Understanding Calibration Frames in Astrophotography: A Beginner's Guide
When you get your start in astrophotography, the complexity can quickly become overwhelming. You’re often happy if you can capture an image at all, let alone a bunch of calibration frames. But calibration frames are an important aspect of astrophotography. And their purpose is to make your images of the sky a lot better by removing the unwanted parts of your images such as dust, heat noise, camera bias noise and amp glow, so that your final deep sky image is free of artifacts and imperfections.
How are calibration frames used?
Calibration frames are used to calibrate against each other, as well as your light frames. A series of 25-50 of each calibration type (Bias or Flat Dark, Dark, and Flat) are typically taken then stacked together to end up with a single master image for each calibration type. Each type of calibration frame serves a purpose, and I’ll outline them all below.
Light frames
Light frames are the images you take of deep sky objects. These are not calibration frames, but the frames you’re going to calibrate with calibration frames. Taking long exposures of deep sky objects will reveal many issues with your equipment. This includes issues introduced by the telescope, image train (reducer, spacers, filters, etc), and camera. Each type of calibration frame plays a role in removing these issues. It’s important when taking light frames to remember the gain/offset or ISO value you used, as well as the camera temperature if it has temperature control, and the exposure length. All these values apply in some way to your calibration frames.
Bias Frames: The Quick Fix for Sensor Noise
A bias frame is an image taken while the camera is covered, with a minimum exposure set. Typically, the bias frame is capturing noise generated by the camera. When electricity flows through the sensor to take an image, a certain amount of noise is generated from that flow of electricity. This is called the bias signal. All cameras produce some type of bias noise from that electrical charge, and the goal of bias calibration frames it to remove that low level noise. To take this type of calibration frame, you cover your camera so that no light can leak into the image, and set it for the shortest possible exposure time. I’ve generally settled on 0.05 seconds as my exposure for bias. It’s best to take between 25-50 of these frames.
Taken in the dark (with lens cap)
Same gain/offset or ISO as your lights
Same temperature as your lights
.05 second exposure
25-50 total frames captured
Dark Frames: Eliminating Thermal Noise
These frames are meant to correct out heat noise, in your images as well as amp glow from the sensor itself. As the camera sensor begins to heat up during a long exposure, hot pixels and heat glow from the sensor begin to appear in your images. A dark frame is a frame that is taken with the camera or telescope covered, and exposure and temperature set to the same as your light frames. For these to be effective, no stray light should enter the camera sensor. Dark frames are taken to match the light frames in exposure and temperature so that you can reverse the effects of heat introduced noise from your light frames. It’s best to take between 25-50 of these frames.
Taken in the dark (with lens cap)
Same gain/offset or ISO as your lights
Same temperature as your lights
Same exposure time as your lights
25-50 total frames captured
Flat Frames: Correcting Vignetting and Dust
Flat frames are a type of light frame, taken with an evenly illuminated light panel to correct for minor vignette, sensor mottling (if any), as well as dust and other stray particles that get into your imaging train (on reducers, the sensor window, or filters). This type of calibration frame tends to be the trickiest to capture properly. It’s recommended you use the same gain/offset or ISO as your lights, and expose to a median ADU value of 32,000, or half way on the histogram. What makes these tricky is you have to exactly replicate the placement of filters, and your camera relative to how your lights were taken. This is why it’s always recommended to shoot flats immediately after shooting your lights, and before you dismantle your equipment. Any bump to the image train mounted to your telescope could result in dislodging dust, or rotating the camera just enough that these frames will not correct out the issues you see in your lights. In fact, if something has indeed changed since taking your light frames, flat frames can often make your calibrated image worse.
Taken with an evenly illuminated flat panel (or against the sky at dawn or dusk).
Same gain/offset or ISO as your lights
Same temperature as your lights
Exposure set to median ADU value of 32,500, or the mid point of your camera histogram
For one shot color cameras, you only need 1 set of flat frames, for mono, or filter use (even on color cameras) you need 1 set of flat frames for each filter used with the camera.
25-50 total frames captured
Flat Dark Frames: A Special Case
Some cameras don’t respond reliably to short exposures. A few cameras that suffer from this issue are the cameras that use the IMX294 and IMX492 Sony sensors, but there may be others. On these cameras, bias frames will not properly calibrate against your flats. And this will result in flats that don’t properly correct your light frames. So, a flat dark frame is a substitute for bias frames on cameras that require them. These frames are taken with the camera covered like a bias frame, but instead of a short exposure, you need to take it for the same length of time that you exposed your flat frames. If your flat frame took 2.5 seconds to expose, then you need to take a flat dark at 2.5 seconds as well.
Only used if bias frames don’t properly calibrate your flats (you’ll know if your flats overcorrect your light frames)
Taken in the dark (with lens cap)
Same gain/offset or ISO as your lights
Same temperature as your lights
Same exposure length as your flats
25-50 total frames captured
Best Practices for Calibration Frames:
Bias and dark frames can be reused, after you process and stack them into master frames, for future calibration runs. I generally replace these with new frames every 6 months to a year as the characteristics of your camera don’t often change much over time.
I recommend you take all your calibration frames in the same application that you take your lights. Some programs will write different information into the header of the image files, and when processing sometimes PixInsight or Astro Pixel Processor doesn't know what to do with the frames, and throw an error, or it results in uncalibrated frames.
Always take your flat frames at the end of an imaging run. Dust can move throughout an imaging session. You also run the risk of bumping your camera when dismantling your setup to bring it indoors. So it’s better to take your flats before you move your telescope.
All calibration frames require the exact same gain/offset or ISO, and temperature as your light frames, or they will not work properly.
Here’s a tutorial I wrote on calibrating your images in Astro Pixel Processor.