What is BLIBBER?
BLIBBER—short for Blip Dialogue Generator—is a browser-based RPG dialogue blip generator and pixel-art NPC creator. It brings portrait processing, voice recording, waveform slicing, typewriter dialogue, sprite animation, and game-ready export into one focused workspace. The goal is simple: turn a portrait and a tiny sound into a character that already feels alive inside a game scene.
BLIBBER is made for indie game developers, game-jam teams, pixel artists, writers, streamers, and anyone prototyping a character-driven scene. You do not need a digital audio workstation, animation timeline, or image editor to test an idea. Record a sound, choose a short section of the waveform, style the portrait, write the line, and press Play Scene to hear and see the pieces together.
Unlike a text-to-speech or AI voice generator, BLIBBER does not invent a speaking performance. It uses the voice, sound, or instrument sample you provide and plays the selected micro-slice at each sounded character. The original pitch remains the default, while an optional pitch control lets you make a deliberate variation. Optional AI is used only for local portrait background removal when you choose it.
The result is an original character package built from your own source material: transparent pixel-art frames, a clean WAV dialogue seed, saved character presets, and an optional WebM recording of the complete scene. BLIBBER keeps the creative loop fast enough for experimentation while exposing advanced controls when a project needs precise palettes, dithering, outlines, position, scale, dialogue motion, or two-frame animation.
What is a dialogue blip?
A dialogue blip is the short sound played as a letter appears in a game text box. Instead of recording every spoken line, a game can reuse one tiny vocal sound to suggest how a character talks. The result sits between silent text and full voice acting: it adds rhythm, identity, and feedback while leaving the exact performance to the player’s imagination.
Different sounds imply different personalities. A soft breath can suit a shy companion. A clipped vowel feels alert or mechanical. A round “oh” can make a merchant sound warm, while a dry click can make a robot or menu system feel precise. Type speed matters too: the same sound feels calm at a measured pace and nervous when letters arrive rapidly.
How to make an RPG dialogue voice
A reliable BLIBBER voice starts with a source that is simple enough to repeat but distinctive enough to identify the speaker. The following workflow works for human voices, creature noises, synthetic tones, percussion, and small instrument samples.
- Record or load a clean source. Use the microphone in BLIBBER or import an audio file. A short vowel, hum, mouth pop, consonant, or instrument note works better than a complete sentence.
- Find one stable moment. Move the yellow selection window across the waveform and listen for a compact sound with no room noise, breathy tail, or sudden click at its edge.
- Keep the original pitch first. Test the voice at 1.00× before changing it. Raise the pitch for a smaller, brighter character or lower it for a heavier voice, but make the choice intentionally.
- Write a line with punctuation. Commas and sentence endings create tiny pauses that keep the dialogue from sounding like a machine gun.
- Match the portrait motion. Use one sprite for a clean shake on sounded characters, or add a second talking frame for a more animated mouth-flap effect.
- Export for the destination. Download the package for reusable PNG and WAV assets, or record the complete scene as WebM when you want a shareable performance.
Why short voice slices work
BLIBBER limits the selected dialogue seed to 10–150 milliseconds. That is long enough to preserve a recognizable tone but short enough to behave like a letter sound instead of a spoken word. The whole yellow window can be dragged, so you can keep its length while searching the recording for a cleaner moment.
Start around 40–80 milliseconds. Very short slices feel percussive and suit robots, skeletons, interfaces, and energetic characters. Longer slices carry more breath and vowel shape, which can suit dramatic or gentle voices, but they may overlap when the type speed is fast. If playback sounds crowded, shorten the slice or slow the text before changing its pitch.
Microphone recordings are normalized after capture so quiet takes reach a useful level without adding an 8-bit filter. The dialogue seed keeps its source character and original pitch by default. Normalization cannot remove fan noise, echo, or clipping already present in the recording, so a close microphone and a quiet room still make the biggest difference.
How to turn a portrait into a pixel-art NPC
Import a portrait, illustration, or transparent PNG into the sprite lab. BLIBBER asks whether you want to keep the original image or remove its background because automatic segmentation is useful for photos but can damage artwork that already has good transparency. Keeping the original is the safe choice for finished PNG sprites.
The pixel grid controls how much detail survives. A 32-pixel grid produces a chunky, icon-like character; 64–128 pixels keeps more facial information; larger grids preserve clothing, hair, and silhouette detail. The custom value accepts unusual sizes for experiments, while the live pipeline keeps processing within a safe working resolution.
Next, choose a three-color palette from the visual gallery or create your own. Palette Strength blends the selected colors with the original image, so 100% creates a strict three-color sprite and lower values preserve more of the source. “None / Original Color” skips fixed color quantization while still allowing dithering. Bayer patterns create regular pixel texture, while error-diffusion methods spread color decisions more organically. A configurable outline can separate the finished character from a busy scene.
One sprite or two sprites?
One sprite is the fastest route and the default in Simple Mode. Each sounded character gives the portrait a short, pixel-snapped shake, then it settles back into place. This works especially well for small portraits, creature sounds, and minimalist games because the motion makes the voice feel responsive without requiring another drawing.
Two-sprite mode is available in Advanced Mode. Add an idle frame and a talking or action frame; while dialogue prints, BLIBBER alternates between them and adds the same per-character movement. When the line ends, the scene returns to the idle frame. Two frames are enough for a convincing mouth flap, blink, bounce, or magical pulse when their silhouette stays aligned.
You can drag and scale the sprite directly in the live scene. For clean animation, position both source images similarly before import and keep the face or torso anchored to the same place. Large changes between frames read as jumping rather than speaking.
Writing dialogue that feels good in a game
Game dialogue is read while the player watches motion and hears repeated sound, so shorter lines usually carry more impact. Give each text box one clear idea. Use a comma when the character should hesitate briefly, and a period, ellipsis, or exclamation mark when the beat needs room to land. The type-speed slider runs from slow on the left to fast on the right, making it easy to audition the same line at different rhythms.
Inline effects should support meaning. Wrap one anxious phrase in [shake]...[/shake], use [wave]...[/wave] for dreamy or supernatural speech, and use [color=#ff394a]...[/color] to emphasize a warning, item, name, or choice. If every word moves or changes color, no word feels important. Clean motion is the default for exactly that reason.
When you are stuck, Random Idea inserts editable character dialogue that demonstrates punctuation, color, shake, and wave tags in context. Treat it as a starting beat: decide who is speaking, what they want from the player, and what changes after the line. A useful NPC line either reveals character, points toward an action, or changes the mood of the scene.
Where BLIBBER fits in a game workflow
During early character design, BLIBBER can answer questions that a static concept sheet cannot. Does the merchant feel better with a slow, round voice or a sharp, rapid one? Does a single shaking portrait communicate enough energy, or does the character need a second talking frame? Does the strict three-color palette preserve the expression at the final in-game size? Because the live scene combines sound, text, and portrait motion, those decisions can be tested together instead of in isolation.
For game jams and prototypes, use BLIBBER to create a compact visual and audio kit, then download the character package. PNG frames can be brought into a sprite or UI workflow, while the WAV seed can be triggered by a dialogue system in the target engine. The exported assets are intentionally separate from the BLIBBER preview, so developers can reproduce the timing and presentation in their own code rather than depending on a proprietary runtime.
For pitch decks, devlogs, and social posts, Record Scene captures the composed BLIBBER performance as WebM. This is useful when the feeling of the character matters more than the raw files: viewers can see the portrait react, hear the dialogue blip, and read the authored timing in one clip. Named presets make it possible to return to several characters without rebuilding each palette, transform, and dialogue setup.
BLIBBER is not a replacement for a full animation package or audio editor. It is a specialized character-voice workbench built around the small but important moment when text appears and an NPC seems to speak. Keeping that scope narrow is what makes the tool quick to learn in Simple Mode and deeply adjustable in Advanced Mode.
What local browser processing means
BLIBBER processes portraits, recordings, waveform selections, animation, and exports in the browser. Your media is not sent to an image-processing or audio-rendering backend. Optional AI background removal uses a model on your device through browser-compatible machine-learning tools, with WebGPU acceleration when available and a compatibility path when it is not.
The downloaded model is cached by the browser so it does not need to be fetched again for every portrait. Workspace autosaves and named character presets are also stored locally on the device. Clearing site data or using a private browsing session can remove that saved work, so export a character package when you need a portable backup.
The ZIP package contains game-ready visual frames and a WAV dialogue seed. Scene recording captures the composed canvas and audio as WebM, a format well supported for browser recording and social previews. If a particular engine requires another container, the exported source assets remain available for conversion or direct integration.
Questions about BLIBBER
What should I record for a dialogue blip?
Record a steady vowel such as “ah,” “oh,” or “mm,” a mouth click, a short hum, or a simple instrument tone. Avoid a full word: repeated syllables can accidentally make every line sound like the same sentence.
Is BLIBBER an AI voice generator?
No. BLIBBER turns a short section of your own recording or uploaded audio into a repeatable RPG dialogue sound. It does not clone a voice, generate speech, or upload audio for synthesis. The optional local AI feature is for portrait background removal, not voice creation.
Can BLIBBER make retro, Undertale-style dialogue sounds?
BLIBBER can create the short repeating voice-blip technique associated with many retro and text-heavy RPGs. It does not include protected characters, voices, music, or artwork from any commercial game. Import your own portrait and audio to build an original character with a classic text-box rhythm.
Can I use a transparent PNG without removing its background?
Yes. Choose Keep Original when the import prompt appears. Existing transparency is preserved and the image continues through pixelation, dithering, palette, and outline controls.
Why does the audio sound quiet or distorted?
A quiet source may still contain a loud click or clipped peak that limits normalization. Record close to the microphone at a steady level, avoid touching the device, and choose a section with an even waveform. BLIBBER does not add a bitcrusher to the voice.
Do I need WebGPU to use BLIBBER?
No. WebGPU can accelerate compatible local background-removal work, but the core portrait editor, Canvas effects, waveform tools, dialogue playback, and exports use broadly available browser APIs. Performance and media-format support still vary by browser and device.
What is the best pixel-grid size?
There is no universal best size. Start at 64 for a classic portrait, move toward 32 for a bolder icon, or toward 128 and above when facial or costume detail matters. Judge the result at the size it will actually appear in your game.
What does BLIBBER export?
Download Character Package creates a ZIP containing processed PNG character frames and a source-quality WAV dialogue blip. Record Scene captures the live composition as WebM. Export availability depends on the browser, and you remain responsible for having permission to use the portraits and sounds you import.
Can I save more than one character?
Yes. Advanced Mode includes named presets for different characters and setups, while autosave keeps the current workspace available after a normal refresh on the same browser and device.
Does BLIBBER upload my portrait or voice?
No. Portrait processing, audio decoding, waveform selection, dialogue playback, autosave, and export happen locally in the browser. Clearing the browser’s site data can remove autosaves and cached models, so download important character packages as backups.