-
Notifications
You must be signed in to change notification settings - Fork 0
Capture Guide
Screenshots, posters, videos, and PNG sequences from Fracturing Fog.
Companion pages: User Index · Avalonia User Guide · Regions Guide
There are four ways to save a picture out of Fracturing Fog. Pick the one that matches what you want to do with it, and the rest follows:
| You want to… | Use this button | Format you get |
|---|---|---|
| Post a screenshot to social media | Image | One PNG, ~1-3 MB |
| Print a wall-sized poster, or send to a printer | Poster | One huge TIFF/PNG |
| Share a smooth zoom-in video | Video | One MP4 (or lossless) |
| Loop endlessly between favourite views | Video Slideshow | One long MP4 |
Tip
Before you capture, decide what should be in the picture. Should the grid be on? Should the watermark be on? Should the post-FX sliders (brightness / contrast / adaptive) be applied? All three flow into the saved file exactly as you see them on screen.
- From the Region combo (toolbar), pick Seahorse Valley. The view jumps to the saved spot.
- Optional: switch the Theme combo to Fire for a warmer look.
- Optional: hide the Grid (toolbar toggle) so the image is clean.
- Press
Mto open the Floating Menu, then click Image. - Pick a filename ending in
.png, hit Save. You get a 1920×1080 PNG by default; if your window is bigger, you get the window's resolution.
Note
For a true 3840×2160 (4K) capture independent of your monitor, use Poster instead and dial
in Width=3840, Height=2160, Tile=1024. The Poster path runs the calculator afresh per tile
at full quality — what you save is exactly what the math produced, not what the screen sampled.
- Pan, zoom, or apply a region until you are on the starting frame you want to see.
- Switch on Lock Iterations in the Region Navigation panel if you have already tuned the iter count for the deep target. Otherwise the video may starve detail on the inner frames.
- Open Floating Menu → Video.
- Pick Region = Mini Mandelbrot (or whatever deep spot you want to dive into).
- Duration
30seconds, frame rate30 fps, start zoom0.5. Leave format on Media Foundation MP4 for a quick export, or pick FFV1 (lossless) for archival quality (needsffmpeg.exediscoverable). - Click Render. Each frame is calculated end-to-end; the dialog shows tile-by-tile progress.
Warning
Lossless H.264 / FFV1 / H.264-HQ presets all require ffmpeg.exe. The app looks for it in the
install folder, then Tools\, then your PATH. If it cannot find a working ffmpeg the render
exits with a friendly error explaining what to do.
- Screenshot (Image button)
- Poster (multi-tile)
- Video Zoom (single shot)
- Video Slideshow (continuous loop)
- Recording Formats
- ffmpeg Discovery + Flags
- Resolution Reference
- Batch CLI
- Watermark
- Tips
The Image button in the Floating Menu saves the current view as a single still.
| Behavior | Description |
|---|---|
| Format | PNG (default), TIFF, BMP — chosen by file extension |
| Resolution | Current panel resolution, OR full virtual desktop when Span is active |
| Post-FX | Live brightness / contrast / adaptive applied |
| Watermark | Embedded if the Watermark toggle is on. Contrast-aware text color. |
| Filename | Auto-generated: FracturingFog_Theme_Region_x...y...z...i..._WxH.png
|
The file dialog defaults to your Pictures folder; switch to any path before confirming.
The Poster button renders a tiled composite at print resolution. Each tile is calculated separately at full quality; tiles stitch into one large image.
| Field | Range | Default |
|---|---|---|
| Width × Height | up to 32768 × 32768 | 7680 × 4320 (8K) |
| Tile size | 256 – 4096 | 1024 |
| Format | .png / .tif / .tiff / .bmp | .png |
| Output | file path or folder | (current dir) |
| Apply post-FX | on / off | on |
| Include watermark | on / off | on |
- Pan + zoom to your subject.
- Open Poster.
- Set output dimensions (e.g., 11520 × 8640 = 12K, 15360 × 8640 = 16K wide).
- Pick a tile size — 1024 is the default; smaller tiles parallelise better but increase seam risk; larger tiles use more memory.
- Browse to output path.
- Click Render. Progress bar shows tile-of-total + ETA.
Cancel any time — the partial frame buffer is dropped.
A 64-megapixel ceiling applies by default to protect against accidental OOM (a 32k × 32k render holds ~10 GB resident). Override only if you have the RAM headroom.
Use the Client dialog instead of the local Poster button:
- Pick a saved server connection.
- Mode =
image. - Width × Height = your poster dimensions.
- Pick fractal / region / theme / quality.
- Output: a local file path on YOUR machine.
For huge posters, set Return mode = saved-path. The server keeps the file on its disk and replies with the path; read it later over file share. Inline (default) streams bytes in 1 MB chunks over TLS — fine for a few-MB poster, slow for multi-GB.
The Video button animates a smooth zoom from the current view to the active region's coordinates.
| Phase | Duration | Behavior |
|---|---|---|
| Pan | First 5 % | Pan to target center at current zoom |
| Zoom | Last 95 % | Log-zoom interpolation, center fixed |
Both phases smoothstep-eased.
Calculation-bound, not wall-clock-bound. The loop advances by elapsed wall-clock time so the total duration is honored even if individual frames take longer than 1/fps to render. Drop quality + iter cap if you need consistent fps.
While a video zoom is rendering, three extra sliders appear in the Floating Menu:
| Slider | Range | Purpose |
|---|---|---|
| TAA Alpha | 0 – 1 | Temporal blend strength between successive frames |
| Fade Start | 1e0 – 1e60 zoom | Where the deep-zoom artifact fade begins |
| Fade End | 1e0 – 1e60 zoom | Where the fade reaches full strength |
Use TAA Alpha around 0.3 – 0.6 for cinematic smoothing without ghost trails.
Regions may carry a stored iteration target. During the video leg, MaxIterations is raised to at least that value so the deep target doesn't render as all-in-set black just because the quality preset's iter formula produced a smaller number.
A continuous mode: zoom in → pause → zoom out → next region → repeat.
| Leg | Default | Configurable |
|---|---|---|
| Zoom in duration | 30 s | Slideshow Settings |
| Pause at target | 7 s | Slideshow Settings |
| Zoom out duration | 30 s | Slideshow Settings |
| Inter-region gap | 0 s | Slideshow Settings |
Stops independently from the single-shot Video feature. Esc or the Slideshow button toggles off.
The Video button label flips to Stop while running.
The image slideshow (Slideshow Settings → Type = Image) can now be captured. Tick Record Slideshow in the Slideshow Settings dialog and pick an encode preset. The engine streams every cross-fade interpolation frame into a PNG sequence under your temp folder while the slideshow runs.
When you Stop the slideshow you get a three-button prompt:
| Button | Action |
|---|---|
| Convert | Runs ffmpeg with the chosen preset and saves a video file. Temp PNG folder is deleted on success. |
| Save Frames | Pick a destination folder; the PNG sequence is moved there for offline post-processing. |
| Cancel | Discards the temp PNG folder. |
Encode presets:
| Preset | Codec / container | Notes |
|---|---|---|
| HighQualityH264Mp4 (default) | libx264 -crf 18, yuv420p | Visually lossless, small, broad-compat |
| LosslessH264Mp4 | libx264 -qp 0, yuv444p | Mathematically lossless |
| Ffv1Mkv | FFV1 v3 in Matroska | Archival / editing pipeline |
The Convert path needs ffmpeg.exe — install it via the floating menu's
FFmpeg Setup dialog first. Save Frames works without ffmpeg.
| Format | Container | Encoder | Needs ffmpeg? | Best for |
|---|---|---|---|---|
| None | — | — | No | Live playback only |
| MP4 (built-in) | .mp4 | Media Foundation H.264 | No | Quick exports, browser playback |
| Lossless H.264 | .mp4 | libx264 -qp 0, yuv444p, +faststart | Yes | Mathematical-lossless archive |
| Lossless FFV1 | .mkv | FFV1 v3 in Matroska | Yes | Archival / editing pipeline |
| H.264 HQ | .mp4 | libx264 -crf 18, yuv420p | Yes | Visually-lossless sharing |
| PNG sequence | folder | (sidecar) | No | Offline encoder pipelines |
MP4 (built-in) and PNG sequence can record simultaneously with any video format — useful for keeping a high-quality intermediate while shipping a small MP4 for preview.
For a 60s 1080p 30fps clip:
| Format | Approximate size |
|---|---|
| MP4 (built-in) | 30 – 80 MB |
| H.264 HQ | 80 – 200 MB |
| Lossless H.264 | 1 – 4 GB |
| Lossless FFV1 | 800 MB – 2.5 GB |
| PNG sequence | 5 – 15 GB |
A 4K 60s 60fps clip is ~16× a 1080p 30fps clip.
ffmpeg.exe discovery order:
- The app folder.
-
<install>\Tools\,<install>\Resources\. - PATH.
Missing ffmpeg with a lossless preset selected: exit code 3 (batch) or a clear error toast (UI).
- Render every frame to disk as
frame_NNNNNN.png(image2 demuxer compatible — starts at 000001). - Invoke ffmpeg on the sequence with the preset's argument set. ffmpeg progress feeds a second progress meter.
Lossless H.264 (h264):
-i frame_%06d.png -c:v libx264 -qp 0 -preset veryslow -pix_fmt yuv444p
-movflags +faststart out.mp4
Lossless FFV1 (ffv1):
-i frame_%06d.png -c:v ffv1 -level 3 -coder 1 -context 1 -g 1 -slices 24
-slicecrc 1 -pix_fmt yuv422p out.mkv
H.264 HQ (h264hq):
-i frame_%06d.png -c:v libx264 -crf 18 -preset slow -pix_fmt yuv420p
-movflags +faststart out.mp4
By default --keep-frames is on with none and off with lossless presets (since frames are intermediates). Override:
FracturingFog.exe --batch --mode video … --lossless ffv1 --keep-frames
FracturingFog.exe --batch --mode video … --lossless none --no-keep-frames
In the UI, Slideshow Settings → Video tab has a Keep frames checkbox.
| Target | Pixels | Aspect | Notes |
|---|---|---|---|
| 720p | 1280 × 720 | 16:9 | YouTube minimum |
| 1080p (FHD) | 1920 × 1080 | 16:9 | Standard HD |
| 2K | 2048 × 1080 | ~17:9 | DCI 2K |
| 1440p (QHD) | 2560 × 1440 | 16:9 | 1440p monitors |
| 4K UHD | 3840 × 2160 | 16:9 | Consumer 4K |
| 4K DCI | 4096 × 2160 | ~17:9 | Cinema 4K |
| 5K | 5120 × 2880 | 16:9 | iMac 5K |
| 8K UHD | 7680 × 4320 | 16:9 | 8K TV |
| 8K Cinema | 8192 × 4320 | ~17:9 | |
| 12K | 11520 × 6480 | 16:9 | High-end poster |
| 16K | 15360 × 8640 | 16:9 | Approaching the 32k cap |
| 32K | 32768 × 18432 | 16:9 | Server cap |
Video mode rounds width/height down to the nearest even number (codec constraint).
Poster mode honors odd dimensions but rejects anything above 32768 × 32768 or 64 MP total (hard server cap).
Headless render with full UI parity.
FracturingFog.exe --batch --region ""Seahorse Valley"" --theme Fire ^
--width 3840 --height 2160 --out C:\out\seahorse.png
FracturingFog.exe --batch --mode video --region ""Mini Mandelbrot"" ^
--theme Plasma --seconds 30 --fps 30 ^
--out C:\out\zoom.mp4
FracturingFog.exe --batch --mode video --region ""Seahorse Valley"" ^
--theme Fire --seconds 60 --fps 60 ^
--lossless ffv1 --keep-frames --out C:\out\
FracturingFog.exe --batch --mode video --region ""MyDeepPick"" ^
--theme Inferno --seconds 20 --reverse ^
--lossless h264hq --out C:\out\dive_out.mp4
Drives the same image-slideshow preset the interactive Slideshow Settings dialog edits, then encodes the captured PNG sequence with ffmpeg.
FracturingFog.exe --batch --slideshow ""Default"" ^
--seconds 90 --fps 30 ^
--width 1920 --height 1080 ^
--encode h264hq ^
--out C:\out\slideshow.mp4
Notes:
-
--slideshow NAMEnames a saved preset in%APPDATA%\FracturingFog\slideshow-configs.json. Omit to use the active preset. -
--encodeacceptsh264hq(default),h264(lossless), orffv1. - v1 headless slideshow renders Mandelbrot regions only; non-Mandelbrot regions in the preset's filter set are skipped with a warning.
-
--keep-frameskeeps the PNG sequence at its temp path after ffmpeg succeeds. Default behaviour deletes it.
Every batch render — image, video, and image/video slideshow — paints the watermark (region + theme name over the program label) into each frame by default. This matches what the interactive Save button does, so a poster you render headless looks the same as one you export from the app.
To render without the watermark, pass --watermark (or its clearer alias
--no-watermark). The flag is a switch that turns the watermark off:
FracturingFog.exe --batch --region ""Seahorse Valley"" --theme Fire ^
--width 3840 --height 2160 ^
--no-watermark --out C:\out\seahorse-clean.png
Note
The flag reads backwards on purpose. Because the watermark is now the default,
the only thing left for a flag to do is remove it — so both --watermark and
--no-watermark mean "suppress the watermark". There is no flag to add one;
it is already there.
Scene mode (--mode scene) is the exception: it carries its own watermark
setting inside the saved scene, so this flag does not affect it.
FracturingFog.exe --batch --remote ^
--connection MyDeskstation ^
--render Poster8K ^
--out C:\out\poster.png
Prompts on stdin for the master password (no echo). Exit 0 on success.
Exit codes:
| Code | Meaning |
|---|---|
| 0 | success |
| 1 | unhandled runtime error |
| 2 | bad command-line argument |
| 3 | --lossless selected but ffmpeg.exe not found |
| 4 | ffmpeg encode pass failed |
Toggle from the toolbar Watermark button.
Watermark composition:
| Element | Default | Configurable |
|---|---|---|
| Region name | active region | yes (Slideshow Settings → Watermark) |
| Theme name | active theme | yes |
| Program label | ""Fracturing Fog v0.6.2"" | no |
| Position | bottom-right | yes (corner picker) |
| Opacity | 80 % | yes (slider) |
| Color | auto (contrast-aware) | manual override available |
The watermark is CPU-composited into the BGRA buffer before swap-chain upload, which is why it survives into screenshots, posters, and videos — the GPU never sees an unwatermarked frame when the toggle is on.
For non-watermarked exports, disable the toggle before triggering the capture.
Headless renders follow the same rule but the other way round: the watermark is
on by default for every --batch mode, and you pass --no-watermark to drop
it. See Watermark (on by default) under the Batch CLI
section.
Match poster output dimensions to your printer's target DPI. A 24"" × 16"" print at 300 DPI = 7200 × 4800. At 600 DPI = 14400 × 9600 — approaching the 16K poster ceiling.
For long video zooms, pre-test with low-quality settings. Drop Quality to Draft + Iter to 256, set --seconds short, render to PNG-sequence. Confirms the motion + framing before committing to a multi-hour Ultra render.
ffmpeg's encode pass is single-threaded for libx264 -preset veryslow. Lossless H.264 takes much longer than the render itself for many regions. Use FFV1 instead — it's also lossless, files are smaller, encode is fast.
Browser-friendly MP4 requires yuv420p. The H.264 HQ preset uses yuv420p; Lossless H.264 uses yuv444p (NOT browser-friendly — re-encode to ship to YouTube).
Poster tile size affects memory but not seam quality. Tiles are calculated with consistent coordinate math so a 256-tile poster looks identical to a 4096-tile poster (just renders slower per tile and faster overall on multi-core).
Lock iterations before capture. A capture mid-iteration-promotion can produce subtly different shading frame-to-frame. Tick Lock Iterations + pin a high value before triggering Poster / Video.
Frame rate vs duration trade. Doubling fps = doubling frame count = doubling render + encode time, but only 4 % better motion smoothness at the human-perception level. 30 fps is usually fine; reserve 60 fps for high-motion zoom-outs.
Span mode for wallpaper-resolution screenshots. Click Span → click Image. The output covers the entire virtual desktop, which is usually 2× – 4× a single monitor.
Adaptive sweep on video. Start the Adaptive Sweep just before triggering Video — the sweep animates from 0 → 100 across the recording, giving a slow-build dramatic reveal.
Capture Guide · Fracturing Fog · © 2026