machinema & Dinosaur Toy Films


Top
 Profile  
 
PostPosted: Sat Feb 20, 2021 6:22 pm 
Offline
Stegosaurus
Stegosaurus

Joined: Sun Jul 15, 2012 6:40 am
Posts: 63
Draconisaurus wrote:
This thread sounds very interesting... Unfortunately I am too mentally stressed to go through the hoops to actually hear the voice synthesis. I'm curious when someone might post some clips? Perhaps also recreating famous lines like "Welcome, to Jurassic Park".

Well, there's that :D

Whoops , can't embed it due to age restriction but you can still see it on YT


Top
 Profile  
 
PostPosted: Sat Feb 20, 2021 10:21 pm 
Offline
Brachiosaurus
Brachiosaurus

Joined: Sun Jun 25, 2017 11:48 pm
Posts: 142
I wish I could use something like it for Anne voice overs hehe. I'm not much with script though.

Edit: Lol that's funny. Beware of rough language though.


Top
 Profile  
 
PostPosted: Sat Feb 20, 2021 10:52 pm 
Offline
T-Rex Killer
T-Rex Killer
User avatar

Joined: Mon Dec 06, 2004 5:21 pm
Posts: 12888
Location: Portland, OR
Oh my.........

_________________
T-Script Reference / Trespasser Millennium / JP Dark Secrets / Tres File Box / Tres Machinema & Dinosaur Toy Films


Top
 Profile  
 
PostPosted: Sun Feb 21, 2021 9:53 am 
Offline
-=TresCom Website Manager=-
-=TresCom Website Manager=-
User avatar

Joined: Fri Jun 24, 2005 9:40 pm
Posts: 4379
Location: Sweden
Teromen wrote:
I wish I could use something like it for Anne voice overs hehe. I'm not much with script though.

Edit: Lol that's funny. Beware of rough language though.


It would be cool with Anne. But I think we could just try and ask her on Twitter if she could do a quick video with some lines, maybe she would do it. She does remember Trespasser actually. :)

As I talked with Scall the other day. This is really cool, but be careful how to use it. It is an AI voice that sound like someone else, even if the person is dead. There can still be legal stuff with it, specially in the future. And worth remembering is that David is still alive and sound like Richard, so I'd say remember to mention if you use this in mods etc and it is not his real voice. :)

_________________
"...there used to be more benches, but InGen's workers removed them during the evacuation in the name of framerate."

"The main laboratory and administrative buildings. This is where we made our work, where the real magic trick happen. When they are in need of height fixing, they'll come here." - Hammond


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 22 posts ]  Go to page 1, 2  Next

All times are UTC


Who is online

Users browsing this forum: No registered users and 9 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  

Powered by phpBB® Forum Software © phpBB Group