following junichiro horikawa youtube tutorials. dont think theres a better way to learn procedural than this... he has a whole series on VEX in houdini.... blender will take another ~3yrs (made up number) to catch up to houdini for procedural... houdini only good procedural graphics software for 3d atm afaik.... horikawa has a playlist of 26 vids with length of about ~50 hrs on VEX in houdini

im getting closer to halfway through this 50 hours... this n word not english native speaker so can speed video speed up to 1.25x to make it what... 40 hrs... and then my advice for how i learn from youtube is to have either 2 screen, or 1 screen. and have youtube vid on left half and copy every single action into your own houdini program as you go. often pausing the video and pressing left-arrow-key to rewind time by 5 seconds. this means the 40hr of content will take maybe a total of 50-80 hr depending how much u want. but i cant recommend junichiro content enough... i am only amateur at procedural but he really showing me a lot of ways to make fundamental building blocks of how VEX work in houdini... like i said.. blender not there yet for procedural for another few yr... then blender probably overtake houdini but u gotta know houdini for procedural for next at least 2 yrs in my noob opinion and no one gonna show it better than junichiro. he then has around ~250 hr of livestreams on youtube, divided into 2hr streams where in each stream he'll recreate some 3d graphics implementation of algorithm.

i think along with choosing some of these might need to learn some python->houdini, python also has implementation in blender but python->houdini + VEX looks pretty good to me... future of 3d graphics is prob largely procedural and no one do it better than houdini and junichiro san so jump on this hype train today... i am noob this is financial advice. most u guys probably into 2d. i dont know if theres much interesting work to be done in 2d with procedural... prob not... this more for 3d hype train

Rain25209 days ago

for blender's python id recommend tutorials by youtuber 'CG Python'. goes over the basics of python into blender... really informative for noob like me... often provides a link of code for 'initial code' and 'final code' for a lot of his videos so you don't have to type out 20 lines of code from paused youtube vid at start... also has a couple really good utility scripts hes brought up in a couple of his vids and often included in 'initial code' link for totally clearing blender's scene etc. so you can start each script under the same conditions each time while prototyping scripts....

blender 3.x to blender 4.0 the python hasnt changed at all pretty much afaik... so you can just follow his tutorials (even tho theyre written in 3.x) you can just follow them while using blender 4.0 and unlikely to have to google why something isnt working

theres maybe a couple weird things for getting into python scripting in blender... theres something bout a 'context' which is kind of confusing... blender has 'active object' and multiple selected objects sometimes. i found following CG Python tutorials informative.. goes over a bunch of useful concepts for python scripting blender.. and a bunch of tips and tricks. python can create Geometry Node trees and create nodes in these trees, alter parameter values, create links between nodes... can read in data to blender with e.g. json files... CG Python has vids for these topics as well as more basics of adding and moving cube or assorted other things u can do with blenders python. something like 30 videos in his playlist for blender python... few more of his vids that arent in that playlist which are still blender python tutorial vids

Rain25210 days ago

procedural modeling can be high poly or low poly... scatter meshes on surfaces... particles... low poly modeling... 3d curve modeling.... loads of techniques all adjacent to procedural... relatively unexplored type of modeling... think it would be useful to know how to do a few of these and mix them together with some procedural things u could add to workflow... probably want to make modules could reuse on different ideas

Rain25211 days ago

godot 4 looks cool... i mucked around in godot 3 and that was cool.... has everything u need... and theres great tutorials for it for doing the most-played sort of gametypes u might wanna make... most the ppl who use godot looks like they understand programming rlly well so u can prob get help on forums or browse forum posts etc or youtube vids... doesnt have the visual nodebased creation like in ue5 or unity so might be easier to dip toes in with those node based game creation things first... altho unity did something weird with pricing recently or something... and ue5 overhauled the character animation so now its ridiculously complex maybe... i wonder if godot will get that ue5 renderer thing where it LODs everything on the fly. would that be yrs away or 1 yr away 2yr away or what... photo realism not always dank tho lots of games wouldnt make sense with photo realism guess

think i saw 'stride warping' or one of those other newer character animation techniques in games lately looks kind of funny when ive seen it implemented... looks like the legs are a separate animal to the top half of the body with the legs running sideways and the top half of the body doing something else.......... godot 4 looks pretty sweet and ue5 too...

Rain25211 days ago

even noob ier opinions

theres a guy who does youtube videos on procedural textures in blender... ryan king art... so you can follow along and copy everything he does... i dont find it that useful for this specific thing to copy all the steps he makes.... id rather see him do a video where he outlines the most-common setups of e.g. color ramp into bump node into metallic shader etc.... i dont think thats a real setup.. but i dont think he has done a video going over the basics.... yea he makes nice procedural textures though... procedural textures pretty dank. i think if i wanted to learn procedural textures id more just plug shit into other shit and fuck around with it for hours and hours.... can get some ideas from ryan king art's videos i guess... he kind of moves pretty quick and its hard to follow i find... ye id prob rather learn texture synthesis from messing around with the nodes making more abstract things probably. a lot of procedural textures use a fair amount of computer resources maybe idk... might not be low poly gucci gang master race but ue5 has the crazy new renderer for auto LODing everything now... think 6 months ago it didnt work for foliage though.... if you want to make realistic epic 4000k gta 6 super realistic reflections x69 ray trace graphics with photo scann asset more pixels than real life. my favorite is the videos on youtube 'epic gta 5 ultra realistic graphics' and u click the video and its just like the entire bottom half of the screen is a mirror of the top half of the screen... reflections of the puddles on the road surface ...

Rain25211 days ago

even noob ier opinions

i think theres enough houdini content on youtube if you look around a bit... depends what you want to learn. 3d graphics is an enormous subject so its good to have some specific goal and find tutorials that have some element you want to learn. like junichiro does 130 different videos each on some specific algorithm and how he implements it using houdini and VEX... maybe 90% of those videos wont be something you want to follow... or maybe you would want to follow a lot more... depends what sort of rough goals u have for 3d.... dont go into procedural modeling with absolutely no goals lol caus its not easy to learn... u wn have some kind of rough goals in mind probably i dont think ud last long otherwise learning procedural. lots of ppl have made tutorials on terrain generation... basic plant generators... theres probably plenty of work to be done just on those 2 subjects where u could create your own procedural setups and workflows and go deep into one of those... motion graphics... idk much about animation/rigging so im not sure how that would work in houdini but it might be alright. im noob take advice with salt. i guess u can also learn from forum posts or books or reading manual... i think probably easier to try start from youtube caus u can see the steps being done... depends what u want to learn again and if u can find good book or not idk i cant read so i prefer youtube most of the tmie. there might be paid video courses that are alright. probably get advice from someone who learned more than 15 hrs from youtube tutorial.... but i do think procedural modeling in general gotta be pretty useful for certain applications... most things can be proceduralized and probably in the future more and more things will be proceduralized... its basically just making your own tools for making some type of art.... its basically the least destructive form of modeling so allows a lot more creative decision to be made without having to redo hours of work... is my opinion. .... i saw some video on grasshopper software and they were calling it 'parametric modeling'... just looked like another form of proceduralism... maybe grasshopper is good software... might have been a more limited form of proceduralism or proceduralism with less customization idk just looked exactly the same as procedural modeling. i think horikawa has a few vids on grasshopper or grasshopper to houdini

Rain25211 days ago

i have folder in my houdini projects folder called 'follow alongs' - every houdini file saved in there is listed as "YOUTUBER NAME - 'VIDEO NAME' v1" etc. this way you can save whatever progress you make when following along with whatever youtube video. also caus im pretty adhd i usually only do about 20 minutes of a video or 5 minutes of a video then get bored do another video instead... but this way you can sort of still save progress of what youve done so far if you want to try following a lot of youtube tutorials and bounce around between them if u have the adhd or whatever u wnt....

im thinking for videos where you REALLY want to learn what theyre doing in the video... might be good to follow a video 2 or 3 times... by the 3rd time you can probably quickly remember all the exact steps they did just before theyre about to do them so then i guess u might really understand the content.... idk if ill do these horikawa VEX videos multiple times... ill probably move on to more practical applications instead

Rain25212 days ago

also make a text file where you list all the shortcuts of the program that you use, with brief description. then organise them in bullet points for e.g. node graph shortcuts, viewport shortcuts, etc. then whenever you forget shortcut can use ctrl-f to search for keyword in the description u wrote of the shortcut if the file gets huge. e.g. if you want attribute wrangle node in houdini, you can click on blank spot in nodegraph viewer, then type tab to bring up addnode menu, then type aw{enter} to search for attribute wrangle and hit enter to add it, then can press {f2} to rename node... you want to be using the shortcuts most of the time probably... then just remember u need to type 'aw' to get attribute wrangle node... each node will have some specific shortest thing you can type to bring it up

Rain25212 days ago

i sort of follow 80% what junichiro says and most of the time the 20% i dont understand he says itll make more sense in a later video in the series and 15 hrs in it does seem to be the case..... again its dry as hell building blocks of procedural modeling its boring af but he does it quick and cuts all the bullshit and refers to the structures he uses in his practical applications of the processes as far as i can tell etc i rate a good 10/10

Rain25213 days ago

i dont know if vectors can be 5 numbers yet

Rain25213 days ago

sometiems vector 2 numbers or 4 numbers too

Rain25213 days ago

if uve ever watched a programming tutorial its usually like "this is how to type hello world, and this is the FIFTY FUCKING DIFFERENT DATA STRUCTURES U WONT USE FOR A BILLION HOURS BUT ILL DESCRIBE THEM ANYWAY'... the junichiro cut out most of the pointless shit like that so far from what ive seen... but it is still fucking dry as hell caus theres so many building blocks to make anything work at all in procedural 3d. maybe u do want some basic knowledge in programming or 3d but i rlly dont think entirely too necessary as long as i guesss u google instead whenever u dont understand or not..... all u have to know is a vector is 3 numbers, each one can be positive or negative... i think anyway

Rain25213 days ago

its fucking DRY content just watching 50 hrs of how to learn fundamental building blocks of houdini but junichiro not slipping on the dry shit or whatever the fuck

Rain25213 days ago

actually ive only done first 15 or so hrs of his VEX tutorial but he has been droppin bangers all 15 hrs so i think its enough cred to drop endorsement personally

Rain25213 days ago

i know im talkin up the junichiro san a lot... but i rlly think his tutorials are approachable to both programmer and non-programmer. programmer will breeze thru a lot of the concepts he is bringing up. non programmer will find it a little bit overload... but junichiro san really explain everything in detail. most programmer have no idea how to explain wtf they doing but he somehow can explain this strong endorsement

More submissions by Rain25 for Daily Art Club

drumming to haterade

i am calling it as corporate dystopia for now for muh country shitpost in comment

remade input systems shitpost

ppl on this streak be giving texture to things with shading

yung lean yoshi city

enjoying the possibilities symmetric drawing opens up


merry xmas and new years

adobe after effects has paint bucket tool WORKS on video footage. need to keyframe things. this is no keyframing. this is also 8 fps and quarter quality. looks alright in real render. might be helpful to use camera in 4k filming instead of 1080p for crisper lines. mirror phone app im using doesnt use 4k it only goes up to 1080p but i dont think its too difficult to make a phone app that films 4k mirror image and displays on screen. filming pretty far back atm phone is bout 1cm from eyeball


this was before i got the script working. the script is working now it selects based on colors ive inputted into arrays that are too similar. i have colorblindness. the boring colorblindness 5% males have. everything redgreenbrownorange look like shit brown to me i think is formal definition. it makes making color palettes easier for me. i can draw with any color and it look like any other color. this was colors generated before i fixed terrible coding i wrote on original script. script should start spitting out good color palettes in couple days under conditions im encoding for which is 'unique from every other color to me'

wrote script for cutting out colors from randomizer. selecting 13 colors per palette atm, then drawing all 13 colors on page, then typing each clashing color into each corresponding array that was talking about yesterday. seems to work alright atm. selecting 13 colors from initial 70 colors it gets down to ~45 colors atm left in the array at the point after last color is chosen... probably still a lot more clashing colors. would probably be faster if i inputted non-clashing colors and had randomizer temporarily NOT select any confirmed non-too-similar-color-pairing. half of this tedious grind is memorizing where each marker goes in the box irl. also small amount of fun looking at the colors next to each other... i been drawing with just pen for looooong time so it amuse to see color

grinding 4-6 color palette creation

from 100 alcohol markers. 30 was too dark didnt pop off the black lines. 70 colors. for greens might have 10 greens some are more green-blue some are more green-brown some are lighter green some are darkergreen. thing with the palettes i want is i want 4 to 6 colors per palette where each color is unique enough from all other colors in palette. this is easy to do in programming if each color has an 'array' (list) a list of colors its TOO similar to... only tricky part is 70 * 69 colors is 4,830 combinations. so need to draw a lot of colors on the page to create array of similar colors for each color, yellow prob gonna look too similar to light yellow, recently i found grey look too similar to magenta

once i have it coded in arrays for each color i can create randomizer and it might spit out half-decent color palette every 2/3 runs of script.

i think this is halfdecent way of creating color palettes for specifically... 4-6 colors where u have fixed-e.g. 70 colors. might be slightly relevant concept for ppl who work in low poly or pixel art. pick 70 colors and painstaking create array for each color where if its too similar to another color it doesnt get selected in color-palette-randomizer generator. idk why the im always talking about randomizer JUST IF I LOOK AT A BOX FUCKING OF 70 MARKERS I CANT PICK 5 RRRRAAAAAANNNNNDOOOOOM markers every time but that just randomizers are pretty lit wig why so aggressive idk why is this page turn into useless tutorial page rip i use to talk about fun stuff lizard ppl

love the sounds of painting in this. why is wiggly paint more fun than regular paint? maybe less apprehension since more uncertainty. idk dont philosophize about fun app why u gotta ask question. nice amount of random vs choice! sounds for each brush

attempt selling out to addiction machine i become zombie

December 12, 2023.mp444mb

Daily Art Club

Make art every day, how long can you last?

joined 835 days ago