Astrophotography storage calculator: FITS file size per session

One 16-bit FITS frame from a 6248 × 4176 sensor is 52.2 MB. Four hours of 180-second subs plus 90 calibration frames come to about 8.7 GB from the camera, and processing can multiply that several times over: debayered 32-bit copies of a colour camera's lights are six times the size of the originals.

Published October 15, 2026 . Sources are listed at the end of the page.

Camera

Imaging area = pixel count × pixel size where both are published; otherwise the published sensor size.

Session
Processing
One frame (FITS)
52.2 MB
From the camera
8.66 GB166 files from the camera
Including processing
56.3 GB
Light frames76 lights (19 per hour)
Capture rate992 MB per hour of lights
Processing copies47.6 GB

Decimal units: 1 GB = 1,000,000,000 bytes, as drive capacities are stated.

What it estimates

The calculator answers a practical question before a long session: how much disk space will this night take, and how much more will processing need? It counts three groups of files:

  1. Light frames: the number that fit in the hours you enter, given the sub-exposure length and the gap between subs (download time, dithering, settling).
  2. Calibration frames: darks, flats and bias or dark-flat frames at the same resolution and bit depth.
  3. Processing copies: the intermediate files a stacking program writes, typically one calibrated and one registered copy of every light frame, stored as 32-bit floating point.

How a FITS file’s size is set

Astronomy cameras normally save FITS files. The FITS Standard fixes the layout precisely:

  • every part of a file is made of whole 2880-byte blocks; the image data is padded up to the next block;
  • each header block holds 36 keyword records of 80 characters;
  • BITPIX = 16 means 16-bit integers, 2 bytes per pixel. Cameras that produce unsigned 16-bit values store them in the same 2 bytes, with the standard’s BZERO = 32768 convention, so the file is the same size;
  • BITPIX = −32 means 32-bit IEEE floating point, 4 bytes per pixel. The calculator assumes this format for processing copies; if your stacking program saves 16-bit intermediates, halve that part of the estimate.

frame size = header blocks × 2880 + ⌈ width × height × bytes per pixel ÷ 2880 ⌉ × 2880

The calculator assumes two header blocks (up to 72 keyword records). Capture programs that write more keywords add 2880 bytes for each additional 36, a difference of a few kilobytes per frame.

Worked example

A ZWO ASI2600MC Pro (6248 × 4176 pixels, 16-bit) taking 180-second subs with a 5-second gap, for 4 hours:

ItemCountSize
One light frame152.2 MB
Lights (19 per hour × 4 h)763.97 GB
Darks, flats, bias30 + 30 + 304.70 GB
From the camera166 files8.66 GB
Calibrated + registered copies, debayered to RGB, 32-bit2 × 7647.6 GB
Total56.3 GB

Two things stand out. Calibration frames can take as much space as the lights on a short night, because their count does not depend on how long you imaged. And processing dominates: each debayered 32-bit copy of a colour frame is 6248 × 4176 × 3 channels × 4 bytes = 313 MB, six times the original. A mono camera’s 32-bit copies are twice the original, so the same session in mono would need about 15.9 GB of processing copies instead of 47.6 GB.

Other sensors

SensorResolution16-bit FITS frame
IMX455 full frame (e.g. ASI6200MM Pro)9576 × 6388122.4 MB
IMX571 APS-C (e.g. ASI2600MC Pro)6248 × 417652.2 MB
IMX585 (e.g. ASI585MC)3840 × 216016.6 MB

Planetary imaging is a different pattern: short videos of thousands of frames, often in 8-bit. At 8 bits an IMX585 frame is 8.3 MB, so a capture of 6000 frames is about 50 GB before any frames are discarded. Use the calculator with your frame count and 8-bit setting to plan these runs. Capturing a smaller region of interest, where the capture program supports it, reduces the size in proportion to the pixels kept.

Assumptions

  • Sizes use decimal units (1 GB = 1,000,000,000 bytes), the way drive capacities are labelled.
  • Compressed formats (some programs offer compressed FITS or XISF) are not modelled.
  • DSLR raw files use their own compressed formats; the calculator’s FITS sizes do not apply to them.
  • The number of processing copies depends on your software and settings. Enter 0 to count capture files only.

For what the calibration frames are for and how many are typically taken, see the calibration frames guide; for how integration time relates to signal-to-noise, see the integration time guide.

Questions

How big is a FITS file from an astro camera?
Roughly width × height × 2 bytes for a 16-bit image, plus a header of at least one 2880-byte block, with the data rounded up to a whole number of 2880-byte blocks. A 26-megapixel IMX571 frame is about 52 MB; a 61-megapixel IMX455 frame about 122 MB.
Why do processed files take so much more space than the raw frames?
Calibrated and registered copies are often saved as 32-bit floating point, twice the size of 16-bit data. A colour camera's frames are also debayered into three channels, so each copy can be six times the size of the original.

Sources

  1. Definition of the Flexible Image Transport System (FITS), version 4.0 IAU FITS Working Group (hosted by NASA HEASARC). 2880-byte blocks (Sect. 3.1), 36 keyword records per header block, BITPIX values (Table 8), unsigned 16-bit via BZERO = 32768. Accessed October 12, 2026.
  2. The FITS Standard document NASA HEASARC FITS Support Office. Accessed October 12, 2026.