#using gimp a bit more than usual

[eeldritchblast]

Making Baldur’s Gate 3 GIFS Tutorial

TOOLS USED:

Nvidia or alternate game video recording software

PotPlayer

GIMP (Optional)

Adobe Photoshop

PART 1: Recording

I personally use the Nvidia recording software that came with my computer to capture videos of the game, but there are plenty of video recording options out there. I know OBS Studio is a popular one, for example. Whatever your choice may be, here are three important tips for when capturing video:

Use the least amount of compression settings your recording software can output, and I recommend at least 30FPS, though 60 is better

Press F10 to hide the Baldur’s Gate 3 UI and capture the entire screen

Give about three seconds worth of video before and after the section you want to record, as safety buffers, if you can

PART 2: Frame Extraction

 Open the video you’ve recorded in PotPlayer in window mode (not full screen), and pause a little bit before where you want to start your GIF. Press Ctrl + G to open up the Consecutive Image Capturer window, and drag it to the side of the video player. These are the settings I recommend in general:

HOWEVER, I suggest playing around with the “Every # frame” option to your own liking. I have a pretty high framerate for this video I’m using, so 5 is fit for me, but if you have a lower framerate, you may want a lower number, like 2.

With your settings in place, click “Start” and unpause the video. Pause the video again after you’ve captured all the frames you want, then click “Stop”. If you’re making a gifset, I recommend then storing each set of frames in a single folder at a time.

PART 3: Resizing

I personally use GIMP to resize/crop the images because I find it easier, then export the image as a PSD for the next steps in Photoshop. But you can use Photoshop for this step too, if you’re more comfortable with that.

The number of frames you can get away with in a single GIF for Tumblr depends on the aspect ratio you plan on saving it as—for example, 270x200 obviously can handle more frames than 540x300 while still fitting within Tumblr’s size limit.

PART 4: Sharpening

In Photoshop, use the Filter > Sharpen > Smart Sharpen tool on each layer. The settings are your choice, but I usually use 200 for the Amount and 0.5 for the Radius.

PART 5: Animation

If you don’t yet have the timeline set up in Photoshop, click on Windows > Timeline to bring it up. Then from the arrow, choose “Create Frame Animation” and click on the button to do so. This will insert the first frame into your timeline. Click on “Make Frames from Layers” in the timeline options menu to add them all.

Click on frame 1, then hold down Shift and click on the last frame to select them all. Then click on the drop-down arrow from the last frame and set the delay between each frame changing. I usually use 0.1 seconds.

Change the “Once” option to “Forever” to make the GIF repeat itself ongoing.

PART 6: Editing/Colouring

Click on frame 1. Then click on the top layer, while still keeping frame 1 visible.

You can use a variety of tools in Photoshop to editing the colours, brightness, contrast, etc. of the gif. Make sure the filters are the top layers to apply them to all the frames. If you test out your gif and see that the filters are not applied to all the frames, toggle the visible option on them off and on again.

I’m not going to go into detail of how to edit your GIF here, because each one is different. But I will say the tools I most typically use are Brightness/Contrast and Curves.

PART 7: Export

Click on File > Save for Web. These are the settings I like to use:

Congratulations, you’ve made a GIF!

#bg3 #baldur's gate 3 #tutorial #gif tutorial

[soundbluster]

for a while since i got into transformers 3 years ago and i loved the franchies instiantly even since i was a kid especially the designs of transformers and Seeing fan arts and fan Designs and fans make thire own continuity made me me want to start learning to draw when im ready but i don't know how to draw cybertronians so do you have any tips and advices for someone who want to learn to draw transformers characters both Traditional & digital art?

Biggest advice I can give is: "Break a character apart into simple components!"

Drawing bots is honestly not that difficult, I find it much easier than people because you can more easily break a character apart into simplified blocks.

When it comes to designing characters it can take a few times to get one you're happy with so doing lots of basic sketches can help. References are always useful too! Every artist uses references.

Here's how I do it (In this case I'm doing an alt design for tfe Nightshade):

(It's not a definitive guide as you kinda have to figure out what works best for you both in terms of technique and art style. I'm aware my art style is definitely not for everyone, as it veers more on the cartoony side. I've had comments about how my faces all look effed up etc XD)

Initial sketch

Break a bot down into basic shapes, circles and sausage shapes for more rounded characters and boxes for squarer ones. I normally start with the head and draw the rest of the body down from it (torso and arms then legs then any back kibble that might be visible) Think of it like the protoform beneath the armour.

To help with figuring out a pose I'll often use my toys. It can help you visualise where their arms and legs etc go along with whether part of a bot is actually visible from a certain angle. For example in this case:

(Yes that's my cat chilling in the background)

This can help you to figure out if a certain design can pull a given pose, for example would a bot with kibble on their hips or arms be able to move their legs or arms a certain way. I used POTP Elita one as a reference for my SG Megatron a lot for example as they've got very similar builds. But this is not essential.

(There are mannequin apps on mobiles etc. you can use to help with pose references too, and image searches are always useful for references.)

For more dynamic poses it helps to think about how the character is moving, so add a curve to the torso/spine to add to any implied motion etc.

Details

Once I'm happy with the basic shape and pose I typically add the armour over the top. Again the head is usually the first bit I do. You can see I got a bit lazy with their left hand and just did a scribble here XD

Inking

Normally when inking I do the edges of each armour piece first then add any details on afterwards in a finer pen. Generally I'll do one body part at the time eg. fully ink the arm before doing the head etc. Its best to make sure you do parts that are in front first!

Here's the finished picture:

Though it'll need some digital cleaning up at another time.

As for digital art...

I can't really give much advice on starting a drawing digitally as I never quite got the hang of it (colouring and corrections I do digitally but all my line art is traditional), I need the feedback of a pen in my hand and pressure on the paper to get it right.

These days the Hardware I mainly use is a refurbished huion art tablet (it was about £120 when I bought it) that's probably the cheapest your going to find a graphics tablet with a screen. BUT it's really not essential to get one especially if you're just starting out. Basic graphic tablets without a screen (just a pen and a pad) are much cheaper or you can use a mouse. I used a mouse for ages when I was first starting out.

In terms of software, for colouring and edits on the PC I use GIMP (I've been using it since like 2009 so I'm not really going to change any time soon!). It's freeware software and can do a lot of stuff, but it's user interface isn't the most user friendly especially if you're not familiar with graphics software but there are plenty of tutorials available for it. I'll try to remember to make a post showing how I use that another time :) as I'm not at my computer at the moment.

There are plenty of alternative image manipulation/graphics software to use as well, some free, some free but with ads, and some paid for (either one off payment or subscription). It's best to see what other artists have to say about them though as I've never really used them!

#my asks #my ask box #fan art #art stuff #how i draw #my ask is open #ask me stuff #nightshade #transformers nightshade #earthspark #transformers Earthspark #tfe #tfes #nightshade malto

[simbouquet]

TS3: Walk Cycle Edits

This mod edits a few of the locomotion animations in TS3 to get rid of the creepy perennial smile that forms on a Sim's face every time they go somewhere. 🚶

More info and download under the cut!

In the base game, an animation oversight that exclusively affects male Sims aged YA and up and all Elders causes them to have a constant smile drawn on their face while they're walking, even if their mood doesn't necessarily warrant it. All other Sims (YA+ females and Sims younger than teen) don't have any facial expressions baked into the animations, so they remain neutral when necessary while still being able to make any of the in-game facial expressions when walking.

The original version of this mod took care of this issue by eliminating the facial animations in the affected animations. EA never learned, apparently, and continued to bake in more forced smiles into their walk cycles in each Expansion Pack, so I've decided to hunt down even more locomotion animations and remove those perma-smiles once and for all. 🙂

This mod will completely neutralize the facial animations for the following walk cycles:

Base: Walk (YA+ males, all Elders)

Base: Ghost walk (All ages and genders)

Base: Ghost float (All ages and genders — not really locomotion, but an idle animation that I found to have a perma-smile, so neutralized it too)

Base: Wading (All ages and genders)

Supernatural: Fairy run (All ages, both male and female variants)

Supernatural: Flight of Felicity Fairy Gift (Teen and older Sims of any gender)

University Life: Backpack walk (Teen and older male Sims)

Into The Future: Plumbot walk

Into The Future: Plumbot hover

On my to-do list are the witch broom animations from Supernatural, suggested by chokolady at Mod The Sims, and the hoverboard animations from Into The Future. They're a bit of a bigger challenge for me because they do involve objects, so my initial attempts weren't good. I will update the mod with these animations next when I get them working. 🙏

Compatibility

This mod edits some animations from expansion packs, but you can still use it normally if you don't have them installed; you'll simply never see the animations in action.

This mod will conflict with any other mod that replaces the CLIP resources pertaining to the walk cycles listed above.

This mod is perfectly compatible with mods that edit other resources related to Sims' locomotion and walkstyles, such as Twoftmama's Walk Not Run or Fryesims' No Female Walk Style, which I used in the preview image to allow the female Sim to use the male walk animation and the male fairy to use the female fairy run animation.

Download: SFS / MTS

Place the .package file in your usual Packages folder or in your Overrides folder.

Additional Credits

s3pe: Exporting/Importing resources, creating the package file.

Blender 2.79, Clip Tools: Importing/Editing/Exporting animations.

Gimp: Creating the preview image.

Thanks to TheSweetSimmer for helping me figure out and get rid of the wonkiness with my edited animations. You rock!!!

Thanks to @nixcorvus09, @xiasimla and @thornowl for their feedback during testing!

Thanks to @thornowl's Jack for modeling the preview shot! 💘

Elder Tommy died of old age in between the original release of this mod and the update, woopsi...!

#mycc:ts3 #ts3:walkcycleedits

[sunsetsands]

Almud Masterpost

Seemed like a good idea to create a place to compile information about the main planet project I have going.

All of the pictures here are hand-drawn. If one day I ever become good or maybe even just passably okay at digital art, I may redo these in gimp or something, but until then, have these messy pencil illustrations.

The planet’s name is Almud (pronounced “awl-MOO-duh” (yes, the D at the end is its own syllable)). It is the second planet from its star, an orange dwarf. Conditionswise, Almud is very similar to Earth, just a lot warmer and wetter, and without a single large moon. Instead, it has a somewhat recently-formed system of rings. These rings are made of the debris from the planet’s former moon, which floated in past the Roche limit a few dozen million years ago and got torn apart. Almud may or may not also have a smaller moon or two somewhere further out. I haven’t decided on that yet.

This is a map of Almud’s entire geography, which is slightly outdated. I’ll probably make an updated version at some point eventually possibly maybe. If you’re curious, those numbers on the continents were so I could keep track of continental drift to make sure everything made sense. I care way too much about tiny details.

This is a rough (and I mean very rough. Not proud of my craftsmanship on the outlines here) approximation of what Almud’s surface looks like. The foliage uses a teal pigment to photosynthesize, and the sky appears pink during the day. Obviously, not all of the planet is wetlands, but there are definitely more wetlands than there are on Earth thanks to the much higher humidity.

I have spent a lot of time trying to figure out what Almud’s animal-equivalent life should be like. After several failed attempts, however, I think I have gotten it to a point I am satisfied with. Below is a phylogenetic tree of all of the “animal” phyla present on Almud, and an overview of what each phylum has going on. I tried not to rip off Earth's phylogeny too much, but there are some notable parallels.

Vaxistoma (roughly "vaccine mouth"): Small fishlike creatures that feed through a proboscis with an extendible needle-tooth-thing inside of it. The rest of their body is covered in thick, hard scales. They mainly inhabit deeper waters where aquatic duossei are less common.

Duosseus ("two skeletons"): The closest equivalent to vertebrates. First members were fishlike, with calcified plates covering the outside of their bodies and rod-shaped bones supporting the inside. The outer skeleton atrophied mostly in the terrestrial members, but most classes do still have notable remnants, as shown in the small drawing above.

Limosus ("muddy, slimy things"): Soft, squishy creatures without much in terms of an internal skeleton, but most groups do have some external armor like their relatives listed above. Can be accurately summarized as "molluscs, but more alieny", though a few members are more like worms or sea stars.

Jocomodivirae (very roughly "funny little guys"): Small velvet worm-esque invertebrates with a thick, leathery pad over their back. A very diverse phylum with many, many members. Definitely not just insects with no exoskeletons.

Planagelattae ("flat jellies"): What if flatworms had three eyes?

Xenigmalus ("strange, mysterious things"): I can't think of a good way of describing these, which is pretty fitting for what they are. Body plans vary wildly here, but are almost always some combination of fins, tentacles, and a big translucent sack. Like the vaxistomans, they usually inhabit deeper waters.

Cornivermia ("horned worms"): Pretty self-explanatory. The flat, hard bits at the fronts of their heads help them dig through softer areas of soil. Some groups use these growths instead as something more akin to pincers, fins, hooks, or shells.

Carniherbae ("meat plants"): You know those animal-fern things from the Ediacaran era? These are just those, but not extinct.

Vivitria ("living glass"): Soft, feathery insides protected by a crystalline silicate shell. Many species in this phylum are colonial, which tend to look like colorful, floating geodes. These colonies often have surprisingly complex sensory capabilities, and some have been found to be about as intelligent as Earth cats.

Xylovitria ("wood glass"): Terrestrial relatives of the vivitrians, almost all of which are colonial. The defining feature of this phylum, besides their terrestriality, is their symbiotic relationship with a wide range of plant-equivalent species. The xylovitrian colony forms a protective, glassy wall around the plants' branches, as well as a system of feathery roots beneath the soil which serve to both gather nutrients for their plant partner and exchange gametes with other colonies to create new, empty xylovitrians for the plants' seeds to land in. In return, the plant gives the colony some of the byproducts of its photosynthesis.

Chiforma ("X-shaped"): Four-sided radially symmetrical creatures. Contains such captivatingly creative groups such as "squids, but four", "clams, but four", "eels, but four", and, most creataculiciously of all, "coral, but four". A shining example of the innovation that specbio nerds are capable of.

Nodovellis ("tangled hair"): Formless, sessile filter feeders. Basically a slime mold trying really hard to be a sponge, but the closest it could get was becoming a loofah.

(Feel free to give critiques or advice on the scientific names I made. All I really did here was mess around with google translate. I know there are guidelines and policies for what is and isn't an acceptable phylum title, but I've never been able to understand what any of them mean. If anything immediately makes you go "That's not how that works!", let me know)

For some additional information, I imagine that life on Almud began in freshwater rather than saltwater. This made the transition to land pretty easy for most of the animals, since they could afford to just flop around in muddy wetlands without any risk of drying out. This does mean that their skin is very, very sensitive to salt, however.

There is one sapient species on Almud: These slug things. Their actual species name is Akada, if that's important. Akada are descended from a social burrowing species that learned how to cultivate the many plants and molds that thrived in the dark, wet conditions of their tunnels. They are herbivorous, have a herd animal-like social structure, and currently have a level of technological advancement similar to ours. For more miscellaneous and mostly jokey info on them, please look here.

I will expand on all of this when/if I find the time and motivation.

#art #pencil on paper #alien #slug #akada #almud #spec bio #speculative biology #spec evo #speculative evolution #xenobiology #creature design #drawing #sci fi #guess who just learned that the name I gave to this planet is also the name of an antiquated unit of measurement #good to know it's not #like #an obscure slur or something

[murfeelee]

Get to know you - Sims Style

Thanks @descendantdragfi here and @obscurus-noctem here for tagging me! ❤️

What’s your favorite Sims death? Every death is hilarious--fire and drowning are iconic, ofc. My fave sims GHOST, however, is the Haunted one from SN, and the Hologram one from ITF.

Alpha CC or Maxis Match? Alpha CC. I don't mind Maxis Match in TS3, but EFF NO in TS4--get that clayfied crap out of my face, alpha all the waaaaaay~!

Do you cheat your sims weight? YES, I use a weight gain mod.

Do you move objects? MOOOOOOOO! 🐄🐮

Favorite Mod? I juuuuust made a whole post drooling over my faves.

First Expansion/Game Pack/Stuff Pack? TS1 Makin Magic. Iconic.

Do you pronounce live mode like aLIVE or LIVing? aLIVE.

Who’s your favorite sim that you’ve made? My avatar sim, Sakura Rain, the time-traveling Japanese cherry-tree faery of the Unseelie Court. (She's inspired by all my fave pink-haired anime characters.)

Have you made a simself? Yes, but only when I play The Sims with my family members, and we take turns making each other and our home. It's so much fun.

Which is your favorite EA hair color? White. Especially after TS4 came out with no CASt or color wheel and I couldn't even recreate my main sims like wtf. I have a type, EA!

Favorite EA hair? PFFT! 🙄 I can't stand most of the hairs EA makes in general, especially not the ones for male sims, as most of mine have very long hair.

Favorite life stage? Young adult, by default. But as more mods have been made for kids (No Stretch, Child Magic, Toddler Interactions) I have been enjoying child sims FAR more than I used to.

Are you a builder or are you in it for the gameplay? I'm a builder--that's how I got into TS1 way back in the day. I like decorating interiors--my exteriors are usually sad, I suck at using CFE--and just cluttering up my lots. As a fandom-inspired storyteller, I rely on in-game interactions to give my sims the poses/scenes I need, so having deep and immersive gameplay that allows me to recreate whatever fandoms I want is key.

Are you a CC creator? Unfortunately.

Do you have any Simblr friends or a Sim Squad? A "sim squad"???

Do you have any sims merch? No, but I want one of those plumbbob headbands so I can dress up as a sim on Halloween. XD

Do you have a  Youtube for sims? Technically no. I got into making TS3 machinimas inspired by my favorite music a few years ago, and would upload them to Youtube cuz that's what everyone did, but omfg all the recent changes in copyright and music policies has made it SUCH a headache uploading ANYWHERE--including Tumblr!

Like, my most recent IWTV insp. machinima on Tumblr has that stupid Copyright Guidelines thumbnail on it, even though Tumblr ACCEPTED the video and let it go through--it plays! It's not removed, so wtf??? No matter how many times I tried reuploading it I got the same problem--this BS is more trouble than it's worth, tbh.

How has your “Sims style” changed throughout your years of playing? My gameplay will always look a bit pudding, but I think/hope I'm getting better at recreating different fandoms, and getting the level of detail down just right.

Who’s your favorite CC creator? It takes a village. There is no ONE creator--especially since all my OG faves retired a long time ago: Luna, @qingshuangtongzi, Hekate999, etc. And new simmers pop up everyday that blow my mind, so yeah, it's just the community at large, really. 💚💚💚

How long have you had Simblr? 2013--this July will make it a full decade. Tumblr's trash but it's also the absolute best, LOL.

How do you edit your pictures? I suck at editing--in GIMP I just tweak the "Curves" brightness & saturation when the in-game lighting is hot garbage, then call it a win.

What expansion/ gamepack is your favorite? (TS1) Makin Magic, (TS2) Apartment Life & Night Life are tied for witches & vampires, (TS3) Supernatural, (TS4) Vampires. Like, Supernatural changed my life, I seriously doubt I'd've kept playing TS3 as long as I have without SN. I also love World Adventures and Into the Future of course, but Supernatural gave us Witches, Faeries, Werewolves, (I'm not even gonna talk about TS3's busted Vampires like how the EFF were they "updated" or "fixed" when we STILL didn't get a functional coffin, like what frikkin vampire would sleep on a frikkin altar exposed to the sunlight like a effing BUFFOON, EA?!) and all kinds of fanciful content that I just can't live without. Hands down. 🧙🪄🧙‍♀️✨

________________________________________

Y'all know me--if y'all want to do this, too, just say I tagged you! <3

#questionnaire #simblr #the sims 3 #mini rant

[egg-emperor]

We all know eggman has an obsession with creating theme parks but what do you think would happen if the doctor made a strip club? Would it just be countless holograms and robots in his image, doing erotic dances?

I can see him having one even just personally as another private room in his living quarters, I feel it's what was missing in Adventure. I like to think he'd have an area for his more adult interests and entertainments, as well as a gaming room to indulge in his more "childish" interests like in the game. Maybe because he was on the move in his Egg Carrier it didn't seem like a good place to keep it, so the fun happened in his bedroom area and maybe his library had some dirty books to keep him entertained, and the very sus statues lol

I like to imagine that he gets creative with all the different ways his creations can entertain and pleasure him because the man has needs hehe. Wherever and however he makes it, it would definitely be sure to be themed to be all about himself first and foremost because he isn't attracted to anyone quite like he is to himself, and if it were his choice then absolutely everything in the world would be all about him! Though he also seems to have a thing for buff guys, so maybe he'd also throw that in there a little bit but he'd be sure to be the focus! XD

It would be a useful place to keep his sexbots stored and give them another purpose. He probably realized it wasn't a great idea to keep them amongst his other robots for battle in storage, in case Sonic and co would stumble across them and make it very embarrassing as it just reveals the things he's into just like that. (They'd never think of him the same way again if they were to those GIANT extensions he has attached to them hehe) A much better hidden area for that purpose alone that's off limits and soundproof would be much more practical and fun.

He has various robots that almost always share physical features with him in one way or another, though some may be built to look more buff as he's also into that. Same goes for holograms he might also add to show off the kinds of technology he's capable of beyond robotics and if the robots are out of service or in use, they'd suffice. And the idea of him going public with it could be interesting, as a way to bring in an audience in an adult way if his deadly theme parks can't, and to get them to finally acknowledge his sex appeal!

Some are given clothes just so they can strip as he's known to make outfits his Egg Pawns for example, some can dance erotically, and some can pole dance, the holograms can do all this visually too. Some robots can offer more private services behind the curtain. A lot are built to have more exaggerated features than his usual robots for battle, but some are repurposed battle robots that he adds a love mode to and attachable extensions when he's attracted to those too and he can get creative with various toys they can be equipped with.

Eggman himself is also good at swinging on poles clearly, sooo 👀

On rare occasions the real Eggman himself might show up to show off his skills with that and maybe even more if he's feeling naughty. He could show up wearing one of his leather daddy outfits, or he reveals that his full bodysuit can function as a gimp outfit. Maybe he will strip and show off and his beautiful titties, belly, butt and more. They'd be rarer special appearances so nobody outside of the place that day will believe anyone who says they saw the real Eggman in there, they'd just say it was probably just a convincing hologram XD

It'd also be interesting to think of how he could use it for evil if he opens one to the public because of course he always takes the opportunity and malicious intentions are behind it every time. Aside from another way to get his superiority and importance into their minds beyond propaganda by sex appeal which is already a scheme, it could serve as a distraction for people while he's out doing evil, which lessens the amount of people to stop him in the area. Wouldn't give away what he's into if he doesn't admit he has some for personal use ;)

#ns4w #asks #dr eggman #eggman #dr robotnik #dr. eggman #headcanons #my post

[abyssalaerlocke]

Durge of the Ice-Hells

Concept Durge (#steel durge) for a story I'm working on

I rarely do art, usually concept stuff for something I'm writing, so I have limited experience and no stylus or anything.

This was made in GIMP (free) mostly using the path tool — which is a pain with tapering brush dynamics, because each taper was a separate line, and needed the fade length adjusted.

I can't tell at this point if I bit off more than I can chew — didn't always know what I was doing, but I did my best. And it was really cool seeing the idea gradually come together 😃

#dark urge #dragonborn dark urge #white dragonborn #draconic bloodline sorcerer #dragonborn #dragonborn sorcerer #steel durge #my art

[morkofday]

before & after coloring tag

tagged by the amazing @snimeat ♥ thank you so much! also have to say that i love your very dark coloring and i admire you for how well you can pull that off ^^

tagging: @luna-lina @nanons @machikeita @badprophetvx @danyokz @liyazaki @dimpledpran @patprans ♥ no pressure but i am very curious and i love this tag a lot myself so i thought maybe you all would have fun with it too ^^

some thoughts/explanations and a list of the original posts under the cut!

pictures taken from these edits:

seanwhite + lang leav’s three questions x

pride month + fave thai drama characters x

patpran +  forugh farrokhzad’s on loving x

pingxie (reboot) + olivia gatwood’s the lover as a cult x

patpran’s love through the seasons x

pingxie (ultimate note) + joan didion x

bad buddy’s main six + sun & moon signs x

this tag is always super hard for me bc a) i don’t usually keep the gimp files and b) i obsess over “cleaning” my folders so i usually delete all “used” screenshots as soon as am done with my edit. i got lucky this time when i realized i had actually kept some of the shots used in these :’D

i tried to pick different colorings i've used but with such few options, these all became kind of similar. also my basic coloring is always kind of the same. but i’ll still blabber about each of these a bit:

the seanwhite one has a lot of red/magenda bc of the color scheme i used in the edit. it’s also a lot darker than the original picture bc i wanted the layers to show up some more. i also tried bringing in more warmth by adding more of yellow to their skin etc. it’s a very basic thing i do for pictures in general. there’s also the added noise i love using for edits to bring that rougher feeling to the images.

for both of the pride month post’s pictures i tried picking one color and switching that into the color i had chosen for each character. for shin i’ve turned green into red and for khai i remember turning some of the background blue towards green. so these two show that process more than just simple coloring.

this edit’s main colors ended up being a soft coral pink and a pastel orange. in this picture i’ve tried taking away some of the green and yellow by adding more magenda, red, and blue. i don’t think i’ve done anything more complicated for this. there’s also the added noise once again to bring in more texture.

reboot is very... green originally. so i struggled a bit when i was trying to take some of that away while still trying to keep the green in the background bc i needed that. zql got more blue tones this way (his hoodie). also this pic obviously has an added texture layer which makes the colors brighter/more muted plus brings in the folded paper effect.

the seasons one is my personal favorite when i think about my patpran edits. it was so much fun to do and i love the final edit to bits. it’s very similar to the pride month edit with how i used black and white pictures plus colored pictures. in this one i just kept pran’s skin color (which i made a lot brighter and added more red and yellow) and then turned everything else towards pink.

this pingxie edit has very muted colors and i also wanted to lean more towards blue bc the second picture is Very Blue originally. so i took away most of the green and yellow, added more blue and magenda, and made the picture a lot darker. i think the color change has happened mostly through the hue and saturation options.

all the pictures in the sun & moon signs edit also follow the same pattern i’ve introduced before: make all the colors turn towards the one desired. on the left image i had to work on the background colors separately to make them red without losing wai’s skin color and on the right i got lucky with his bright blue shirt which i could easily turn red. otherwise it’s mostly just curves for brightness and darker shadows etc.

if you have any questions, feel free to ask ^^ i love sharing my editing process but i rarely feel like it’s something i am good at explaining. mostly i just throw things together and see what works. am getting better at it tho and have learned a lot during all this time i’ve been around :’D

#tag game #thank you so much!! #i enjoy these tags every time #i just wish i could show more #am just awful when it comes to saving anything on my laptop #i get so anxious about using too much storage haha #also yes am a neat freak when it comes to my folders #but am happy i got to show at least something! #the b&w + one color has really been my thing lately #it's just so much fun to do imo #tho i hate it when i need to try and #keep the skin color for some pics #bc then i need to go select things by hand #and oh boy that is a struggle sometimes #do not look at my edits with a keen eye #or i will burn down

[knackfandomarchive]

Talking about composition in visual art. I hope you guys like when I ramble about random stuff. Something my brain kinda latched onto are these two images:

I'm not sure what to do for alt-text since I'll be explaining what they are anyway. They both have a similar arrangement, with Lucas, in the icy caves, walking beside a much-larger Knack who is covered in icicles. The poses of the characters aren't innately exciting, as they're just walking, and their expressions are in a neutral range. I would probably only need one of these if I were to use them as a pose reference. However, the lighting, shot distance, and silhouettes are very different, which is why I can't seem to bring myself to delete one. One thing I like to do when comparing images is make them grayscale.

Usually I just use the desaturate function in GIMP, but with Abominable Knack, I had to use the mono-mixer and tweak the settings, or else the values would change and he'd be darker than he is to my eyes. Losing my train of thought already. Uh, but you can sort of see why they're so different.

In the first, they're both in a dark part of the cave, and you see them in a more dramatic profile view. Knack is a bit like a shadow, but because he is so large and spiky in comparison to Lucas, he takes up more attention. Plus he is the one speaking at that time.

In the second, they've almost reached the mouth of the cave, the lighting is much brighter and Lucas takes up more visual space. Knack is sort of cut away, like the camera is no longer paying attention to him because Lucas is talking. Knack also looks at him while he is speaking, and because the camera is closer, the scene becomes more intimate.

So it's like, they go from an action-hero type composition to something else.

#knack 1 #talking about art #knack knack #lucas knack #video game screenshots #*gestures vaguely* #you know what I mean? #I should probably eat before I post. But I'll post before I eat. #knack #ice knack #discoknack (me)

[encyclopika]

Animal Crossing Fish - Explained #212

Brought to you by a marine biologist and the Kraken itself...

CLICK HERE FOR THE AC FISH EXPLAINED MASTERPOST!

I’ll admit I was saving this one. I liked the theme I had going starting and ending the original Fish Explained and the Museum Tour series with the deep sea. Except, now ACPC has bestowed upon us another denizen of the deep, so I want to cover this one RIGHT NOW SO BAD. It’s the Giant Squid!

The giant squid first appeared in ACPC for the Fishing Tourney #41 in August 2021. The theme was “the deep sea”, but besides the giant squid itself, nothing else about it really screamed “deep sea” to me, but I’m also obnoxious. Anyway -

The Giant Squid is giant and also a squid - it’s in Phylum Mollusca with our other mollusk friends in the games, like the vampire squid, the clam, etc. It’s an incredibly diverse group - just the fact that cephalopods - the class within Mollusca that includes the squid, nautilus, and octopus, the latter of which are lauded as the smartest invertebrates - are related to bivalves (not known at all for any thought) is astonishing to me. What complexity! 

Squid differ from octopuses in a number of ways, but most strikingly in the arrangement of their arms, of which they have 8, and two tentacles which are usually longer than the arms, and distinguished by only have suckers at the very end. And I mention that because the Giant Squid (Architeuthis dux) is one of the largest invertebrates by length because of its two long tentacles. Females of the species have been estimated to grow to a maximum total length of 43 ft (13m), with males falling a bit shorter. It’s perhaps the most famous example of deep-sea gigantism, along with the giant isopod. Despite how big this thing is, the thing it’s most notorious for, besides the size, is its elusiveness.

By http://www.pref.kyoto.jp/kaiyo/2-topicnews/news/2002/02-02-01/mega-squid/mega-squid-01.html (web archive: [2]), Fair use, https://en.wikipedia.org/w/index.php?curid=6280565. "Copyright (C) Kyoto Prefecture"

The picture above should give you an idea of how big this thing is (there’s a person above it). This is also the very first photo of a live specimen ever photographed, which was captured in 2002. Up until that point, giant squid were rarely seen washed up on shore or as bits and pieces within the guts of deceased Sperm Whales. There was a great effort in the 1990′s to find live specimens, especially those in their natural habitat, to learn anything about them.  

By Report in The Wall Street Journal, 28 December 2015 (full size image); image was auto-normalized in GIMP., Fair use, https://en.wikipedia.org/w/index.php?curid=62371432. Original copyright for this image belongs to Mizuhashi Fisherina, Toyama.

(^Here’s another great image to show you the massive size of just the mantle, on a juvenile, taken at the surface in 2015.) In 2004, the very first video of an adult giant squid in its natural habitat was filmed. I actually remember when that happened; it was a huge sensation in the marine bio world. Since then, a couple dozen giant squid have been found, and we have learned a lot about them - like that they have a worldwide distribution, they have a complex nervous system with a big brain, and it’s prey consists of deep sea fish and other squid species it ensnares with its long tentacles and uses its powerful beak to hold and crush them. They also have more predators than just the Sperm Whale, including pilot whales and sleeper sharks. Unfortunately, it’s really, really difficult to learn more about them without expensive equipment and dedication, since their habitat is at such dark, crushing depths. (If you wanna learn more about why we know more about the Moon than the deep sea, take the Museum Tour into The Abyss).Luckily, they sometimes wash ashore or come to the surface or end up in fishing gear, but, that can’t confirm or tell us as much as we’d like to know, like their behavior, if they have migration patterns, their reproduction habits and cycle, etc.

Although the Giant Squid may be the largest mollusk by length, it’s beat out in the weight department by the equally elusive and fascinating Colossal Squid (Mesonychoteuthis hamiltoni). These two squid are big, intimidating, and rarely seen - a perfect combination to spark lasting tales of the Kraken in our history and stories. 

And there you have it! Fascinating stuff, no? 

#giant squid #deep sea gigantism #deep sea #marine biology #sealife #squid #animals #animal crossing #animal crossing pocket camp #acpc #science in video games #animal crossing fish explained

[theseventhoffrostfall]

You do make a good point about the raiders! One of my earlier ideas, I said, had them working with the Railroad as part of "cooperation" (Railroad sets back the Institute's plans, raiders may or may not get some high-tech gear for their own plans). But looking on it now, you have a point: for 200 years these gimp-suited lunatics have been attacking little farms and for what? Some ears of corn, a box of tomatoes? 50 caps and a pile of dirty rags?

Taking inspiration from your idea of them being "bandit/survivalists" instead of dirty leather-wearing imbreds, there would be those who realize "Hey, raiding on farms isn't working anymore. Farmers aren't just carrying pipe pistols anymore, and whatever we do get isn't even enough to get us through the week. Scruffy Ed and his gang starved to death last week because they were too busy huffing Psycho to realize they were running out of food. None of us here can farm half a damn, but those Gunner freaks seem to get a lot of money just by providing protection and carrying out attacks for a bunch of rich guys-- or at least as far as I know. What if we did something like that, just to earn some extra scratch?"

This would lead to "mercenary groups" (or at least what raiders would think mercenary work is). These "groups", having years of combat training and know their way around a gun, advertise themselves as "affordable security." Sure, they may not be as loaded up with combat armor as the Gunners, but they're far cheaper. They'll do any job for the right price, protection, scare tactics, killing off your rival, deliveries, you name it. To the raiders, it's usually the same kind of "work" they used to do, but now they're getting good money for it. With this money, these former raider groups move out of their shacks, purchase new weapons, armor, food, water, medicine. Leadership is no longer based on how can screw the most people or huff the most chems, but based on loyalty, camaraderie, intelligence, skill, and knowledge. Butcher Bill could be as violent as a drugged-up pitbull, but his dumbass would get them all killed considering he doesn't even know which direction left is. Professionalism trumps wanton violence. They won't attack on sight, you could talk to them, maybe even gain favor with them, or just wipe them out if you want

Granted, there would still be raiders, those who didn't want to move on and still cling to the idea that their violent, savage life of drugs and wearing fish-hooks is going to work out for them one day. As such, they're way more insane than the other raiders but are thinning in number each day

Honestly, one of the few shreds of cool ideas in Fo4 was how a decent number of the raider gangs (unnamed and still homogenous, granted) raider gangs had named leaders with bits of politicking and bickering going on between them. While the "doing drugs and eating people" bit is still dumb, and everyone just using pipe guns strikes me more as lazy design than anything, I like the idea of these various petty warlords with gangs full of tough, gritty sonsofbitches that source any professionally-made gear they can and fill in the gaps with cunningly-handmade weapons, armor, and surprisingly advanced traps, turrets, etc. I actually like how, while the AI doesn't really bear it out, voice lines suggest that Raiders fight smart, albeit mostly being taught by experience and loose, rough mentorship (High level raiders will admonish lower-level ones to get behind cover, flank, ect) In theory, it's an evocative concept.

I kind of picture Raiders as being two-tiered. On the top, you have your lifers, your hard-bitten professional fighters like the Nuka-World gangs, and below them you have 'common' Raiders--your raider boss/warlord type is a pro whose gang is a rotating cast of part-timers who supplement their normal hardscrabble wasteland survivalism with a bit of "Fuck it, Farmer Joe over there's crops didn't fail this year like mine did, guess I'll see if Teeth-Taker Tom wants to bring a few boys over and get a cut of that action".

That's what was interesting about Sinjin, the raider boss of the Shroud quest. His whole threat (albeit one we're told and not shown) was that he was taking these two-bit raider gangs and training and organizing them into professional fighting forces.

So yeah, while Raiders also doing mercenary work (probably outside their own turf, within which they don't want to be seen as subordinate to anyone) would be cool to see also happen, I like the flavor of Raiders as raiding.

[tista-bie]

Seeing your scanlation efforts, even being 90% google-sensei and 10% intuition, is inspiring me to want to take the plunge and give a crack at translating the raws of my fave mangas which have (tragically) been dropped by their scanlation groups. If you don't mind me asking, what's your process for isolating and translating the text, what tools do you use, any tips and tricks, etc?

it's not efficient by any means, but i actually do know katakana and hiragana (and the manga i've been scanlating has furigana on all kanji) so i actually just set my keyboard to japanese ime and type what i read into google translate. if there's kanji that i can't pull up through this method (either by being a rare kanji, or one of those forced furigana readings) i use this site to look it up by its components (it also has one with hand-drawn inputs but you have to have at least a basic understanding of how things are written to use it), and in my brief explorations of manga that doesn't have furigana, while it is more arduous, you can repeat the same process on... all kanji

while i do say i use google translate (which i do), if i need a second opinion, i use deepl, which honestly usually provides a better translation, though sometimes, it instead simply provides a different translation. so, again, it's mostly useful as a second opinion if you're not clear on what something means. if neither result is as clear as you'd like, it's always worth googling to see if it's a japanese turn of phrase you've never heard of

ultimately, the most important aspect of the translating part, especially if you're like me and using machine translation, is not leaning too hard on a literal translation. if i just plug what google gives me into the word balloons, the result will be something like this

which, aside from sounding somewhat off (which is the main trip-up of machine translation), is also wrong; japanese doesn't explicitly state the subject/object of something most of the time. so google presumes the feelings he's conveying are hers instead of his, because it doesn't have context. that's where the 10% intuition comes in; it's your job to fill in the blanks and figure out those subjects/objects of the sentences that google can't. and it's also your job, as an english speaker, to tweak the final result into something real people say.

(one thing that helps with that is to have a "voice" you write the characters with; i usually write oscar very formally unless context clues suggest he's being informal, for example, because it's just how he is. finding that voice in each character helps you write their dialogue well in ways that reflect the original intent better than just a straight translation)

for cleaning/typesetting i use clip studio paint, mostly just because it's the tool i already use for my art. i've only done a tiny bit of redrawing because it's a pain in the ass, but if i were to do it more regularly then csp would definitely be the most useful tool for the job; it's my favorite art software i've ever used. but for the sake of just cleaning/typesetting, photoshop or gimp can fill much the same role. just make a new layer, cover as much text as you can without looking ugly with white, and then add the new text. if you want to get REALLY fancy, then look up tutorials from someone else, because i don't have the patience for proper cleaning/redrawing as a one-woman scan group

that's... pretty comprehensive for how i do it? when a new chapter releases, i grab the raws, read through them, translating page-by-page (i translate all lines on a page at once, sometimes even grabbing lines on the next page if i need more context to understand properly), typing it all up in a script as i go (i just split the script up by pages and who's speaking, relying on my memory of what goes in which balloon, but if you have a hard time tracking it it probably helps to put each balloon in a different line), and then, once i have the script ready, i clean and typeset each page at a time. and while i typeset, if there's anything i wasn't happy with in the script, or that doesn't fit where i need it to, this is where i put any final spin on it i need to. get through that, make a funny credits page, and bam. it's not what you might consider a proper scanlation, but if everyone's dropped the series, it's not hurting anyone to give it a shot

#sorry this became an essay i just got really into explaining this #i put a lot more thought into my process behind this than you would think given i'm a filthy machine translator

[solvicrafts]

Sølvi's Ongoing Resource List

NOTE: I am only adding resources that I have personally used and had good experiences with, so with that in mind your mileage may vary and you may find some resources better/more useful than others, or you may know of a different resource that you prefer. Also, this list is far from finished and I will be updating it frequently.

🖨️ Printing Services 🖨️

GotPrint (x)

So far, I have only used them for business cards, so I can't attest to their other products. I've used them for 4 different business card orders, and all of them turned out pretty good! There's almost always a good discount code somewhere out there for them, so they're a pretty good option price-wise.

2. Vistaprint (x)

I've used VP for a few different things, and in fact they used to be my go-to for business cards. I prefer to use GP now, but they're not a bad option. I've also used them to order custom mugs and have had good results.

3. Printful (x)

This company is often used for drop-shipping businesses, but I prefer to use them for personal designs as I'm not a huge fan of the drop-shipping business model, myself. Overall, their product quality varies quite a bit; I've had some stuff turn out amazing (the blankets for example have usually been pretty good), and some stuff turned out absolutely awful (Christmas stockings... OOF) Use at your own risk if you're trying to start a business around them; make sure you order sample products for yourself first before putting them up for sale so you can verify their quality.

4. WackyButtons (x) and WackyPrint (x)

Different websites, but they offer similar services and belong to the same company. WB is where I get my buttons (shocker, I know) and I've also gotten a few other things from WP like bottle openers, stickers, and magnets. Overall the quality for both is pretty good; that said if you're ordering bottle openers keep in mind that the color you pick does influence the quality of your print. While my stickers and magnets turned out great, I could barely see my design on the bottle openers. Make sure you use bright images with good color contrast.

5. Spoonflower (x)

I use SF for almost all of my custom fabric printing unless I only need a small amount (in which case, I buy printable fabric sheets from Jacquard). Primarily, I'm using them for my plushie pajama lines, and their print quality has always been pretty good for me. Their minky fabric is soft and easy to sew, but quite flat compared to other minky fabrics, so I wouldn't expect to use it for plush designs. Their fleece is very soft and pretty high quality, but tough to sew. Overall, they're a good option and the only one I've used for large projects, but their prices a bit steep. If you're not sure whether or not you want to take a risk on ordering large amounts of fabric from them, I'd recommend getting their sample kit first so you have an idea of what to expect.

🎨 Aristic Resources 🎨

Lunapic (x)

Lunapic is a web alternative to GIMP and Photoshop, and it's also free! I sometimes use it to make really quick edits/add simple filters like soft lighting to some of my pictures. I've used this website since, uh... my high school days. xD They've been around a long time.

2. Canva (x)

Canva is technically free to start designing stuff with, but a lot of things are locked behind the Pro membership. TBH I do think it's worth it if you plan on using the service a lot. You can also just do it for whichever months you want to, honestly. I use Canva for things like creating shop advertisements and banners, designing business cards, and just a whole bunch of other little things. Hell, I even use them for cleaning up my shop images and making avatars/icons and stuff :D Canva has their own assets that you can use, but you can also upload your own, too, which I strongly recommend. I would not use Canva for professional graphic design, but if you have assets of your own to contribute, it's helpful for lots of little things.

3. GIMP (x)

GIMP is a free alternative to Photoshop, and also something I've used for several years. TBH it is not a very easy program to learn if you're fairly new to these types of things, but that also kinda depends on the person, I think. I ended up downgrading to an earlier version because I wanted something a bit simpler and more user-friendly.

💻 Web Design (and Learning) Resources 💻

Udemy (x)

There are a looooooot of courses to find on Udemy. Quality can of course vary quite a bit, but odds are you'll find something good if you look around. That said, I highly recommend waiting for a sale because they tend to happen ridiculously often, so why spend $259.99 on a course when you can get it for $11.99? Anyway, I learned a lot of basics from Rob Percival's Digital Marketing course, and it covers quite a lot of different areas if you're looking to start a small business online and don't have a lot of experience. I started from there, took a few other courses in other things I had an interest in, looked around some more on YouTube to fill in some of the gaps, and just kinda went from there.

Udemy can be a great resource or kinda 'meh' depending on what you're trying to do, and it's very much a YMMV kinda thing. I've had pretty good experiences with their courses due to my own learning style, but not everyone will find them helpful. Definitely wait for the prices to drop if you wanna try them, though. Also, the one time I had to get a refund on a course, I found the process really simple and didn't have to fight for it, which is a HUGE plus for me.

2. Divi (x)

Gonna be honest, I'm not a fan of WordPress in general, but if you're not very experienced in webdev and just want to put together something fairly simple and nice looking, Divi's theme is highly customizable. I'd still recommend learning some HTML and CSS on your own to fill in the gaps, but if you want something simple and can afford it, Divi might be the way to go. I use it myself for all of my WP-based sites and really don't have any serious complaints about it. The nice thing too is that you can use it on unlimited websites, unlike most WP themes, and you can pick from starter templates and edit them as you see fit.

For example, I used Divi for both The Felted Dragon and the Bregan D'aerthe Online Fan-Site. (Granted, they are both severely in WIP status because I just do not have the time this month to work too much on them, but uh... if you ignore that, they do look okay xD)

3. Namecheap & EasyWP (x)

I have bundled these into one because there actually IS a way to bundle them! I use Namecheap for my domain name registrations, and there is almost always a discount code of some sort floating about, so they're a fairly inexpensive option. Maybe there's a better registrar out there, maybe not. Either way, I like them for their simplicity. If you're really really new to website building and want as simple of a process as you can realistically get, you can also get EasyWP hosting after purchasing a domain. I do think there's better hosting out there tbh, but for a fairly simple/user-friendly experience, it can be a good option, and you do get your first month free. It's also really easy to cancel an EasyWP subscription from your Apps on the Namecheap dashboard, so you could always try it out and then cancel if you'd rather go for something else.

4. DepositPhotos (x)

If you're looking for stock images and trying to decide between DepositPhotos and ShutterStock, I recommend DP. For one thing, their prices are generally better, and for another... I've actually seen a LOT of the same images on both sites! I'm also a fan of being able to buy extra images on top of the subscription for a lower cost, which I've definitely used a few times for web design. That said, definitely shop around because there are a lot more options for stock images now, and there might be better deals out there now.

------------------------------------------------------------------------------

That's about it, for now. I'll be updating this list as I think of more stuff to add to it; I know I'm definitely forgetting a LOT of things lol

#save tag #solvi's resources

[goodgifswithfreetoolschallenge]

How to Reduce The File Size of a Gif

NO MATTER WHAT METHOD YOU USE TO MAKE GIFS

You can make the most amazing looking gif in the whole wide world, but if it has a gigantic file size (greater than 10-10.5MB), it is never going to upload!

To understand how to reduce the size of a gif, it helps to understand how exactly a gif works.

NOTE: If you have an existing gif you are looking to reduce in size, check out Gifski or ezgif.com's Optimize pane.

There are a few different ways to reduce the file size of a gif.

A) Keep it short.

If a gif is nothing more than a bunch of frames from a video packed together in an image stack, keeping a gif short means fewer images in that stack!

Exactly how short your gif needs to be in order to be small enough to upload depends on many factors (FPS, scale, color, aspect ratio) but generally, most gifs made by conventional means are going to be too big to upload if they are longer than 5 seconds in my experience—and even that may be too much depending on how you make your gifs and their scale. 3 seconds is generally a good length to shoot for.

B) Scale it down.

Besides shortening the length of your gif, scaling your gif down is the most effective way to reduce its size. There is (usually) no point in uploading a gif to Tumblr that is 1080 pixels wide. Tumblr will compress your image down to 540 pixels to fit it into the post window. If you scale down that gif it would fall in file size significantly!

For example, the following gif is scaled down to 540 pixels, and is 9.3MB in size.

When this same gif is full size (1080x1080) it's 47.9MB!!!! That's 5.15 times larger!

C) Decrease frames per second (FPS)/the number of frames captured per second.

24 FPS is the standard frame rate of most live-action movies and TV shows. There is little point (usually) in setting your frame rate above 24 FPS when giffing shows and movies, as this will only result in having a lot of duplicate frames taking up space in your gif. However, you can also reduce FPS below 24 FPS and generally you'll still be happy with your gif as long as you don't reduce the FPS too much and there isn't a lot of movement in your gif. I wouldn't personally decrease FPS below 18 FPS, as I think after this point, gifs look too choppy for my taste, but you may feel differently and think a lower FPS than 18 fine, or think 18 FPS is too low—and that's okay!

NOTE: When converting straight from video to gif, decreasing FPS and the number of frames goes hand and hand. If you are screencapping frames and then converting that stack of frames to gif in a program, you should capture your frames at a different rate per second or delete frames and change your frame delay to reflect the new FPS you need to make your gif run at a "normal" speed.

D) Mind the aspect ratio.

Think about surface area here (in pixels squared). Consider three gifs that are all 540 pixels wide: A rectangular gif taller than it is wide (ex: 540 x 720) will have frames with the greatest surface area and therefore the highest file size. A square gif (540 x 540) will have the next largest file size, and a rectangular gif wider than it is tall (ex: 540 x 320) will have the smallest file size because it's going to end up having the smallest pixel by pixel surface area.

This does not mean that all gifs should be rectangular and wider than they are tall. In fact you often times absolutely will NOT want that. But it is worth remembering when you are gauging how long the gifs you are making can be while still being small enough to upload. For example, when I make 3:2 aspect ratio sets, I know my gifs can be a little bit longer than they can when I make 1:1 (square) sets.

E) Reduce the number of colors.

A gif can only have 256 colors, but if you reduce that number down further, you can also reduce your gif's file size. Applications like FFmpeg, DaVinci Resolve, GIMP, ezgif.com, and ScreenToGif can reduce the number of colors in your gif to reduce the file size.

F) Cut duplicate frames and change the delay on remaining frames.

Suppose your gif contains a long pause where a character doesn't move for a good half a second. If you captured that gif at 24 FPS, then there are 12 frames that are all exactly the same in your gif! Suppose you could delete all the duplicate frames, and just make the one remaining frame stay put for 0.5 seconds???

Most gif-making tools can't do this for you. However, you can do this in applications like GIMP, ScreenToGif, PhotoScape X, and on ezgif.com. I think Gifsicle also does this.

[tastydregs]

While Google and OpenAI Fight for the AI Bone, the Open Source Community Is Running Away with It

“Hey, did you hear? they say Google and OpenAI don’t have a competitive advantage in LLM.” “Yeah, sure… who said it?” “Google.” “Wait a minute…”

Photo by Kai Wenzel on Unsplash

One week ago SemiAnalysis released a real shocker when they made public a leaked document from Google titled “We Have No Competitive Edge, And Neither Does OpenAI.” While we can’t be sure if the document is legit, it brings up some thought-provoking points about the real struggle in the world of large language models (LLMs). It’s not Google vs. OpenAI; it’s more like open-source LLMs taking on their closed-source counterparts.

This leaked document hints that both Google and OpenAI might be losing their edge against the ever-growing open-source LLM community. The reason? It’s pretty simple: open-source projects are moving at lightning speed. Faster than large corporations or corporate-backed companies can match, especially since open-source projects don’t face many reputational risks. Apparently written by a Google researcher, the document emphasizes that even though Google and OpenAI have been working hard to create the most powerful language models, the open-source community is catching up at an astonishing speed. Open-source models are quicker, more adaptable, and more portable. They’ve managed to achieve great results with much less resources, while Google is grappling with bigger budgets and more complex models.

What’s more, having tons of researchers working together in the open makes it tougher for companies like Google and OpenAI to stay ahead of the game in terms of technology. The report says that keeping a competitive edge in tech is getting even more difficult now that cutting-edge LLM research is within reach. Research institutions around the globe are building on each other’s work, exploring the solution space in a way that’s way beyond what any single company can do. Turns out, being able to train huge models from scratch on pricey hardware isn’t the game changer it used to be, which means pretty much anyone with a cool idea can create an LLM and share it.

Alright, we’ve seen open-source projects trying to outdo their corporate counterparts before, but let’s dig a bit deeper to see if this is a genuine threat in the AI world.

Ups and downs in the world of open collaboration

Open-source software has always seen its ups and downs. Some projects like BIND, WordPress, and Firefox have done really well, showing that they can stand up against big-name enterprise products. On the flip side, projects like OpenOffice, GIMP and OpenSolaris faced struggles and rapidly lost ground. Regardless, open-source software is still popular, with many websites using Apache web servers, BIND servers, and MySQL databases.

Now, the problem is that keeping open-source projects funded and maintained can be tricky. It takes solid planning, the right resources, and a real connection with users. If a project has a dedicated user base and passionate developers, it’s more likely to stay on top of its game and keep getting better. Back in 2018, OpenAI faced some of these hurdles and decided it was time for a change. They started looking for capital and eventually became a capped-profit company. That means they could get investments and offer investors a return capped at 100x their initial investment.

OpenAI said they needed to make these changes to fund research, support big companies, and keep things safe. So, you could argue they did what they had to do to dodge the usual open-source pitfalls. But that did not come for free, since, while OpenAI has made impressive progress in AI development, its increasing secrecy, lack of transparency, and limited customization options have alienated the very community it once aimed to serve.

On the other hand, Google is really into open-source software, and they are involved in quite a few open-source projects. Just look at Android, their mobile operating system. It’s built on the Linux kernel and has been a game-changer in making open-source software popular in the smartphone world. Today, most smartphones run on Android. Another awesome open-source project from Google is Kubernetes, which has become the top choice for container orchestration. It helps developers automate things like deployment, scaling, and managing containerized applications. Last but not least, let’s not forget Chromium. Google’s Chrome is built on the open-source Chromium project, and it has become super popular since its launch.

By being part of open-source projects like these, Google shows they are really into transparency, openness, and working together to create innovative and flexible software solutions. They are dedicated to making the tech world more inclusive, diverse, and accessible for everyone. For this reason, I wouldn’t be too shocked if Google decided to make their next big language model an open-source project. It could be a clever move because they’d have all their brand, marketing, and developer muscle behind it, giving OpenAI some serious competition. Of course, that’s assuming the model’s quality would be on par, which hasn’t been the case so far. What’s even more crucial is that someone else might snag that spot first. As we’ll see next, there’s a long list of newbies waiting in line.

A guided journey through the open-source LLM boom

One of the coolest aspects of the SemiAnalysis document is the recent timeline highlighting key milestones in the open-source community, particularly in the area of large language models (LLMs). It all starts with what might be considered the “big bang” of recent open-source LLM advancements — the release of LLaMA by Meta on February 24, 2023 . LLaMA is a LLM with sizes from 7B to 65B parameters, claiming to require less computing power, making it ideal for testing new approaches. Actually, it was not released as an open-source model, but one week after the release, LLaMA’s model weights leaked to the public, and everyone got a chance to play around with it. That’s when things started snowballing.

Here is a quick summary of the milestones described in the document:

Artem Andreenko’s Raspberry Pi implementation of LLaMA (March 12, 2023)

Stanford’s Alpaca release (March 13, 2023)

Georgi Gerganov’s 4-bit quantization of LLaMA, letting it run on a MacBook CPU without a GPU (March 18, 2023)

Vicuna’s 13B model release (March 19, 2023) was trained for just $300.

Cerebras trained an open-source GPT-3 architecture that outshines existing GPT-3 clones.

LLaMA-Adapter (March 28, 2023) set a new record on multimodal ScienceQA with just 1.2M learnable parameters using a Parameter Efficient Fine Tuning (PEFT) technique.

UC Berkeley’s Koala (April 3, 2023) releases a dialogue model trained entirely on free data, costing $100 and scoring over 50% user preference compared to ChatGPT.

Open Assistant release (April 15, 2023) offers a complete open stack for running RLHF (reinforcement learning from human feedback) models with a 48.3% human preference rating.

And the list goes on…

Yes, you got it right. All of this happened in just over two months, proving how lively the open-source scene is. However, most of these open-source models might be known only to insiders and haven’t hit the mainstream (I have to admit that I learned about the majority of them through the leaked document too). But just some days ago, on May 5, a possible game-changer arrived: MosaicML released MPT-7B, setting the bar high for open-source competitors (and maybe even OpenAI). What’s more, it’s licensed for commercial use, unlike LLaMA. The MPT series also supports super-long inputs with context lengths of up to 84k during inference.

MosaicML put the MPT series through rigorous tests on various benchmarks, showing it can match LLaMA-7B’s high-quality standards. The base MPT-7B model is a decoder-style transformer with 6.7B parameters, trained on 1T tokens of text and code. MosaicML also released three fine-tuned versions: MPT-7B-StoryWriter-65k+ for super long context lengths in fiction; MPT-7B-Instruct for short-form instruction following; and MPT-7B-Chat, a chatbot-like model for dialogue generation.

MPT-7B was trained on the MosaicML platform in just 9.5 days using 440 GPUs, with no human intervention. The perks of these improvements are clear in the numbers. For example, MosaicML claims that the MPT-7B model gives a competitive performance with only 7 billion parameters, which is 28 times smaller than OpenAI’s GPT-3 with its 175 billion parameters. This size reduction means big cost savings since fewer resources are needed for both training and deployment. Plus, the smaller MPT-7B model is more portable, making it easier to incorporate into various applications and platforms.

So, the million-dollar question is: can MPT-7B match the quality level we’ve come to expect from ChatGPT? We’ll see. But one thing is for sure — the open-source world of LLMs is buzzing with excitement and innovation, and it won’t be long before we’ll find out.

So, do we have a clear winner?

No, actually we don’t. I don’t think that all the fears expressed in the article are necessarily well founded. Many of the concerns raised in the article might not be well-founded. However, this isn’t exactly great news for Google either, as OpenAI still remains by far the undisputed leader in the LLM market. OpenAI made a clever and bold move in November 2022 by launching ChatGPT for public use, totally free and perhaps not yet fully secured. This move garnered massive traction, making OpenAI the fastest company to reach one million users in just five days, and a whopping 100 million by the end of January 2023.

Impressive numbers aside, there’s an important point to be made here: OpenAI is collecting a huge (and I mean HUGE) amount of user data. While the leaked document claims that faster, cheaper algorithms provide a competitive advantage, that’s only part of the story. In the realm of AI, what truly matters to users is the quality of information that models offer. To make better inferences, more data and feedback are needed, and that’s exactly what OpenAI is collecting.

Additionally, it’s worth noting that OpenAI has Microsoft’s backing, granting them access to a massive cloud of user data. When your AI can whip up a stunning PowerPoint presentation, a comprehensive Excel spreadsheet, or the perfect LinkedIn profile just by describing them in words, the algorithm used to accomplish that becomes irrelevant. But when you ask questions and receive incorrect responses, knowing the algorithm was trained on a mere $300 budget is hardly comforting.

While Google and OpenAI Fight for the AI Bone, the Open Source Community Is Running Away with It was originally published in Towards AI on Medium, where people are continuing the conversation by highlighting and responding to this story.

[nenekiribookwyrm]

April 2023 Update!

Happy end of the month! I've got quite a bit of news to share so let's get into it.

I finished my project for the Werewolf Jam and I'm pretty pleased with how May The Hunt Be Everlasting and May Nature Guide You Home turned out. I think I did something neat within the time frame of the jam and I got a chance to jump back into Twine development after many years away. I also used public domain art and some creative edits in GIMP to make the cover art for the game. I didn't have quite enough time to draw or paint an image like I wanted to, so I felt like this was a good compromise. If you're interested in playing the game yourself, it's available for free and runs in your browser over at: https://nenekiri.itch.io/may-the-hunt-be-everlasting-and-may-nature-guide-you-home

The voting period for the Werewolf jam runs until May 6th, so if you have the time I'd appreciate a review and what you thought about the game!

I was lucky enough to have a poem of mine featured on the Voice of Dog this month for National Poetry Month and it turned out really good! This piece of writing is a really personal one to me and I'm glad that Rob let me read it for the podcast this month. If you'd like to listen to it you can do so here: https://thevoice.dog/episode/forest-by-nenekiri-bookwyrm

There's going to be a few more of my poems planned to be read out on the podcast at a later date, so I made a poem section for the website. If you'd like to get a sneak peek of what could be featured you can read my poetry in my collection I have on my SoFurry page: https://www.sofurry.com/browse/folder/stories?by=520851&folder=66067

I've sent in my author bio for the Post-Self short story anthology titled Clade and gotten word that Madison is targeting a summer time release. So get excited for that later in the year! I wrote my bio in-universe and I really like how it turned out. Since the story focuses on more human characters rather than the usual furry animals, I opted to use my human name. Don't cast any binding spells on me with that deep knowledge! It's also a bit of an experiment to see how using that name works out. If I end up writing more fiction that deals with humans primarily over animals, I may end up using that name for those kinds of stories in the future. Would be an easy way to let potential readers who are familiar with my work engage with the kinds of stories they're looking for.

I've started working on a new big project! Until I hear back about the short story anthology with dragons, I'm kind of at an impasse with working on that for the time being. So I'm switching gears for a while and I'm going to try and write a story featuring butterflies. I was very much into butterflies as a kid, but haven't found a good angle to write a story about them now that I've fallen into furry writing. This project is an attempt to do that. The basic premise is that I'll be following a group of butterflies as they perform their annual migration from one side of the world to the other. This migration can take a very long time, depending on the butterflies you're looking at, it can anywhere from 3-6 generations to make that trip. I think this is the angle I've been looking for. What would it be like to be a butterfly on this journey knowing that you will never see the destination that you're traveling towards? How do they get through their day to day? What hardships do they face in their travels? All of these questions excite me terribly and I'm itching to get started writing.

But before I fully commit, I want to do some preliminary research. This is going to be a lot bigger of a project than I'm used to. Depending on how I end up tackling it, it could very well end up being novel length. And if not that, at least the length of a novella. Which means I should get a decent baseline of information on butterflies before I get too far into the weeds in writing a bunch of characters for the story.

Thankfully, my local library had a ton of books on the subject and I'm working on reading through a bunch and taking copious notes. One of the books, The Language of Butterflies has had a very interesting first couple chapters. There seems to be a concerning amount of butterfly researchers who had some sort of experience when they were growing up that led to them spending their lives studying them. In an obsessive fashion. The author makes references to this phenomenon as a kind of “Butterfly Madness” (I am paraphrasing, she doesn’t actually call it that) and I’ll be honest that it gave me pause as I was reading through the first few chapters, lest I fall into this pattern myself. I was enamored with them a lot when I was a kid, so maybe I’m immune? Or was it a long dormant madness waiting to re-emerge at the correctly precise trigger? This is the closest I have come to actually experiencing the unease that is supposed to be common in cosmic horror. Perhaps, I too, will become an acolyte of Prodryas before long.

Speaking of the library, they had a big book sale this weekend and I picked up a good few books that I'm excited to dig into at some point. The Dragon's Luck book is probably the one I'll dig into first, but the animal anatomy and painting with acrylic book are going to be invaluable for me to practice painting more.

Work has unfortunately been very busy as they shift my job duties around. I don't see it slowing down any time soon unfortunately, so I'm hoping to stave off burnout for the time being with being able to work on my personal projects in my free time.

On a brighter note to end out this update post, the weather has been good enough to take walks again and I've really missed being able to do that.

That's about everything I have for this month. Hoping that the next month treats you well!

#updates #werewolf jam #werewolf #butterfly #voice of dog