It is shown that photon shot noise and radiation-pressure back-action noise are the sole forms of quantum noise in interferometric gravitational wave detectors that operate near or below the standard quantum limit, if one filters the interferometer output appropriately. No additional noise arises from the test masses' initial quantum state or from reduction of the test-mass state due to measurement of the interferometer output or from the uncertainty principle associated with the test-mass state. Two features of interferometers are central to these conclusions: ͑i͒ The interferometer output ͓the photon number flux N (t) entering the final photodetector͔ commutes with itself at different times in the Heisenberg picture, ͓N (t),N (tЈ)͔ϭ0 and thus can be regarded as classical. ͑ii͒ This number flux is linear to high accuracy in the test-mass initial position and momentum operators x o and p o , and those operators influence the measured photon flux N (t) in manners that can easily be removed by filtering. For example, in most interferometers x o and p o appear in N (t) only at the test masses' ϳ1 Hz pendular swinging frequency and their influence is removed when the output data are high-pass filtered to get rid of noise below ϳ10 Hz. The test-mass operators x o and p o contained in the unfiltered output N (t) make a nonzero contribution to the commutator ͓N (t),N (tЈ)͔. That contribution is precisely canceled by a nonzero commutation of the photon shot noise and radiation-pressure noise, which also are contained in N (t). This cancellation of commutators is responsible for the fact that it is possible to derive an interferometer's standard quantum limit from test-mass considerations, and independently from photon-noise considerations, and get identically the same result. These conclusions are all true for a far wider class of measurements than just gravitational-wave interferometers. To elucidate them, this paper presents a series of idealized thought experiments that are free from the complexities of real measuring systems. DOI: 10.1103/PhysRevD.67.082001 PACS number͑s͒: 04.80.Nn, 03.65.Ta, 42.50.Lc, 95.55.Ym
I. QUESTIONS TO BE ANALYZED AND SUMMARY OF ANSWERSIt has long been known that the Heisenberg uncertainty principle imposes a ''standard quantum limit'' ͑SQL͒ on high-precision measurements ͓1-3͔. This SQL can be circumvented by using ''quantum nondemolition'' ͑QND͒ techniques ͓2-9͔.For broad-band interferometric gravitational-wave detectors the SQL is a limiting ͑single-sided͒ spectral density
͑1.1͒for the gravitational-wave field h(t) ͓10,11͔. Here ប is Planck's constant divided by 2, m is the mass of each of the interferometer's four test masses, L is the interferometer's arm length, and f is frequency. This SQL firmly constrains the sensitivity of all conventional interferometers ͓interferometers with the same optical topology as the Laser Interferometric Gravitational Wave Observatory's ͑LIGO's͒ first-generation gravitational-wave detectors͔ ͓12,13͔. LIGO's second-generation interferometers ͑LIGO-II; ca....