Make soldier say stuff, Yas Forums

Deep fake voices are government only right now because Juden Peterson was so pissy about the deep fake for him and the lawsuit is pending

uhhh
vocaroo.com/5wa4KGf40xi

LOOK AT ME TEST SUBJECT I'M A PICKLE, I'M PICKLE GLADOS
old.vocaroo.com/i/s0n9YLhoaBG0

Attached: file.png (582x393, 185.42K)

I man I got excited when I saw this had mlp characters but it has no trixie lulamoon or tempest shadow so whats the point, those are the only characters with interesting voices

anyway as for soldier heres mine
snup.net/369g

put a d between the n and u

oops
vocaroo.com/nXdxlM21mop

vocaroo.com/eUfzjN2PvNV

This is going to be amazing for YTPs and porn, holy shit.

Attached: delicious2.gif (640x360, 431.49K)

>15 can do TF2 voices now
Oh my...
The Smexual revolution is upon us.

Attached: 1563046269123.png (944x849, 644.82K)

vocaroo.com/a7854WD7RcM

vocaroo.com/8IGZbrfvUiJ

vocaroo.com/aa9vEtG6Lqd

Attached: abd.jpg (647x497, 49.52K)

vocaroo.com/9yjVOam3W4Y

Attached: 1443613732591.png (759x718, 384.13K)

WE ARE NEARING FULL AI WAIFUS
vocaroo.com/2sUNBv09fHa

Attached: 1550314114627.gif (312x213, 972.92K)

vocaroo.com/mtXj9BOVOvH

vocaroo.com/fyT0Jh5MjlR

Damn, gonna have to learn Japanese if I don't want an english dub waifu.

vocaroo.com/tMFgrSxoETg

Attached: b0d.jpg (1751x1796, 147.33K)

That's a good point, do you reckon Japanese is less difficult to synthesize?

Vocaloids were a thing over 15 years ago, so I'd say yeah.

posting a good one
youtube.com/watch?v=W60LgWhTKPo

Attached: glados_wgw.jpg (1204x1204, 481.06K)

>pangus cupcake
Is this Dr Brule Soldier?

Check /mlp/'s voice thread, we have a bunch of different voice models and there is a tutorial for making and using them.
Pretty sure Trixie exists, it's just not as "plug in and play" as 15's website.

To put it simply the language has significantly less sounds.
Meanwhile English has shit like "Ghoti" which is pronounced like the word "Fish."

vocaroo.com/5d0CGd7MfW3

vocaroo.com/l5RnLF44t1J

Attached: 1570472932509.png (512x512, 207.4K)

That sounds better than some voicefags.
It's fucking on now.

Attached: 1549308712070.png (430x473, 310.43K)

vocaroo.com/nSkyYbfpQsn
"suk a dik" turns into "suk the dik"

Attached: watt.gif (500x281, 91.5K)

>Deep fake voices are government only right now because Juden Peterson was so pissy about the deep fake for him and the lawsuit is pending
but they're not though?
youtube.com/watch?v=IM3Q6wfdD9Q
youtube.com/watch?v=_citG7kWBnI
youtube.com/watch?v=aPPaChTRbCA
youtube.com/watch?v=XETSTdHP6GQ
youtube.com/watch?v=6OXAMY1PQr0
youtube.com/watch?v=G5nb2oyiw8Q
youtube.com/watch?v=ExUyyfT-DdM
youtube.com/watch?v=hDVuh4A-q3Q
youtube.com/watch?v=BkH2zEHJ2Ng
Here's a link for this exact software:
github.com/CorentinJ/Real-Time-Voice-Cloning

Absolutely. There was a unfinished Four Swords battle game, and you could pick any name you wanted and the announcer would still say your name because of the way Japanese's phonetics works. At least this is from what I heard.

It's really not that great

vocaroo.com/iv8dCGTGlz0

Attached: Bloody Jack.gif (600x252, 2.33M)

I will have to check that out thanks
I would love to see Kathleen Barr do the trixie voice for some sort of boisterous villain in bayo 3

It’s really hard to make anything with soldier.
After adding more than 3 words the whole thing turns into a garbled mess.
However the word “America” sounds clear as day wherever you put it

vocaroo.com/mmZC5zoctIY

Honestly it's more fitting for him to say it like that.

What tricks are you using to make it sound like that?
That's nearly indistinguishable from human speech.

Attached: 1569010808460.gif (223x222, 147.07K)

Is that supposed to be the shy horse?

Attached: 1584634795157.jpg (400x400, 25.49K)

10/10 friend

Attached: ha.gif (300x237, 2.66M)

vocaroo.com/cglOFFsKVNA

Attached: 1444914579366.webm (640x480, 93.25K)

voca.ro/4EZLk5vUBNK

Attached: me.jpg (589x800, 542.04K)

Dunno, just typed what felt natural for that voice.

Attached: anigif_enhanced-buzz-30614-1412787156-6.gif (400x225, 3.4M)

vocaroo.com/4WM7bKOyoFN

voca.ro/aouXpUcrexi

vocaroo.com/ghtybWx3Q1j

This is amazing.
vocaroo.com/6btexh5p5Od

Attached: 1550871093600.png (545x434, 268.27K)

vocaroo.com/c61LiqFzny6

Attached: 1547388879222.gif (219x199, 1.12M)

the sentence mixing people make from the 15 lines soldier says is better and more natural than this shit lmao

Other characters are good, though. Especially Gladass and Flutterschwein.

>i love busy sea

It's BZC, faglord

your a bzc, dumbo

>he nails half of screaming eagles shout
good enough

vocaroo.com/58wS8K2rH9v

If you plan on making your own voices, most of the work needed is gathering "clean" voice samples (no background noise) and transcription. Clean audio is also of utmost importance as background effects and even noise reduction affect the quality of the resulting voice. Not the guy who posted it, and I haven't tried it yet, but this should make things not as tedious for the second bit.
file.house/5TjW.zip
Make SURE the output file has semicolons at the end of each line for the next part, otherwise it won't work right. Using find and replace to add ; to the end of each line should work, making sure the final line has one too.
After that, you'd need to download this to convert the transcriptions into ARPAbet. Instructions are included.
github.com/AmoArt/ARPAbet-dialogue-converter
Take the ARPAbet translations, copy/cut them from the output file, and paste them starting on a new line under the last transcribed line, and save it. Then copy the file with combined transcriptions, paste it, and delete every other transcribed line besides 1/6 of them (including their ARPAbet transcriptions for the lines you saved), saving it as something like (charactername)-verify.
Remember the most important rule: the voice you choose NEEDS to have a good amount of audio in total minutes. It depends on how much inflection in their voice there is, but generally speaking you need about an hour or more of audio.

vocaroo.com/kjndfv0WRBp

Attached: 1571838544591.jpg (200x200, 17.15K)

How come the Mlp voices are the best available? I take it that there are more voice samples compared to tf2