Glitch Image Maker — RGB Split, VHS, Pixel Sort & Datamosh Effects

What is glitch art?

Glitch art is a visual aesthetic that intentionally reproduces or imitates the appearance of digital errors — signal loss, compression artefacts, datamoshing, corrupted files, and RGB channel misalignment. It emerged in the early 2000s alongside internet culture and remix music, and is associated with artists such as Rosa Menkman, Iman Moradi, Ant Scott, and the collective JODI. The look spread into mainstream design through album covers, film title sequences, fashion photography, and social media filters.

Two broad approaches exist. Actual glitching means physically corrupting the bytes of an image or video file — usually with a hex editor — and letting the decoder produce whatever unpredictable result comes out. This is how the earliest glitch art was made, and it remains the purist's method. Simulated glitching means deliberately producing the look of corruption using a controlled rendering pipeline — the approach this tool takes. Simulated glitches are repeatable, adjustable, and safe: you get the artistic result without risking damage to a file or having to learn a hex editor.

The vocabulary of glitch art is drawn from real technical phenomena: RGB channel misalignment from misregistered printing or analog video, scanlines from cathode-ray televisions, colour bleed from VHS tape, block artefacts from JPEG and MPEG compression, and datamoshing from corrupted video encoding. Each of these has a signature visual character, and this tool gives you independent control over each one so you can dial the exact look you want — VHS, cyberpunk, corrupted-file, or something entirely your own.

💡 Quick example. Load any image and the Retro preset appears immediately — a balanced VHS-style glitch with mild RGB split, soft scanlines, a small wave, and a warm colour tint. From there you can push toward the more aggressive Cyberpunk or Corrupted File presets, or dial specific sliders to taste. The point of the default preset is that you always see the effect working from the first second.

How to make a glitch image

1. Add your image. Drag it onto the upload area, click to browse, or copy an image and paste it anywhere on the page. The Retro preset applies automatically — you see a glitched image immediately.
2. Pick a different preset if you want. Ten presets cover the most common glitch looks: Retro (default), VHS 1987, Cyberpunk, Datamosh, Scanlines only, RGB Split, Pixel Sort, Corrupted File, Subtle, and Custom (blank slate). Click any preset to load its parameter set — you can then adjust individual sliders on top.
3. Dial the master intensity. The master slider at 0% disables all effects (identical to the original). At 100% each effect runs at its configured value. Up to 150% exaggerates everything. Use it to fade the whole glitch in or out without touching individual sliders.
4. Fine-tune individual effects. Eight sliders: RGB split amount and angle, pixel sort threshold (with horizontal/vertical orientation toggle), block displacement amount and size, wave amplitude and frequency, scanline intensity and spacing, noise, colour bleed, and vignette. Each updates the preview live.
5. Lock or reshuffle the block layout. The seed input controls the random block displacement pattern. Increasing the seed or clicking Reshuffle gives you a new layout. Everything else — sliders, colours, and intensity — is completely deterministic, so once you find a layout you like, slider changes will not scramble it.
6. Use Randomize if you are stuck. The Randomize button rolls new parameter values within curated ranges and applies them, and also bumps the seed so you get a fresh block layout. Click it a few times to explore the space.
7. Compare with the split view. Drag the vertical divider to see the original on the left and the glitched version on the right. Use the view tabs to see either side full-size.
8. Download or copy. Choose PNG (best for fine RGB fringes), JPEG (universally compatible), or WebP (smallest modern format), then click Download or Copy to clipboard. Copy settings copies a JSON object with every parameter and the seed so you can replicate the exact effect on another image later.

The eight effects, explained

Each slider controls one specific technical phenomenon. Understanding what each one reproduces helps you know when to reach for it.

1. RGB channel split

Also called RGB shift, chromatic aberration, or colour fringing. The red, green, and blue channels are sampled from different positions — usually R shifted right, B shifted left — producing coloured fringes along edges. This mimics the misregistration of analog TV signals, cheap lenses, and misaligned printing plates. Small values (1–3 px) suggest a slightly-off camera; large values (15–30 px) look like the signal is falling apart. The angle slider lets you rotate the split direction — 0° is horizontal, 90° is vertical, 45° gives a diagonal that suits portraits with strong vertical features.

2. Pixel sorting

Pixels within a row (or column, via the orientation toggle) are reordered by brightness while staying in place within that row. The result is long streaks that look like a software bug scrambling the scanlines. Popularised by artist Kim Asendorf's ASDF Pixel Sort algorithm (2010). A threshold slider controls the range of brightness values that get sorted — a low threshold sorts only the brightest pixels (subtle streaks over highlights); a high threshold sorts almost everything (aggressive smears). Horizontal sorting produces horizontal streaks; vertical sorting produces vertical streaks and works better on tall images or portrait-oriented subjects.

3. Block displacement

Random rectangular blocks of the image are horizontally shifted by a fixed amount. This is the visual signature of datamoshing — where a compressed video's motion vectors cause blocks of one frame to persist into the next. Small block sizes with small shifts look like compression artefacts; large block sizes with large shifts look like the image is being torn apart. The layout is set by the seed, and both block size and displacement amount scale proportionally to the image width so the visual look is identical between the preview and the full-resolution download.

4. Wave distortion

Each row of the image is shifted horizontally by a sine-wave offset that varies with the row's position. This mimics the tracking errors and vertical hold problems of analog tape — the wobbling, wave-like displacement you see when a VHS tape plays at the wrong speed or a signal is weak. Larger amplitudes produce more dramatic wobbling; higher frequencies produce more oscillations per image height.

5. Scanlines

Horizontal darkening lines are drawn at regular intervals — mimicking the visible scanlines of a cathode-ray tube (CRT). The spacing determines how dense the pattern is (3–5 px is typical for a VHS look; 1–2 px looks more like a CRT screen photographed close up). Intensity controls how dark the lines are — a value around 30–50% looks like a real CRT; 90% makes the image almost unreadable. Intensity is capped internally so that combining a high value with a high master multiplier never reverses into pure black.

6. Noise / grain

Random pixel-level brightness variation is added across the image. This mimics the sensor noise of high-ISO film or early digital cameras, and the analogue noise floor of tape formats. Small amounts (10–20) add subtle texture; heavy amounts (60+) make the image look actively damaged. Combined with scanlines, noise is the fastest way to make a modern photo look thirty years old.

7. Colour bleed

Each colour channel is multiplied by a per-channel factor, shifting the whole image's colour balance. This mimics the chroma bleed of VHS tape — where the colour signal lags slightly behind the brightness signal, resulting in colour smears to the right of edges. In a still image it manifests as a global colour cast that intensifies toward saturated areas.

8. Vignette

Corners and edges are darkened relative to the centre. Vignetting is a lens artefact — the natural darkening that occurs toward the edges of an image produced by a wide-aperture lens, especially on cheaper cameras. It also mimics the way CRT screens look brightest at the centre. Vignette is the subtlest effect here but it ties a glitched composition together by directing the eye inward.

The ten presets and what they look like

How to pick a preset for your specific image

The right preset depends on three things: what the image is, where it will be viewed, and what mood you want. Here is a starting heuristic:

Why some effects are directional

Most glitch art is directional — it has a horizontal bias. Scanlines run horizontally. Tape tracking errors cause horizontal displacement. Chroma bleed travels horizontally because the colour signal lags the brightness signal along the scanline. Datamosh blocks displace sideways because inter-frame motion prediction works along the horizontal scanning direction.

This has a practical consequence: if your source image is a portrait with strong vertical features (tall buildings, standing figures, doorway framing), horizontal glitch effects will cut across those features in a way that can look awkward. You have three options:

• Use the RGB split angle slider. Setting it to 90° pushes the RGB channels vertically instead of horizontally, giving you a vertical split with the same visual character as the horizontal one.
• Use the pixel sort orientation toggle. Switch from horizontal to vertical sorting — vertical streaks work better on tall images with strong vertical edges.
• Rotate the image before glitching. A 90° rotation turns a horizontal glitch into a vertical one — the effects then run along the natural axis of the subject. Rotate back after.

VHS vs glitch vs datamosh — three different aesthetics

These three words get used interchangeably but describe different things. Knowing the difference helps you pick the right preset or slider combination.

Every VHS effect is a glitch effect — but not every glitch effect is a VHS effect. If you want the 1987 camcorder look specifically, start from VHS 1987. If you want modern glitch art, start from Cyberpunk or Pixel Sort. If you want video corruption, start from Datamosh.

Real glitch art vs simulated glitch art

The tool you are using produces simulated glitches. This is a legitimate and widely used approach — virtually all glitch-art generators on the web work this way — but it is worth understanding the difference from the original technique.

Neither approach is "better" — they produce different aesthetics. Real glitching produces the specific artefacts of a specific decoder handling a specific corrupted stream, which can include effects no simulator produces. Simulated glitching gives you control, speed, and repeatability, which is what most creators actually need.

How glitch art moved from the fringe to the mainstream

Glitch art was a niche internet subculture for its first decade. It started moving into commercial design in the late 2000s, when three things happened at once: music labels began using glitched album artwork, fashion photographers started incorporating digital artefacts as a deliberate aesthetic, and film title sequences (notably the Ghost in the Shell television intro and later the Black Mirror opening credits) demonstrated that glitched visuals could signal a modern, self-aware tone.

By the mid-2010s the aesthetic had become a fixture in three distinct commercial contexts:

• Music artwork. From Kanye West's Yeezus (2013) to the entire visual language of hyperpop, glitch art is now the default visual signature of electronic and experimental music.
• Fashion and editorial. Vogue, Dazed, and 032c have all run glitched photo editorials. The aesthetic reads as "digital-native" without looking dated.
• Brand identity for tech, gaming, and crypto. A subtle glitch treatment on a logo or website hero signals that the company is fluent in the internet without having to say so explicitly.

Twenty years ago, glitching a photo required a hex editor and a lot of patience. Today it is a slider on a browser tool — which is exactly why the aesthetic is everywhere.

Best glitch settings for Instagram, TikTok, and Twitter

Social platforms re-encode every uploaded image, and their compression can destroy the delicate parts of a glitch effect. Instagram, Facebook, Twitter, and LinkedIn all run uploads through a JPEG compressor that prioritises smooth gradients over sharp edges — which is exactly the wrong behaviour for preserving RGB fringes and thin scanlines.

• Keep effects strong enough to survive. RGB split below 3 px often smears to nothing after the platform compresses. 4–6 px survives.
• Use wider scanlines. A 3 px scanline spacing (dense pattern) can blur into grey after re-encoding. A 4–5 px spacing stays crisp.
• Avoid single-pixel noise. Fine grain is the first thing a compressor removes. Either commit to bold grain (30+) or leave it off entirely.
• Download as JPEG at quality 90, not PNG. Counter-intuitive, but uploading a PNG that the platform immediately re-compresses to JPEG produces worse results than uploading a clean JPEG that survives the pass.
• Post to the platform, not a link to the platform. Some sites detect certain file characteristics and apply even heavier compression. Exporting at 1080×1080 or 1080×1350 stays inside the "normal" size range.

⚠ Do not use CMYK JPEGs on social media. CMYK is a print colour space. Social platforms expect sRGB and will produce colour-shifted results if given CMYK. This tool always outputs sRGB, so this is not an issue here.

Performance and image size

Different effects cost different amounts of processing time. On very large images, the slowest effects dominate the total render time. Here is what to expect at different resolutions:

Because pixel sort dominates the total cost, the tool automatically skips rows when the image is very tall, which is why pixel sort at full resolution may show a slightly different (chunkier) pattern than the live preview. The preview runs at a capped resolution so that all sliders update smoothly; when you click Download, the tool re-renders everything at the original resolution and the pixel sort algorithm runs with row skipping at very large sizes.

If you find the tool is slow on a specific image, the two levers that help most are: (1) reduce the source image dimensions with the Image Downscaler before glitching, and (2) temporarily set the pixel sort threshold to 0 while tuning other effects, then raise it once everything else looks good.

Glitch art history and artists worth knowing

If you are drawn to this aesthetic, the following names and works are foundational.

• Nam June Paik (1960s–70s) — a Korean-American artist credited as the first to manipulate a television signal for artistic effect. His distorted TV installations preceded the term "glitch art" by decades but established the foundational idea that a malfunctioning signal could be art.
• JODI (Joan Heemskerk and Dirk Paesmans, 1990s) — the Dutch-Belgian duo whose early website art used misconfigured HTML and broken browser rendering as the medium itself.
• Rosa Menkman — Dutch artist and theorist whose 2011 Glitch Studies Manifesto argued that glitches reveal the invisible structure of digital systems. Her essay remains the most-cited theoretical text on glitch aesthetics.
• Iman Moradi — author of Glitch: Designing Imperfection (2009), the first book-length treatment of glitch as a design language.
• Kim Asendorf — created the ASDF Pixel Sort algorithm in 2010, which is the technical basis of the pixel-sort effect in this tool and in nearly every contemporary glitch-art generator.
• Kanye West — Yeezus album artwork (2013) — a clean example of mainstream glitch art, with the CD case photographed as if the file were physically damaged.

Where people use glitch images

• Album covers and single artwork. The aesthetic has been a staple of electronic, industrial, and experimental music artwork since the mid-2000s.
• Social media posts and stories. A single strong glitch effect makes a photo stand out against a feed of clean photos — a strong signal in the attention economy.
• Event posters and flyers. Glitch is closely associated with electronic music parties, cyber events, and independent film festivals.
• Brand identity for tech, gaming, and crypto projects. The aesthetic signals "we are digital and we know it" without having to state it explicitly.
• Photo illustration for articles about AI, cybersecurity, and digital culture. A glitched portrait or data visualisation is a visual shorthand for the subject matter.
• Personal photo editing. Glitching a favourite photo as a distinct visual treatment rather than another filter.
• Music video stills and animation frames. Stills from animated sequences are often glitched frame-by-frame.
• Website headers and hero images. A subtle glitched background image reads as modern and technical without being noisy.
• T-shirt and merchandise design. Glitch patterns and text treatments print well and are immediately recognisable.
• Experimental film and video art. A frame of a video is glitched, printed, and then used in an installation or film print.

Common mistakes when making glitch art

• Cranking every slider to 100. The most common mistake. Extreme values on every effect produce a chaotic mess that reads as noise, not as glitch art. The best glitch images usually have one or two strong effects and everything else subtle. Use the master intensity slider to reduce the whole effect rather than reversing individual sliders.
• Using glitch on images with no clear structure. Glitch art works best when it is disrupting something recognisable. A subject with clear edges, faces, or text gives the glitch something to interact with. A cluttered or low-contrast image just looks noisy.
• Ignoring the direction of the effects. Horizontal glitch on a vertically-oriented subject creates a mismatch. Either rotate the image first or use the RGB split angle slider and pixel sort orientation toggle to match the direction to the subject.
• Expecting pixel sort to work on every image. Pixel sorting produces the strongest visual effect on images with distinct bright and dark areas. On very noisy or uniformly textured photographs the streaks disappear into the texture. Lower the pixel-sort threshold to focus on bright highlights.
• Not accounting for platform re-encoding. Instagram, Facebook, and Twitter all re-encode uploads. Delicate RGB fringes and fine scanlines can smear away. Either keep effects strong enough to survive (RGB split > 3 px) or download as PNG and post to a service that does not re-compress.
• Forgetting the copyright question. Glitching someone else's copyrighted image does not automatically make the result your own work. For commercial use, start from images you own or from sources explicitly licensed for derivatives.
• Saving as a low-quality JPEG. JPEG compression will blur the exact RGB fringes and pixel-sorted streaks that make the effect work. Save as PNG when you care about the visual result, JPEG only when you need the smallest file.
• Assuming the "real glitch" method is better. Actual file corruption (via a hex editor) produces unpredictable results and can break files irreversibly. Simulated glitch — what this tool does — is controllable, repeatable, and produces the same aesthetic in a fraction of the time.

Privacy and data storage

This tool runs entirely in your browser. Your image is decoded from a local file (or paste event), rendered on the HTML5 canvas with vanilla JavaScript, and re-encoded with the browser's native toBlob() encoder. Nothing is uploaded to any server, logged, cached, or stored anywhere. There is no account, no login, and no network request after the page loads. When you close the tab, the image is dropped from memory. Because the canvas API automatically strips EXIF metadata during the render, GPS coordinates, camera model, and timestamps in the original file are also discarded.

Limitations

• This tool produces a static glitched image, not an animation. Animating a glitch requires generating a sequence of frames — a different workflow that this tool does not attempt.
• The preview is capped at a moderate resolution for performance. When you download, the tool re-renders the effects at your image's full original resolution, so the output is not reduced.
• Pixel sorting is computationally expensive — a full-resolution sort on a large image can take several seconds. The tool automatically skips rows on very large images, which may produce a slightly chunkier pattern than the preview shows.
• The Randomize button produces parameter values within curated ranges, not arbitrary random numbers. This is deliberate — pure randomisation usually produces unusable chaos. If you want genuine chaos, load the Corrupted File preset and push the master slider to 150%.
• This tool does not create real byte-level glitches. A genuine corrupted file behaves according to your specific decoder's tolerance and can look different on every viewer.
• Animated GIF and APNG inputs are handled as static images — only the first frame is processed. For animated glitch, use a dedicated GIF editor.
• The RGB split angle slider and pixel sort orientation toggle allow vertical effects, but wave distortion, scanlines, and block displacement are always horizontal — they cannot be rotated. If you need vertical versions of those effects, rotate the image 90° before applying them and rotate back after.

Frequently asked questions

Related tools on AIToolsPros