Controlled Glitch Art

HΣγ ιΛ!η!! I saw an old thread on glitch art but it was lost in time. This is my favorite databend I have done. How does LAIN obtain their controlled glitches? I use audacity and word pad. How does one glitch two photos together for something unique? How do we feel about pixel sorting?

>How does LAIN obtain their controlled glitches? I just write a bunch of python scripts. Here's a good library to get you started in that direction: https://imageio.github.io/ Most of my effects are just normal image manipulation algorithms without any clamping or restrictions on the control parameters. For example, pic related is just a box blur where I don't check for overflow or renormalize after the convolution. Rather than saturating to white, the pixel values "wrap around" modulo the bit depth. >How does one glitch two photos together for something unique? Look into how layers are handled in image editing software. You have a ton of options here. https://en.wikipedia.org/wiki/Blend_modes Boolean blend modes are particularly interesting for glitch.

some more of mine

Lainzine Vol 3 has the guide

>audacity It is nice and i m using it too but it is not that good I have started to blender and looking to create 3d glitches because they are cooler

>>1999 >3d glitching Mind posting some?

>>2015 Unfortunately no because i am still at the learning part

>>1999 Does blender let you write your own glsl shaders? I imagine that could lead to some very nice results.

>>2018 It seems like you can write your own glsl shaders, blender is really huge program i think if you dive deep enough there will be wonderfull glitchy projects i cant understand why so many peoples hate it

pixel sorting is actually my favorite effect. I'm not actually sure how it's related to glitch art aside from aesthetic though, as I simply do it in either photoshop using the wind effect on my schools lab computers and with gimp using a plugin at home. Is there a way to get the effect with typical databending methods? Although actually after learning gimp and photoshop pretty thoroughly i'd rather create glitch like effects in them then open images in audacity or a hex editor and fuarrrk around corrupt it 5 times to get something that may not even look good. I def appreciate it, but like I know I could create OP's picture by playing with color channels and doing some stretching and color tweaking. And I would be sure I would be getting the outcome I want instead of hoping and praying.

>>2020 gimp is actually quite good for glitch art because the built in effects are kinda buggy, and it has some kind of scripting support (which I haven't actually used so ymmv). check out the wonderful "general convolution" filter or whatever it's called. lets you do arbitrary 5x5 filters. i tend to use gimp to "clean up" images after i've glitched them, or to quickly render masks and such to be used as glitch sources. >i'd rather create glitch like effects in them then open images in audacity or a hex editor the audacity/hex editor method is really more of a proof of concept than a practical tool. as i said here >>1888 the best way to do controlled databending is to use scripts. if you don't know how to program, glitch art is a decent way to learn. a lot of the algorithms you'd use are Computer Science 101. the general idea is to design a function that iterates over the pixels in your image and does *something* nonlinear to them. it helps to know a bit about how normal image processing works, so you know which parts of the standard algorithms to break in order to create the effect you're looking for. well, that's the surface of it anyway. you can of course go deeper with convolutions, merging multiple images, animation/time-dependency, intentional bugs with casting and overflow, etc. >pixel sorting i'm fairly certain that's not what the wind effect does, but i could be wrong. here's how pixel sorting works: >pick a threshold value, and optionally a "brightness function" >calculate the "brightness" of each pixel of each row of the image raster >partition the pixels into two groups based on whether or not their brightness exceeds the threshold you've defined >sort one of the groups based on brightness, don't touch the other one >merge the two groups back together (preserving the positions of the unsorted pixels) the brightness function is arbitrary. anything that will assign a number to each pixel in the image will work. you can even sort each color channel separately. pic related is a gif i made with that method.

>>2022 Yeah it probably isn't actual pixel sorting just looks like it, if i'm remembering the effect correctly. Thanks for the information though. I actually had trouble sticking with computer science because I was more interested in art, so it might be cool to refresh my knowledge on it as a project over this summer focusing on this.

I used audacity and gimp to slowly create a few gifs of the lain wallpapers floating around. One I'm not posting on here. It is on /lain/ already and it is almost 60mb so, yeah. around 124 images in that one.

I have some of this

_.3._.y._.3._

audacity notepad## paint.net _.3._.y._.3._

>>2081 I tried the audacity method from Lainzine3 with no success. https://lainzine.neocities.org/all-releases/lainzine03.pdf page 14 In troubleshooting, I confirmed the tiff is valid. Audacity imports A-law raw data, and the (unchanged data ) exported as headless A-law. I change the extension to .tif and it is not recognized as an image. Comparing the initial part of the file between the original .tiff and after it comes out of Audacity it has changed, and no longer viewable as an image. I was autistic about following the instructions to the letter, so what did I miss? Was something left out of the instructions*? Or maybe the method doesn't work anymore with Audacity? (v.2.1.2) *I found one thing missing from the instructions: - "Do not give a file extension, so Audacity will set it as .raw, otherwise it will automatically pick .aiff EVEN WHEN you have Header=header-less and Encoding=A-law"

>>2082 I use a different method with Audacity 2.2.2 Save the desired image as .BMP Import as raw data into Audacity Select U-LAW encoding and BIG-ENDIAN byte order Select everything past the 0.5 mark Mess with the effects menu (Amplify, Bass, and Treble, Echo, Invert there are a bunch that have cool outcomes) Export (crt+shift+e) audio as a .RAW file Convert .RAW file to .JPEG Hope that's of any help example provided is a result of the Bass, and Treble effect.

>>2082 >tiff if you convert to .ppm you really don't have to worry about the header. it's literally four space-char terminated strings: magic number, height, width, bit depth, then raster. if you fuarrrk it up, you can just put it back in with any text editor. it's huge of course, but so are tiff and raw. http://netpbm.sourceforge.net/doc/ppm.html

on the subject of conversion: i'd argue imagemagick and ffmpeg are the most important tools in the glitch toolbox. a lot of what makes glitch art look like glitch art is carefully selected encoding so that when you get in there with the hex editor or whatever it actually looks interesting and not just waves and cuts and color de-sync. the possibilities are endless... interleaved color channels vs separate rasters, messing with color tables in bitmaps, fuarrrking with keyframes in animation. combined images/blend modes. not to mention the crazy stuff you can do with extreme compression. at the very least, imagemagick makes it easy to stick a header back on if you accidentally corrupt it.

>>2084 >>2083 >Hope that's of any help Yes it was! Thank you lainons! I used PPM, imported as A-law, Little Endian (Big Endian didn't work for A-law, but I didn't try U-law), exported header-less A-law .raw file. Some effects operations seem to invalidate the file (even if the header is not changed). I found imagemagick convert is unhappy unless the extension is one it agrees with, similar to the discussion here https://www.imagemagick.org/discourse-server/viewtopic.php?t=29025 Though I did discover an IM tool I'd not used before (identify). There is another tool I use (montage) to make tiles of many images. montage laintest.jpg -duplicate 4 -geometry 302x -tile 2x2 laintile.jpg -geometry WidthxHeight :Set the size of the tile, using <width>x or x<height> will maintain the aspect ratio of the original image. -duplicate N :Use the initial image N times (if you're not using *.jpg in a directory as usual) Thank you both. test images attached. :)

Sup

>>2086 Another test. >>2087 impressive. ffmpeg I presume? Although I have no idea how.

>>2088 using convert to do duotones & montage for tiling: for color in red green blue yellow ; do convert laintest3.jpg -fill $color -tint 100 laintest3$color.jpg;done montage laintest3red.jpg laintest3green.jpg laintest3blue.jpg laintest3yellow.jpg -tile 2x2 -geometry 604x laintile2.jpg

>>2089 blending in the glitch at 40% to the original image..

>>2090 A higher tint of 80% (than previous 50%) brightens the yellow tile significantly.

>>2091 ...and a wallpaper sized version for you to conclude (for now). toilet -Generate ascii, then screen capture an image of the text convert -Change the background image of the ascii & overlay it on a wallpaper sized version, so no image editing software was used (apart from screen capturing).

>>2093 very nice >>2086 Good good, I remember that feeling when soykaf finally started working.

>>2094 >>2094 Yes, I was frustrated with trying to make the glitch with Audacity and also making glitch art only using IM convert. Now both are working. :)

anyone clued in to some less conventional ways to get glitch effects? i just mean things outside of scripts, imagemagick, hex editing, audacity, etc. those various online format converters/editors are sometimes good for interesting bugs because they're usually CS101 javascript hack-jobs. i can't think of any one in particular; i usually just search for them and pick whichever one looks the most like a relic. pic unrelated

>>2096 oh i thought of another one, rom corruption. this guy makes some cool soykaf https://corruptionasart.tumblr.com/ a little kitschy for my taste, but it's a neat idea.

thx ctrl-c ctrl-v p&i frames

heres a few more.

>>2099 This is what it looks like when I try to watch HD videos on my oldest running computer.

I just connect to BBSs with an actual modem and wait for line noise.

I've only just started getting into this stuff

helo frens >>1888 deep frying memes is fake-glitch, more akin to image "noise" >>2020 read or dl the ready-made package  https://github.com/forkerer/Sorty/blob/master/Data/sceneClass.pde here's the script redditors use https://github.com/satyarth/pixelsort >>2098 hmm >>2099 gj on the 2nd gif

>> 2304 whoops,  s/2098/2099 and then >>2302 not 2099

I have been doing image manipulation in racket including working on a glitch image generation tool. However i am not very familiar on what patterns glitch art has. I got some primitive stuff like contrast, inversion and pixel color value shifting. But im short on ideas, any suggestions welcome.

>>1883 I have this pretty neat technique for glitching videos. I use ffmpeg to stream a video to localhost, then use netcat to caputure that stream and put it into a video. Sounds like it shouldn't be glitching, right? The thing is, my laptop is absolute soykaf in terms of specs so i get this datamoshing-like effect. I discovered it when i was attempting to stream a movie from said laptop to a client on my phone (VLC) but my soykafty performance in combination with my soykafty bandwidth made it almost unwatchable - a side effect was some nice glitching effects. My question is how can i convert this video (the file is basically a broken .mkv) into a non-broken webm but maintain the glitch effects? When i use ffmpeg it converts the video fine but the video is still 'broken' (mpv displays errors while i play it). Is it worth actually making the video a proper file, or do you think that defeats the purpose? I would assume different players/codecs wouldn't be able to play it or the glitch effects wouldn't show the same (correct me if i'm wrong) PS: Send me some links to videos you think would be cool to glitch, i've got no ideas.

>>2302 looks nice, keep at it. >>2304 >deep frying memes is fake-glitch, more akin to image "noise" who's deep frying memes? were you trying to reply to >>2096 ? i don't know what makes it "fake glitch" anyway, aside from it being easy to do. >>2408 try some convolutional algorithms like sharpen, blur, edge-detection/outline effects.

>>2604 Update: using ffmpeg to convert the videos into their same format (i.e. ffmpeg -i file.mkv file2.mkv) does work for some videos but for others audio delay still exists. I think this is because the extent of the 'glitch' isn't constant. So some glitches might fuarrrk up the file more than others. I'll try extracting the audio and video to separate files then putting them back together.

>>2611 Update: >Re-encode the glitched video to a separate file without audio >extract audio from the glitched video to a .aiff >combine into a .mkv This works perfectly. Any pointers to a better audio format? .aiff was chosen because it was the first to pop into my head. aac didn't work, and i don't have lame for mp3.

>>2613 Example

>>2614 The Real Deal

>>2614 Examples of individual frames i turned into .gifs in gimp (added some dithering).

>>1883 What is your personal website / gallery?

First time trying this in a while

Having noticed differences in how JPEG decoders handle out-of-bounds data, I recently made a file that decodes to three different images (or sometimes muddled combinations) depending on what you view it with. I'm curious how it will thumbnail here. code to make it: https://pastebin.com/Zyc3jqBq

Been using paint.net plugins when I want extra control over the artifacts.