Frequently Asked Questions

macOS 15.6 (Sequoia) or later.

It will run, but performance will be significantly lower. FITS Blaster uses Apple's Metal GPU framework for stretching, star detection, and metrics computation. Metal works best on Apple silicon (M1 and later). Intel Macs with integrated graphics lack the GPU memory bandwidth to keep up with large image sets; loading times will be several times slower.

No. FITS Blaster is a native macOS application and uses macOS-specific frameworks (Metal, Accelerate, AppKit). There are no plans for a Windows or Linux port.

No. Image loading, stretching, and all quality measurements work entirely offline. However, FITS Blaster checks for updates once every 24 hours in the background. This is a lightweight check and can be skipped or disabled in the update prompt. If no connection is available, the check is silently skipped and the app works exactly the same.

Yes. Download the M31 sample set (255 MB zip): eight raw 120 s light frames of the Andromeda Galaxy from one night, taken with a one-shot colour camera (ATR2600C, RGGB Bayer pattern) on a William Optics Zenithstar 73.

Unzip it, click Open Folder… in the toolbar, click Open Folder… again in the options popover, and select the FITS-Blaster-Sample-M31 folder. The included README.txt lists how PixInsight's SubframeSelector rates each frame, so you can compare its ranking with FITS Blaster's. Two frames were taken through thin cloud.

There are three ways to load images into FITS Blaster:

1. Open Folder… button or ⌘O — a small popover appears with two options: Include files from subfolders and Include REJECTED folder (loads previously rejected frames, marked as rejected). Click Open Folder… in the popover, then choose the folder. The options start from your defaults in Settings → Files & Folders; changes in the popover apply to this folder only.
2. Drag & drop onto the window — drag a folder or any number of FITS files from Finder and drop them onto the FITS Blaster window.
3. Drag & drop onto the app icon — drag files or a folder onto the FITS Blaster icon in the Dock or in Finder to open them directly, even if the app is not running yet.

FITS files with integer pixel data: BITPIX 8 (8-bit), BITPIX 16 (16-bit), and BITPIX 32 (32-bit). Floating-point FITS (BITPIX −32 or −64) are not supported and will be skipped with an error message. The supported extensions are .fits, .fit, and .fts.

Yes. FITS Blaster installs Quick Look extensions, so Finder shows thumbnails for FITS files, and pressing the spacebar on a file shows a stretched preview. This also works for files the app itself doesn't load: floating-point FITS (BITPIX −32 and −64) and 3-channel colour FITS such as stacked integrations, including files over 1 GB.

FITS files in iCloud Drive only get a thumbnail once they are downloaded to your Mac. To keep thumbnails for an iCloud folder, choose Keep Downloaded on it in Finder.

Not at this time. FITS Blaster is purpose-built for FITS files as used in astrophotography. RAW camera formats (CR3, NEF, ARW, etc.) are not on the roadmap.

In practice, as many as your RAM allows. It has been tested with up to 592 images without issues on a MacBook Air M1 with 16 GB — see the Speed & Performance section below for details.

Yes. Click Open Folder… (or press ⌘O) and turn on Include files from subfolders in the popover. FITS Blaster will scan recursively and display each subfolder as a separate section in the sidebar. By default, calibration folders (FLAT, DARK, BIAS, CALIB) are skipped automatically; you can customise the exclusion list in Settings → Files & Folders.

When FITS Blaster loads images from a session with multiple filters (for example Ha, OIII, and broadband frames in the same folder), it reads the filter name from the FITS header and automatically groups frames by filter in the sidebar. Each group gets its own colour-coded section, making it easy to assess quality separately per filter — a frame that is fine for Ha may still be worth rejecting for OIII.

If no filter information is present in the headers, all frames appear in a single group.

Greyscale loads and renders much faster and is usually all you need for judging star shape, focus, and trailing. FITS Blaster always loads greyscale first, so you can start culling immediately regardless of display mode.

Colour mode debayers OSC (one-shot colour) images and shows them in full colour, which can be useful for checking gradients, satellite trails, or colour artefacts. Colour rendering takes roughly 2–3× longer per image. Both the colour and greyscale versions are kept in memory once loaded, so switching between them is instant.

Mono (greyscale) cameras are always displayed in greyscale regardless of the display setting.

Never. FITS Blaster is strictly read-only. Pixel data is loaded into a separate display buffer; the original files are never written to. The only file operation FITS Blaster performs is moving rejected frames into a REJECTED/ subfolder — and that move is fully undoable.

Reset clears the current session so you can immediately open a new folder. It does not delete or move any files.

Three ways:

• Zoom slider — the slider in the display controls bar below the image (range 0.25×–4×). Click the zoom icon or the percentage label to jump to 1:1 (100 %).
• Zoom to fit — the fit button (↖↘ arrows) next to the zoom percentage scales the image to fill the viewer without scrolling.
• Trackpad pinch — pinch to zoom in or out directly on the image.

The slider and pinch keep the current centre of the viewport fixed as you zoom, so you don't lose your place.

Use the zoom anchor. Click anywhere on the image to pin that point — a small crosshair appears. All subsequent zoom changes (slider or pinch) will keep that point centred in the viewport.

• Move the anchor: click a different spot.
• Clear the anchor: click the crosshair again (within about 20 px of it), or press the Defaults button in the display controls bar.
• Re-centre after scrolling: click the lit scope (⊕) icon in the zoom controls — it scrolls the anchor back to the centre without clearing it.

The anchor persists across image navigation. Because all frames in a session cover the same field, the same position points to the same star or region in every frame — useful for comparing star shape across a run.

The Target Background slider sets the desired brightness of the sky background in the displayed image, on a scale from 0.05 to 0.50.

FITS Blaster's automatic stretch (STF) always targets a background level of ~0.25, so the default value of 0.25 produces no visible change. Adjust it when you want a different look:

• 0.10–0.15 — darker sky, more contrast; good for images with a relatively dark background.
• 0.25–0.40 — brighter background; useful for images with large extended objects (wide nebulae, galaxies) where you want to preserve mid-tone detail.

The value persists across app restarts. This control is analogous to the Target Background parameter in PixInsight's Multiscale Adaptive Stretch.

The Stretch slider is a post-stretch contrast multiplier (0.5×–3.0×, default 1.0×). Increasing it brightens mid-tones and brings out faint detail; decreasing it flattens contrast. Unlike Target Background, it resets when you open a new session.

Each frame receives a coloured badge based on how its metrics compare to the rest of the session:

• Green — the frame is within normal range for this session
• Amber — one metric is noticeably worse than average (trailing, focus drift, low star count)
• Red — the frame is significantly below average and is a candidate for rejection
• Grey — metrics could not be computed (no stars detected, or saturated field)

The badge priority is: trailing stars → focus failure → low star count.

Full Width at Half Maximum — a measure of how sharp (tightly focused) the stars are. It is the diameter of a star profile at half its peak brightness, in sensor pixels. Lower is better: a well-focused session might have FWHM 3–5 px, while a poorly-focused or turbulent frame might be 8–10 px or more.

FITS Blaster fits a Moffat β=4 profile (the same default as PixInsight's SubframeSelector) independently along the X and Y axes, then reports the geometric mean. Results agree with PixInsight Moffat4 within ±10% under typical conditions.

Eccentricity measures how elongated the stars are (0 = perfectly round, 1 = infinitely elongated). A value above ~0.5 usually indicates trailing from poor polar alignment, wind shake, or a mechanical issue.

FITS Blaster fits a 1D Moffat β=4 profile along four directions (0°, 45°, 90°, 135°) and uses the measurements together to reconstruct the true major and minor axes of the PSF ellipse analytically — at any orientation. Eccentricity is then e = √(1 − (FWHM_minor / FWHM_major)²). This approach detects elongation at any angle with equal sensitivity, including diagonal trailing.

Colour (OSC/Bayer) cameras: each star crop is bilinearly demosaicked using adaptive per-channel sky weights before the four-direction Moffat fit. Adaptive weights balance Ha and OIII signal in dual-narrowband images, so the axis-ratio measurement is accurate regardless of filter type. Eccentricity precision on OSC cameras is equivalent to a mono camera.

Signal-to-Noise Ratio — the peak star signal divided by the background noise level. Higher is better. SNR depends heavily on exposure length, sky background, and seeing conditions. It is most useful for comparing frames within the same session rather than between different nights or setups.

PSF Signal Weight is a metric computed alongside the Moffat PSF fit. It measures peak amplitude per unit of sky noise, rewarding both good transparency and good seeing.

The formula is PSF-SW = mean(A) / σ_sky, where A is the background-subtracted sub-pixel peak brightness of each star and σ_sky is the sky noise from Step 1.

Good transparency delivers more photons, raising A. Good seeing concentrates photons into a narrower PSF, also raising A. As a result:

• High PSF·S — good transparency and/or good seeing. Stars are bright and sharp relative to the sky noise.
• Low PSF·S — haze, thin cloud, poor seeing, or high sky background. The frame was photon-starved or the PSF was bloated.
• Unlike SNR, PSF·S uses the star's peak rather than total aperture flux, so it is not diluted by a large aperture or high background.

PSF Signal Weight appears in the session chart (PSF·S button), the inspector panel, and the sidebar sort menu. It is a supplemental metric and does not contribute to the composite quality score.

FWHM is not a standardised measurement — every tool makes different choices about the PSF model, fitting method, and unit convention. Agreement within ~10% is normal. The most common causes of larger differences are:

• Different PSF models: FITS Blaster and PixInsight use Moffat β=4; Siril fits β as a free parameter (typically 3–4). When β < 4, Siril's free-β fit finds a tighter core and reports a lower FWHM.
• Gaussian vs Moffat: A Gaussian PSF has lighter wings than a Moffat profile, so Gaussian FWHM estimates are often slightly larger.
• Units: Siril reports FWHM in arcseconds by default. To convert: FWHM (px) = FWHM (″) / plate_scale, where plate_scale = 206.265 × pixel_size_µm / focal_length_mm.
• 1D vs 2D fitting: FITS Blaster fits two independent 1D profiles and takes their geometric mean. PixInsight and Siril fit a full 2D PSF simultaneously.

For frame ranking (the primary use case), absolute accuracy matters less than consistency. FITS Blaster is self-consistent within a session and correlates reliably with actual image quality.

Yes. FITS Blaster reads the BAYERPAT FITS header (also COLORTYP and CFA_PAT) and, when a Bayer pattern is detected, bilinearly demosaicks each star crop to a full-resolution luminance image before fitting the Moffat profile.

FWHM and eccentricity use different luminance weights, chosen to match how each metric is effectively measured by PixInsight SubframeSelector:

• FWHM uses ITU-R BT.709 weights (wG = 0.7152). Because G pixels make up 50 % of any Bayer mosaic, PixInsight's PSF fit is naturally G-dominated. Matching this weighting gives FWHM rankings that agree closely with PixInsight's (Spearman r ≥ 0.95 on tested narrowband and broadband datasets).
• Eccentricity uses adaptive per-channel weights estimated from the outer-border sky brightness of each crop. This correctly balances Ha and OIII signal in dual-narrowband images (such as an Antlia ALP-T filter), where all three Bayer channels carry real star flux.

FWHM values are at native pixel scale — no scale-back factor is applied.

1. Open your folder — use ⌘O or drag it onto the window. FITS Blaster starts loading and measuring quality metrics immediately; you can begin reviewing before the full set has loaded.
2. Scan the session chart — the horizontal strip across the bottom plots all frames in time order, coloured by quality. A cluster of red or amber dots usually points to a specific problem period (clouds, a focus drift, a gust of wind).
3. Review suspect frames — click a dot in the chart to jump to that frame, or step through the sidebar with the arrow keys. The main view shows the stretched image; the inspector on the right shows the exact metric values.
4. Collect suspect frames — drag across the chart, or ⇧+click a range in the sidebar then ⌘+click to add them to the Flagged view. Auto-Flag can also flag frames below a quality threshold automatically.
5. Inspect and reject — switch to the Flagged view in the sidebar to see only your flagged frames. Step through them, then press your reject key (default: X) to move the bad ones to REJECTED/.
6. Hand off to your stacking tool — point PixInsight, Siril, or AstroPixelProcessor at the same folder. The rejected frames are now in REJECTED/ and will not be picked up by a folder scan.

The session chart is a horizontal strip showing one dot per frame, in the order they were captured. Dot colour represents the filter group (Ha, OIII, SII, Lum, RGB, and so on) — each filter gets its own consistent colour so you can see at a glance whether a quality problem affects one filter or all of them. You can:

• Click a dot to jump to that frame (the dot turns white to mark your position)
• Drag across a range of dots to add those frames to the Flagged view for closer inspection
• Filter by folder or filter group using the pill buttons above the chart, to focus on one subset at a time
• Choose the metric to plot with the buttons above the chart: FWHM, Ecc, SNR, Stars, Sky BG, PSF·S, or Score
• Switch between dots and bars with the small button at the end of that row; your choice is remembered

The chart makes it easy to spot patterns — a dip in a single filter's dots usually points to a filter-specific problem, while a decline across all colours suggests clouds, wind, or a tracking issue.

Dot colour — filter group
Each dot is coloured by its filter group. This lets you compare filters side by side and immediately see whether a quality dip is filter-specific or affects the whole session.

Dashed horizontal lines — per-group medians
A dashed line is drawn at the median value of the active metric for each filter group. Use it as a reference: dots noticeably above or below the dashed line for their group are outliers worth inspecting.

Dot colour and size — cursor and flagged state

• White dot (larger) — the frame currently shown in the viewer (the cursor). Only ever one at a time.
• Full brightness (filter colour) — flagged frames and the cursor when a flagged set is active; all non-rejected frames when no flagged set exists.
• Dimmed (default 50 %) — non-flagged frames when a flagged set is active, so the flagged frames stand out clearly.
• Strongly faded (default 15 %) — frames that have already been rejected and moved to REJECTED/.

Both opacity levels are adjustable in Settings → Display → Session Chart Brightness.

Select a frame in the sidebar and press the reject key (default: X). You can customise this and all other key bindings in Settings → Interface.

Yes. Use ⇧+click or ⇧+arrow keys to select a range (highlighted in orange), then press the reject key. The range can be built in the All view or the Flagged view — the behaviour is the same.

The Flagged view (the segmented control at the bottom of the sidebar) shows a curated pile of frames you have collected for closer inspection, without committing to rejection yet. Think of it as a review queue.

Selection model — two independent concepts:

• Cursor — the single frame currently shown in the viewer. Plain click or arrow keys move it. It never affects the review queue.
• Range selection — an orange-highlighted group of frames built with ⇧+click, ⇧+arrow keys, ⌘A (select all visible), ⌘D (clear), or ⌘I (invert). Used for batch operations (add to queue, remove from queue, reject).

Building the review queue (Flagged view):

• Press F to toggle the flag on the current entry (or the entire orange range if one is active). Press D to deflag everything at once.
• In All view: ⇧+click or ⇧+arrow to select a range, then ⌘+click or F to add it to the queue.
• Drag across the session chart to add a run of frames to the queue directly.
• Auto-Flag — adds all frames that fall below your chosen quality thresholds.
• The (De)flag toolbar button toggles the flag state of the current entry or range at any time.

Working inside the Flagged view:

• Plain click — move the cursor (view that frame); clears any orange range.
• ⇧+click or ⇧+arrow keys — select an orange range for batch operations. ⌘A selects all visible, ⌘D clears the range, ⌘I inverts it.
• Reject key — rejects the orange range (or the cursor if no range). Rejected frames stay visible so you can undo immediately by pressing Reject again.
• ⌘+click on orange — removes the entire orange range from the queue (no rejection).
• ⌘+click on a plain frame — removes that single frame from the queue.

The review queue persists when you switch between All, Flagged, and Rejected views, and is cleared only when you open a new folder or use Deselect All.

They are moved into a REJECTED/ subfolder inside the folder they came from. No files are deleted. You can undo a rejection at any time.

It depends on the Single key reject/undo (toggle) setting in Settings → Interface:

• Toggle on (recommended): The same key both rejects and un-rejects. Press it once to reject a frame, press it again to undo. This is the simplest workflow — many users assign the spacebar.
• Toggle off: Reject and undo are separate key bindings, each configurable independently.

In either case, if the REJECTED/ folder becomes empty after an undo, it is removed automatically.

Use the sort dropdown in the sidebar toolbar to choose between filename (the default, reflecting capture order) or any quality metric: FWHM, Eccentricity, SNR, Stars, Sky BG, PSF Signal, or Score. Rejected sorts rejected frames to one end. The up/down arrow button next to it reverses the sort direction. Sorting by quality score with the worst frames at the bottom makes it quick to select and reject the tail of the list as a group.

Simple mode hides all numeric metrics and shows only the visual quality badge. It is useful for a quick visual pass or for users who prefer not to interpret numbers.

Geek mode computes the quality metrics, shows them in the inspector panel on the right, and enables the session chart. For a compact row of metric values above the chart, turn on Show metric chips above session chart in Settings → Display. Toggle between modes with the configured key (default: G) or the toolbar button, which shows the current mode name (Simple or Geek).

Auto-Flag lets you set quality thresholds and preview how many frames match before applying. The matching frames are added to the Flagged view so you can inspect them individually and decide which ones to actually reject. Access it from the toolbar.

Thresholds work in one of two modes:

• Relative (default) — compared with the median of the frame's filter group, e.g. flag frames whose FWHM is more than 1.5× the group median. This adapts automatically to each night's seeing.
• Absolute — fixed values, e.g. flag frames with FWHM above 3.5 px.

Available criteria: FWHM, eccentricity, star count, SNR, sky background, PSF Signal Weight, and (absolute mode only) quality score. Sky background and PSF Signal Weight are off by default; they are the most useful criteria for catching haze, thin cloud, and moonlight, which barely change star shape.

Point your stacking tool (PixInsight, Siril, AstroPixelProcessor, etc.) at the same folder you opened in FITS Blaster. Rejected frames have been moved to a REJECTED/ subfolder and will not appear in a normal folder scan. No additional steps are needed — your stacking tool will see only the frames you kept.

If you change your mind, reopen the folder in FITS Blaster with Include REJECTED folder turned on in the Open Folder popover, select the rejected frames, and undo the rejection before you start stacking.

Yes. Click Export… in the toolbar to save the loaded frames as a file. The export sheet lets you choose:

• Format — CSV, tab-separated (TSV), or plain text. CSV and TSV are for spreadsheets and scripts; the plain-text report is padded into aligned columns for reading in a text editor.
• File paths — absolute paths, paths relative to the folder all frames share, or filenames only.
• Include rejected frames — adds rejected frames with a status column (kept / rejected).
• FITS header columns — any header keys you want as extra columns, such as DATE-OBS, EXPTIME, FILTER, FOCUSPOS, or HUMIDITY. Keys that don't appear in any loaded file are dimmed.

In Geek mode the export includes FWHM, eccentricity, SNR, star count, PSF Signal Weight, and score for each frame. Numbers always use a decimal point, whatever your region settings, so the files open correctly in any spreadsheet. Set your defaults in Settings → Export; changes in the export sheet apply to that export only.

Tests run on a MacBook Air M1, 16 GB, images on internal SSD. FITS Blaster was restarted before each test.

592 colour images (IMX571 OSC), 51.9 MB each, 30.72 GB total:

Memory figures are peak during loading. In greyscale modes the footprint drops significantly once loading completes (to ~655 MB for Geek, ~590 MB for Simple) as intermediate GPU buffers are released.

290 greyscale images (IMX585 mono), 16.8 MB each, 4.88 GB total:

You can start working immediately once loading begins — you do not need to wait for the entire set. In colour mode, FITS Blaster renders the full set in greyscale first so you can start culling straight away; both versions are kept in memory so switching between them is instant.

1. Load the folder in greyscale + geek mode. Greyscale loads faster than colour; geek mode computes quality metrics and populates the session chart.
2. Glance at the session chart — drag-select any obvious bad runs to flag them, then use Auto-Flag to catch anything you missed.
3. Switch to the Flagged view. Step through the flagged frames and press the reject key on the ones you want to move to REJECTED/.
4. For the remainder, switch the sidebar sort to quality score. Your worst frames are now at one end; ⇧+click to select them as an orange range, then press the reject key.

If you only want a quick visual pass without metrics, simple mode loads faster but the session chart is not available.

Almost. Arrow keys navigate; ⇧+arrow extends the orange range selection; ⌘A / ⌘D / ⌘I select all, clear, or invert the range; F toggles the flag on the current entry or range; D deflags everything; your configured reject key rejects or toggles. The only things that require the mouse are drag-selecting a range in the chart and clicking filter pills.

Yes — FITS Blaster holds no locks on your files. However, avoid undoing rejections while a stacking tool is actively reading the same folder, as moving files mid-stack may confuse the stacking tool.

Yes, in the version downloaded from this website (not the Mac App Store version). Run the app's executable with a folder path, and FITS Blaster measures every frame without opening a window and writes a CSV file with the results:

"/Applications/FITS Blaster.app/Contents/MacOS/FITS Blaster" ~/Astro/M31/LIGHT

The folder path must start with / or ~. By default the CSV is written to frames.csv inside that folder. Progress goes to the terminal, and the path of the CSV is printed when it's done, so you can use it in scripts. Your FITS files are never moved or changed.

Add filter criteria to export only the good frames. As in Auto-Flag, thresholds are relative to the filter group's median by default:

"/Applications/FITS Blaster.app/Contents/MacOS/FITS Blaster" ~/Astro/M31/LIGHT -f 1.5 -e 0.5 -p 0.5 -o ~/Desktop/m31-keepers.csv

This keeps frames with FWHM at most 1.5× the median, eccentricity at most 0.5, and PSF Signal Weight at least half the median. The most useful options:

• -o output file  ·  -S include subfolders  ·  -r include the REJECTED folder (marked status=rejected)
• -m rel|abs threshold mode  ·  -f FWHM  ·  -e eccentricity  ·  -n stars  ·  -s SNR  ·  -b sky background  ·  -p PSF Signal Weight  ·  -q score (absolute mode)
• -P rel|abs path style  ·  -H DATE-OBS,FILTER,… FITS header columns

Run it with --help for the full list. For the beta, use FITS Blaster Beta.app in the path. If you get “permission denied” for a folder in iCloud Drive, give FITS Blaster Full Disk Access in System Settings → Privacy & Security → Full Disk Access.

Yes, completely free. No subscription, no paywall, no frame limit. All features — metrics, Auto-Flag, session chart, subfolders — are available to everyone.

If FITS Blaster saves you time, a small donation via Ko-fi is very much appreciated. It helps cover the cost of the Apple Developer account and motivates continued development.

Open an issue on the GitHub Issues page. A short description of what happened, what you expected, and which macOS version you're on is all that's needed.

Use the feature request template on GitHub. Describing the problem you're trying to solve (rather than just the solution) makes it much easier to evaluate.

No. All processing happens on your Mac. Your image files never leave your device.

None. FITS Blaster does not collect analytics, telemetry, or usage data of any kind. The only network request is the daily update check in the version downloaded from this website, which fetches the list of available versions and sends nothing about you or your images. The Mac App Store version makes no network requests at all; the App Store handles its updates.

No. FITS Blaster only reads the folders you open, and the only change it ever makes on disk is moving rejected frames into, and back out of, a REJECTED/ subfolder.

• Mac App Store version — sandboxed by macOS: it can only access folders you choose in the open panel or drag onto the window.
• Version from this website — not sandboxed, because the command-line mode needs to read any folder path you give it. It still only reads the folders you open. macOS privacy protection still applies: for folders in iCloud Drive you may need to give FITS Blaster Full Disk Access in System Settings → Privacy & Security.