Photometric Filter Systems

What a photometric system actually is

Stars emit light across the whole electromagnetic spectrum, but you rarely need the full spectrum to do useful science. A photometric system reduces that flood of wavelengths to a handful of numbers by passing the starlight through a set of filters, each one a piece of colored glass (or a thin-film interference coating) that transmits a known band of wavelengths and blocks the rest. The transmission curve of a filter — how much light it lets through as a function of wavelength — is its passband or bandpass. Measure how many photons make it through each filter, and you have the star's brightness in that band: a magnitude.

The crucial idea is standardization. If everyone agrees on the same filter shapes and the same zero point, then a V magnitude measured in Chile in 1965 means the same thing as a V magnitude measured from a backyard CCD today. That comparability is what turns photometry from a one-off measurement into a science. A photometric system is therefore not just glass — it is a contract: a defined set of bandpasses plus a defined zero point that anchors the magnitude scale.

UBVRI: the Johnson-Cousins system

The most influential optical system is UBVRI. Harold Johnson and William Morgan defined the U, B, and V bands in the early 1950s to match the photomultiplier tubes and 1P21 detectors of the era; Alan Cousins later standardized the redder R and I bands. The letters are mnemonic: Ultraviolet, Blue, Visual (centered in the green-yellow, near where the dark-adapted eye peaks), Red, and near-Infrared.

These are broadband filters — each is tens to over a hundred nanometers wide. Width is a deliberate trade: a wide passband collects more photons, so you can reach fainter stars in a given exposure, at the cost of throwing away fine spectral detail. The U band is special: its blue edge is partly defined by the atmosphere and the detector rather than the filter alone, which is one reason U-band photometry is notoriously hard to calibrate.

Band	Effective wavelength	Approx. width (FWHM)	Region
U	~365 nm	~66 nm	Ultraviolet
B	~445 nm	~94 nm	Blue
V	~551 nm	~88 nm	Visual (green-yellow)
R	~658 nm	~138 nm	Red
I	~806 nm	~149 nm	Near-infrared

Color index: turning two numbers into a temperature

A single magnitude tells you how bright a star looks; the real power comes from color. Subtract the magnitude in one band from another and you get a color index — most famously B-V = m_B − m_V. Because the magnitude scale is logarithmic and inverted (smaller numbers are brighter), the sign of B-V tells you which way the star's energy leans. A hot star pours out blue light, so it is bright in B, has a small m_B, and a small or negative B-V. A cool star peaks in the red, is faint in B, and has a large positive B-V.

The color index is, in effect, a cheap thermometer. It encodes roughly where the star's blackbody peak sits without ever taking a dispersed spectrum. That is why color-magnitude diagrams — the observational cousins of the Hertzsprung-Russell diagram — plot V against B-V: the horizontal axis is temperature in disguise.

Star	Spectral type	Approx. B-V	Surface temp.
Vega	A0V	~0.00	~9,600 K
Sirius A	A1V	~0.01	~9,900 K
The Sun	G2V	~0.65	~5,770 K
Arcturus	K1.5III	~1.23	~4,300 K
Betelgeuse	M1-2Ia	~1.85	~3,600 K

Vega's row hides a convention: in the Vega magnitude system, the star Vega is defined to have magnitude 0.0 in every band, which makes all of its color indices zero by construction. The whole UBVRI scale hangs off this single anchor (in practice a network of standard stars matched to it). Reddening from interstellar dust complicates the picture — dust makes stars look redder than they are — so astronomers distinguish the observed color from the intrinsic color via a color excess like E(B-V).

Magnitudes, zero points, and the AB system

Counting photons gives you an instrumental magnitude, but to put it on a standard scale you need a zero point. Two conventions dominate. The Vega system, used by UBVRI, sets Vega to zero. The AB system, used by SDSS and many modern surveys, instead pins magnitude directly to physical flux density per unit frequency:

m_AB = −2.5 log₁₀(f_ν) − 48.60

where f_ν is in erg s⁻¹ cm⁻² Hz⁻¹. The AB system has the tidy property that an object with a flat f_ν spectrum has the same magnitude in every band. Converting an AB magnitude to a Vega magnitude (or vice versa) requires per-band offsets — small in the optical (a few hundredths to a few tenths of a magnitude) but large and important in the infrared, where they can exceed a magnitude. Forgetting which system a catalog uses is a classic source of error.

From UBVRI to ugriz, Gaia, and LSST

Johnson-Cousins was tuned for photomultipliers; modern silicon CCDs and space detectors have different sensitivity, so new systems followed. The SDSS ugriz system (u ~354 nm, g ~477 nm, r ~623 nm, i ~762 nm, z ~913 nm) was engineered with non-overlapping, near-rectangular passbands placed to dodge bright atmospheric emission lines and to tile cleanly across the CCD response. Gaia uses just three very broad bands — G (a wide ~330–1050 nm white-light band), BP (blue photometer), and RP (red photometer) — to squeeze maximum signal out of a space telescope surveying nearly two billion stars. The Vera C. Rubin Observatory's LSST adopts ugrizy, adding a y band out near 1000 nm to push into the near-infrared from the ground.

System	Bands	Zero point	Typical use
Johnson-Cousins	U B V R I	Vega	Classic stellar photometry, variable stars
SDSS	u g r i z	AB	Large-area CCD imaging surveys
Gaia	G BP RP	Vega (Gaia DR)	All-sky astrometric + photometric survey
2MASS (near-IR)	J H Ks	Vega	Infrared all-sky survey
LSST	u g r i z y	AB	Deep, repeated wide-field imaging

The proliferation of systems is exactly why cross-calibration matters: the same star has a different numerical color in ugriz than in UBVRI, and survey teams publish transformation equations (often color-dependent polynomials) to convert between them. There is also a distinction between natural and standard systems — your telescope plus atmosphere defines a slightly different effective passband than the published standard, and reducing your data onto the standard system is part of the calibration craft.

Why photometric systems matter

• Depth over detail. Broadband photometry reaches stars hundreds of times fainter than spectroscopy in the same exposure — essential for surveying billions of sources.
• Temperature for free. A single color index estimates surface temperature without dispersing the light.
• Photometric redshifts. Sampling a galaxy's spectrum through several bands lets you estimate its redshift from colors alone, the backbone of cosmological surveys.
• Comparability. Standardized passbands and zero points let measurements made decades and continents apart be combined.
• Variability. Repeated photometry in fixed bands is how we catch eclipsing binaries, Cepheids, transiting exoplanets, and supernovae.

Common misconceptions

• Filters are sharp. Real passbands have sloped wings and transmission well below 100%; the "center" is an effective wavelength weighted by the source spectrum.
• The V filter is green. V peaks in green-yellow but is broad; it is called "visual" because it roughly matches the eye, not because it is monochromatic.
• All magnitudes share one zero point. Vega and AB systems differ band by band; mixing them silently corrupts colors.
• Color is intrinsic. Interstellar dust reddens starlight, so observed color must be de-reddened using E(B-V) before it reports the true temperature.
• Photometry replaces spectroscopy. It complements it — photometry gives breadth and depth; spectroscopy gives the detail photometry averages away.