Max comment length: 0/4096

Show Emotes mikuwelcome mikuhappy rinthumbs rinheart lenno lensorry lenthumbs lenshock lukaclap lukasparkle kaitocheer kaitolick meikolol meikojoy meikoangry lukaspin yukarietto
File / Embed / Poll / Tegaki
Drag files here to upload or
click here to select them, or paste an image

Verified — you can post.

Name the character shown in the picture.

Loading image…

(for deletion)
  • Please read the Rules and FAQ before posting.
  • Supported file types are: GIF, JPG, OGG, MP3, MP4, PNG, SWF, WEBM, WEBP.
  • Supported embeds are: YouTube, Nicovideo, Twitter/X, Bluesky, BiliBili, Vocaroo.
  • Maximum file size allowed is 20MB.
  • Images greater than 250x250 pixels will be thumbnailed.

We're aware about all the current bugs. If you find more please post about it on the /meta/ thread.
Expect them to be fixed within the coming days.


[Gallery] [Return] [Catalog] [Bottom]
Denne_Hib….webp - 204.68 KB (1600x1400) Download SauceNAO IQDB Yandex
This is a continuation of the ideas discussed in the Vipperloid thread. We're working on improving Hibiki Denne's voicebank and design. This is her redemption arc.
>>
So how do we improve her exactly? The first step would obviously be to fix the sample rate. But what comes after that? Adobe Enhance + RVC?
>>
To anon who's converting her samples - if it fixes the lag, I can go over her oto.ini again to make it better
>>
>>28948
Lag?
>>
noise_Trac….mp3 - 94.02 KB Download SauceNAO IQDB Yandex
>>28949
>>
>>28950
If you're referring to the noise that sounds like a fart then it's probably the sample itself that's the problem. Oddly enough I only hear it in the ones that end with "u".
>>
>>28951
what's funny is that it sometimes disappears if you pitch her up but might reappear if you do it too much lol
>>
flag compa….mp3 - 649.93 KB Download SauceNAO IQDB Yandex
Did some experimenting and I think her flag in the original song must have been gender -30 using the samples flagged with ↑. I wasn't sure since I only saw mentions of g-20 in the original creation thread. I'm aware the lyrics are probably wrong lol
>>
>>28948
I'm the one converting the samples, i don't think that's necessary
>>
hair.png - 86.09 KB (400x400) Download SauceNAO IQDB Yandex
my idea for the hair
>>
a1a59c4355….jpg - 48.84 KB (420x700) Download SauceNAO IQDB Yandex
83bee60367….jpg - 77.99 KB (736x1107) Download SauceNAO IQDB Yandex
If going for something that leans into the more wild part of the hair over the twin tails could do a sort of deathhawk like hair is what first came to mind seeing the hair.

>>28972
That is cute if using a fairly stylized art style.
>>
>>28972
It needs the dookie
>>
>>28951
Pretty sure that's for the shitting noises in the demo song's lyrics
>>
>>28972
Looks like two lightning bolts. Why not give her twin drills in the shape of poop since she's themed after literal shit.
>>
>>28975
I feel like Denne doesn't really have a punk style for this to work like it would on let's say Flower for example
>>
>>28977
That's intentional?
>>
>>28982
I'd assume so, probably. The VB was made because OP wanted a character for his (literal) shitpost lyrics.
>>
>>28981
Yeah maybe more toned down in a way that is cuter like messy hair.
>>
Hi, i fixed the sample rate of all the audios using ffmpeg and a batch script. I didn't denoise them.
https://files.catbox.moe/nmohol.zip
>>
hibiki_hif….mp3 - 290.82 KB Download SauceNAO IQDB Yandex
>>29001
Thank you so much! That seems to have fixed the fart noises produced with worldlineR.

I messed around with hifisampler and got her to sound surprisingly decent. The issues and how I solved them:
1. Sounds with "u", "e" were much quieter than the rest. Fixed this by enabling the P flag (aka "Normalize" in the expression menu). To use it, you must set wave_norm to true in the config.yaml file located in hifisampler folder. I set the value for everything to 80-100 to make her louder.
2. Some of the sounds ending in "E" and "U" are extremely noisy, to fix that create Hb flag, Range: 0 to 500, Default: 100. That's the flag for noise/breath control.
You can set it all the way down to 0 if the sound is very noisy, that's what I did on "e" in "koe". However, don't set it to 0 if the note contains a hissing/soft consonant, by setting it to 0 on "す" in "usubi" it completely removed the "s" sound. I left it at 25 for that one and it's still a bit muffled. Most sounds are going to be good with the default of 100.
3. Change default value for gender (if you're using OpenUtau) from -100/100 to -600/600. It also works in reverse, so negative will make the voice sound more masculine instead. I used gender 500 for this.

There is still hum present, but I think she sounds miles better. This is without denoising, which probably could improve it even further and make her easier to use with hifisampler. At least she sounds good enough with the default resampler now. Gonna have to look into denoising the samples properly, I heard the default noise removal filter in audacity is considered really bad. I'm still holding out hope someone will run her samples through Adobe Podcast though...
>>
>>29015
if it's just a matter of feeding these >>29001 through then i can try using it
>>
>>29015
she already sounds great with the sample rate change wow
>>
>>29016
adobe podcast de-noising is very hit-and-miss so far (example: does nothing with the noises in い↑). would UVR be better? i've used it in the past for RVC datasets with generally okay results
>>
>>29020
Possibly. I've only used UVR to extract instrumentals so i don't know how well it does in that regard. Some files need manual denoising, for example, there are some files that end with that TV hum noise and you gotta add a fadeout effect manually.
>>
ClipboardI….png - 52.61 KB (576x295) Download SauceNAO IQDB Yandex
Maybe the dookie can just be a bun
>>
>>29039
I think that's how it already was in the original though
>>
>>29041
I unironically thought she had some super saiyan shit going on desu
>>
my-recent-….jpg - 50.79 KB (640x829) Download SauceNAO IQDB Yandex
How about getting rid of the twintails (controversial, ik) and giving her a hairstyle like this except making the bun look more like uh, yknow, shit.
>>
images (1).jpg - 28.13 KB (295x400) Download SauceNAO IQDB Yandex
images.jpg - 39.49 KB (490x325) Download SauceNAO IQDB Yandex
>>29046
>>29041
Her hair was supposed to be this gyaru hairstyle, but made swirly to look more like poop.
https://dic.pixiv.net/a/昇天ペガサスMIX盛り
Imo the poop doesn't have to be on her head lol. If we keep the preschooler theme, she can have a poop charm on her bag or as her badge.
>>
>>29048
Ah, no wonder why people tend to redesign her as a gyaru
>>
Honestly the "poop" hair looks more like fire
>>
>>29048
This looks a lot nicer. I think the issue with the original is that her artwork's just not that great.
>>
>>28972
The art style here reminds me of Sakamoto's Zundamon
>>
ClipboardI….png - 10.54 KB (347x412) Download SauceNAO IQDB Yandex
>>29046
something like this?
>>
ClipboardI….png - 8.86 KB (275x356) Download SauceNAO IQDB Yandex
>>29091
or this
>>
>>29092
i really like this one
>>
>>29091
this one is cute
>>
>>29092
This one is perfect but i'd change the bangs because it doesn't look much like Denne
>>
>>29092
If it were a fully braided bun it could have layers to it.
>>
>>29092
What if you add back the twintails and make them super short?
>>
Before denoising the audio we should normalize it so it doesn't sound so quiet
You can technically do both things with Izotope RX Pro and I've tried but when i tried to batch export it kept crashing.
>>
>>29102
someone suggested uvr earlier, i dont know if you can batch export in that
>>
>>29092
I love this design actually
>>
>>29015
you actually made her sound feminine
>>
>>29100
She has twintails there though
>>
Bump
>>
>>29092
Maybe we should go with this design
>>
evil zako.png - 112.81 KB (376x790) Download SauceNAO IQDB Yandex
where are the drawbros idk what i'm doing
>>
>>29247
kisekae?
>>
>>29248
lol no symmetry tool
>>
ClipboardI….png - 12.40 KB (513x331) Download SauceNAO IQDB Yandex
>>29247
here's my shitty (no pun intended) concept for her outfit. its basically the same as teto's but with buttons and i tried to make her bra more visible. also her arm tattoo is the number 2, you can probably guess why.
>>
>>29253
>also her arm tattoo is the number 2
kek That's genius
>>
>>29253
If your idea is to make the bra more visible you could make her jacket look partially removed and exposing one shoulder
>>
>>29247
It's cute but it kinda feels like im looking at a different character
>>
Thoughts on a Hibiki Denne diffsinger?
>>
>>29247
I was thinking having socks piled up tiered like two turds
>>
Currently attempting something stupid. I'm normalizing all the files with a batch file then running them through BL Denoiser (a FL Studio plugin), then exporting them again.
>>
>>29268
Success!
>>
Alright here's all the normalized + denoised audios. Unfortunately some of them sound like they were recorded on a shitty 2000s laptop and others sound like you're listening to static due to the poor quality.
https://litter.catbox.moe/t12u0q.zip
(catbox didnt let me upload this to the main site for some reason so you're gonna have it to download it within 3 days off of litterbox before its gone)
>>
>>29272
the audios sound better on their own but in utau itself she sounds like she's dying
>>
>>29273
I have no idea why its doing that
>>
ClipboardI….png - 14.67 KB (524x369) Download SauceNAO IQDB Yandex
>>29255
>>
ClipboardI….png - 14.58 KB (524x381) Download SauceNAO IQDB Yandex
>>29275
needs more tummy
>>
>>29277
If you're gonna add a zipper you should probably remove the buttons lol
>>
>>29253
this is kind of basic, doesn't even look like a kindergarten uniform anymore, just standard crypton and maybe we shouldn't be changing her number
>>
8mb.video-….mp4 - 5672.99 KB (1280x720) Download SauceNAO IQDB Yandex
>>29273
Try hifisampler with Hb flag, here's before/after
>>
>>29284
I was curious and put 139 into a gorawase generator and one way to read it is
1 会陰 39 詐欺
Perinium fraud lol
>>
>>29284
Agree. We should keep the sleeves though.
>>
>>29292
Sounds fine but tbh i can't even remember what the original sounds like anymore lol
>>
Oekaki.png - 11.56 KB (500x250) Download SauceNAO IQDB Yandex
>>29299
I was thinking of a shorter sleeve like only starting from mid forearm or a sort of large bracelet. My very crude rendering....be kind I am finger drawing on a phone. Drawn pink here just to be less on the nose
>>
ClipboardI….png - 10.99 KB (354x311) Download SauceNAO IQDB Yandex
>>29304
like this?
>>
>>29309
Nah I was going for making it turd pile shaped thus my comment on why I drew it pink lol
>>
denne2.jpg - 285.25 KB (768x1264) Download SauceNAO IQDB Yandex
my 2nd attempt
the inspiration for this one is blue archive kindergarten outfits, socks and shoes like miku v6, the goddamn belt is from red chimera sv
put the frills from her bra to her sleeves and shirt + detached sleeves to show off her arm tattoo, and she's wearing a backpack with poop charm but i ain't drawing allat so take the bad photobash. the chest area is all warped because i used liquify to make her boobs bigger lol
>>
>>29430
I feel like the blue clashes a little with the orange and the red badge thingy doesn't help. But other than that I really like this design a lot. It reminds me of Voiceroid's Ai/Shouta duo.
>>
>>29284
Does she even have a number?
2 fits her better anyway cause a shit, a dookie, is called a "number 2"
>>
>>29430
It's cute but maybe you should make the outfit yellow instead of blue
>>
>>29471
Not that I'm aware of
>>
>>29430
grown ass woman dressing like this
>>
b20h-cost….webp - 54.31 KB (600x600) Download SauceNAO IQDB Yandex
>>29475
>>29470
traditionally it's blue or pink tho, maybe pink then to still keep it recognizable or change the skirt color? i'll play around with it
>>29476
139
>>29479
tbf she's one of the younger ones lol
>>
>>29430
Even though I like how she looks I feel like for what she is she needs to be more about the hips, ass and thighs than tits logically speaking

>>29470
The colors feel right to me and dont clash. I like the blue and orange.

>>29471
I dont know if "number 2" exists as a Japanese euphemism just English and American specifically likely. Code number 139 so if she does have a number on her I assume it would be that which is why I posted this observation how 139 is accidentally funny >>29293
>>
variants.jpg - 1501.04 KB (5344x1264) Download SauceNAO IQDB Yandex
>>29430
I did more variants, i can release the psd later if someone wants to play around with colors more, i just need to clean it up first
>>
>>29483
Pink ones look best
>>
>>29486
orange or blue skirt?
>>
>>29482
>she needs to be more about the hips, ass and thighs than tits logically speaking
>ass
i see what you did there
>>
>>29487
nta but yellow
>>
>>29495
err, orange...
>>
Is it known why Hibiki Denne looks the way she does? Like what style were the VIPPERs even going for, a gyaru?
>>
>>29497
The canon is
>Kindergarten clothes (summer, wide neckline)
>>
>>29498
Okay the new outfit makes more sense now. Also I like the blue shirt/blue skirt design the best.
>>
I like the idea of the hat hiding the shit hair
>>
>>29482
>I dont know if "number 2" exists as a Japanese euphemism just English and American specifically likely.
I thought about it some more and I think it's fine since she's getting updated by western users. Or maybe her english voicebank can have the number 2 instead (that's assuming we're gonna make a diffsinger)

Speaking of which who's going to train the RVC? I can work on utau renders for that
>>
>>29519
Idk about japan but where i live "number 2" also means taking a shit but it's not used commonly
>>
>>29520
I found this
https://detail.chiebukuro.yahoo.co.jp/qa/question_detail/q1369615218
>>
ilove_New ….mp3 - 3286.73 KB Download SauceNAO IQDB Yandex
I tried using mostly her middle pitch with updated samples from >>29272, sadly she's still super noisy and I had to apply denoising flag manually to a lot of notes.
>>29274
Could it be your software exported it in the wrong sample rate again? Because when I listen to her samples, they don't sound that noisy to me.

She kind of sounds like a raspier Teto to me when I use her middle pitch lol.
>>
>>29519
>Speaking of which who's going to train the RVC?
Eh i can give it a shot but i can't promise i'll do the UTAU->RVC renders myself. I'll see if i can find something to train the RVC, heard Colab has limits now.
>>
>>29527
It's ok, we're still on the improving voicebank part before we even train anything
>>
I'll explain the method i'm using right now. I've merged the 504 audios in those three folders into a single file and i'll be using that to train the RVC. After that I think we can either re-record the 504 audios from scratch which will probably take a lot of time or just make a monopitch CV or even a CVVC/VCV which will take less time.

Side note but while looking for info on her i found out that she has no fanart on pixiv, absolutely nothing, which is pretty shocking for a VIPPERLOID. Even Tabane Roze who's pretty recent has at least something.
>>
>>29525
oh god it is a raspier teto
>>
>>29530
Are you going to train locally?
I was actually thinking of doing it a different way: RVC trained on utau renders with the gender flag applied > apply it to one of the public Japanese datasets (if the TOS allows it) > train a diffsinger
Or skip the RVC step completely and train a DS with renders
>>
>>29533
No I'm gonna use Colab if I can find one that works.
>>
>>29525
Sounds like a chain smoking kid

>>29487
I still like orange skirt with the blue shirt. White shirt looks too plain and pink shirt looks bad with the orange because conflicting vibes it gives me.
>>
Alright i couldn't get the rvc colab to work so have the training data
https://files.catbox.moe/el522u.wav
>>
rvc compar….mp3 - 910.20 KB Download SauceNAO IQDB Yandex
>>29537
i fucked around with RVC, i don't think training it on all samples is the way
1st one is trained on the file you sent, 2nd is just 1 pitch (from >>29272), both are 100 epochs
i think 2nd one sounds much more like our guy, but both are slurring words
>>
>>29550
even after denoising it still sounds really weird wow.
>>
>>29551
it just sounds like it was recorded on a laptop mic ngl
>>
>>29550
I think we should use it regardless
>>
>>29530
The Vippers were allegedly ashamed of her so maybe that's why
>>
>>29555
the first one?
>>
ClipboardI….png - 3575.21 KB (3707x3508) Download SauceNAO IQDB Yandex
>>29557
I feel like Denne might've been one of those things that didn't really catch on within the community and was just randomly discovered one day by a gaijin checking the 2ch archives, which would explain why almost every page about her is in English but you can barely find anything in Japanese. In fact the only person out there who seems to know more about Denne than we do is this MMDLOID guy that keeps coming up in my searches so I'm almost 50% certain that they're who "discovered" her and the reason we know about her existence in the first place.

>>29559
We should go with 2nd, it sounds closer to the original.
>>
>>29560
Just checked and both her VocaDB and UTAU Wiki pages were created in 2023. That explains a lot.
>>
https://youtube.com/watch?v=XChpBYry1fI YouTube
>>29560
>>29561
>MMDLOID
Yeah he's also the guy that made this vcv voicebank that sounds like an rvc
that's why I'm hesistant, you can hear the pronunciation issues
>>29560
>We should go with 2nd, it sounds closer to the original.
I'm actually training a third model right now, let's see if it sounds better or not
>>
Am I the only one who feels more prone to use the Japanese name order? Denne Hibiki sounds better than Hibiki Denne to me.
>>
>>29567
Wait her name is Denne Hibiki and I've been reading it wrong all this time?
>>
>>29567
Yeah it's Denne Hibiki. I blame the Utau Wiki for making Hibiki Denne more common.
>>
>>29589
Meant to quote the anon above mb
>>
>>29560
Yeah I'm pretty sure she was a one off thing that was supposed to be forgotten the next day but someone just rediscovered her while looking through the archives.

I had never heard the name Hibiki Denne or Denne Hibiki before the 2020's and I've been invested in the whole Vipperloid thing for years, I find them fascinating.
>>
>>29612
That being said I wonder if some 4chan equivalent to her exists. I'd be surprised if 4chan of all places has never created an UTAU before.

Technically we are a 4chan offshoot so I guess the leekas would count, but I mean within 4chan itself.
>>
>>29612
I'm glad she was found tho
Her concept is really fucking dumb in a positive way
Makes me wonder if we'll ever have our own Hibiki that we'd rather forget about at some point
>>
>>29619
there actually was a joke voicebank we were making that was a maid robot guy with big boobs and no one liked but the original thread got deleted back when the site was having database issues and it kinda got memory holed.

maybe we could bring it back, maybe not.
>>
>>29620
I rember. I think he deserves a chance so maybe.
>>
>>29613
I doubt it, they usually were the ones to fall for 2ch's bait
>>
>>29624
seconding. even if its literally "draw a girl and call it a guy" we really need another male voicebank.
>>
>>29624
thirding
>>
I think it should have its own separate thread. Also we should start from scratch with a new design.
>>
>>29272
>Alright here's all the normalized + denoised audios.
https://files.catbox.moe/0brafe.zip
reuploded
>>29001
can you run these through the script again please?
>>
>>29658
Aren't they already in the correct sample rate?
>>
Everybody makes her sing in a higher pitch. I prefer my Hibiki sounding like an average Japanese dude, we are not the same.
>>
Bump
>>
>>29717
no they're back to farting mess
>>
>>29895
I'll check the sample rate in a bit but I'm pretty sure its all in 44100hz already
>>
>>29895
>week 2
>shitloid is still farting
>>
Ironic how the "scatloid" literally has farting noises in its bank even though they're completely unintentional and caused by a sample rate bug
>>
>>29550
I tried denoising the training data even further using a combination of Elevenlabs Vocal Isolation + Adobe Podcast + FL Studio plugins. It still doesn't sound great but it should sound better. Try training with this one instead and see how it sounds.

Also i don't know what your input audio sounds like but i think it should be within the vocal range of Denne.

https://files.catbox.moe/eyduve.mp3
>>
>>29911
Checked the sample rates of all the files in the three folders using another batch script and all of them are in the correct rate which is 44100hz. I think this voicebank is just cursed.
>>
>>30011
Another thing worth mentioning is that if you're a girl and want to use yourself as the input data you can lower the pitch within Replay itself (or whatever RVC GUI you're using) to make the output sound more masculine.
>>
>>30013
Huh. I saved the "a" sample in 44.1k 16-bit mono and it fixed the fart noise, so I thought that's what it was. I don't get it.
>>30011
>>30016
Thank you. It kinda sounds weird in places like it changed the pronunciation, but I will cut it into pitches and train the test models. Idk what to use as the input tbh. The anon who trained Sai used Synth V to render her samples, but idk if that's against TOS and I don't have it, using someone's utau feels wrong too.
>>
>>30019
You could just use your own voice. It's not gonna sound like you.
>>
>>30020
It's not gonna sound Japanese though lol
>>
>>30021
Fair
>>
>>29483
looks a bit too much like tsukiyomi ai/shouta imo
>>
>>30038
from voiceroid
>>
>>30038
it's a default japanese kindergarten uniform
>>
I feel like Hibiki is extremely hard to redesign without changing most things about her.
>>
>>30054
I think we should get rid of the weird swirly hair thing and just make her twintails swirly instead.
>>
Hibiki mentioned by Canon in 2012
https://x.com/canon_73/status/217785893245812736 [Embed]
Show the video in this Twitter/X post
>>
>>30057
I wonder if Canon knows the voice provider for Hibiki. Probably not, but she does have a friendship with the other VIPPERLOID VPs.
>>
>>30073
who's going to dm her and ask
>>
>>30074
I believe Figamin and Puzzle might be in contact with her, or perhaps it was the old owner that knew her
>>
>>30075
that was last year before T39 existed. so unfortunately, we're not in contact with her. but as far as i know she responds to anyone who has an inquiry.

i'm not sure if this is worth contacting her over though.
>>
>>30076
Mod-san is right, don't contact her just to ask for Hibiki's voice provider. It's a stupid idea.
>>
>>30056
it's in her character profile
>>
hibiki vcv….mp3 - 648.98 KB Download SauceNAO IQDB Yandex
Tried to do what >>30020 suggested, recorded with real time RVC conversion, maybe it is doable, I need to test more
Sample and renders with hifisampler
>>
IT'S PERFE….png - 359.25 KB (640x480) Download SauceNAO IQDB Yandex
>>30149
IT'S PERFECTO
>>
>>30149
What are we going to call this Hibiki Denne voicebank, and more importantly, how are we gonna get permission?
>>
>>30161
>how are we gonna get permission
we can't
>>
>>30163
yeah this is all just gambling on s_kesai and/or VP anon not coming back online
>>
>>30187
i wonder if anyone from VIP is in contact with him still.
>>
>>30188
The Hibiki account is mutuals with one guy but idk if that means anything
>>
>>30149
She sounds great anon. This is the best I've heard her sound in fact
Glad you guys are still working on this

>>30193
How do you know this? I thought he was banned from Twitter
>>
>>30221
He's referring to the @dennehibiki Twitter which was used to promote the voicebank in 2012
>>
is her first name hibiki or is it denne?
>>
>>30330
It's Hibiki so Hibiki Denne is technically the English spelling
I prefer Denne Hibiki though
>>
visualizat….png - 269.40 KB (2251x933) Download SauceNAO IQDB Yandex
>>30221
Thank you
I'm planning on going through her samples again and try a different approach to improving them. I wanna try manually denoising the samples instead of in bulk since they all have varying amount of noise, I just think it might be better. I also saw this >>30088 and they're using AudioSR, so I was gonna see if I can use that to make it better.
Then I'd train an RVC model for each pitch again and try to record a VCV/CVVC. Recording it is hard since I'm not good at Japanese + RVC does this weird thing where it misinterprets the sounds ("n" becomes "u" and so on), so 1 sample needs multiple retakes. It might be just because of how little data there is.
I was also thinking of releasing the legacy CV version as well if the samples come out good. So yeah, that's the plan for now
>>
>>30341
doing god's work anon. i respect the amount of patience you have. i could not do this manually.
>>
What happens if you pitch slightly different layered info an overtone and delete the difference? Is there a way to do that? I assume this could sound awful or great lol
>>
>>30341
>manually
good luck anon you'll need it
>>
>>30341
how's progress anon?
>>
>>30341
Do you have any examples of this AudioSR thing? I'm curious to hear how the audio restoration sounds.
>>
https://youtube.com/watch?v=qbH_6MqKYB0 YouTube
>>31303
not anon but here. i believe this uses audiosr.
>>
>>31308
Here's non vocaloid examples btw
https://audioldm.github.io/audiosr/
>>
Bumping for interest
>>
>>31298
>>31637
I denoised the audios, but couldn't get audiosr to work lol
Might have to train without it
>>
>>31640
Is it sample rate issues again?
>>
>>31642
No, anon said audioSR is the issue
>>
Trained 1/3 pitches
>>
>>31757
Good luck anon
>>
>>31757
Ganbatte
>>
>>31757
Are you doing three separate RVC models?
>>
>>31796
nta but i think earlier they said they were going to train 3 RVCs on each individual folder and make a voicebank out of it
>>
Bump. What happened to this?.
>>
Imho the original design is already fine, it just has to be drawn by someone who can actually draw
>>
bump
>>
>>33771
This
>>
Bump

[Gallery] [Return] [Catalog] [Top] There are 184 posts made by people with files.

Delete post: [ File only ]

Quick Reply
0/4096
Drop files here, click to select, or paste

Verified — you can post.

Name the character shown in the picture.

Loading image…