We present a revised effective wavelength and photometric calibration for the Wide-field Infrared Survey Explorer W4 band, including tests of empirically motivated modifications to its pre-launch laboratory-measured relative system response curve. We derived these by comparing measured W4 photometry with photometry synthesised from spectra of galaxies and planetary nebulae. The difference between measured and synthesised photometry using the pre-launch laboratory-measured W4 relative system response can be as large as 0.3 mag for galaxies and 1 mag for planetary nebulae. We find the W4 effective wavelength should be revised upward by 3.3%, from 22.1 to 22.8 μm, and the W4 AB magnitude of Vega should be revised from mW4 = 6.59 to mW4 = 6.66. In an attempt to reproduce the observed W4 photometry, we tested three modifications to the pre-launch laboratory-measured W4 relative system response curve, all of which have an effective wavelength of 22.8 μm. Of the three relative system response curve models tested, a model that matches the laboratory-measured relative system response curve, but has the wavelengths increased by 3.3% (or ≃ 0.73 μm) achieves reasonable agreement between the measured and synthesised photometry.