There are at least two common models for calculating the photoemission of accelerated electrons. The 'extended-charge-distribution' method uses the quantum probability current (multiplied by the electron charge) as a source current for Maxwell's equations. The 'point-like-emitter' method treats the electron like a point particle instead of like a diffuse body of charge. Our goal is to differentiate between these two viewpoints empirically. To do this, we consider a large electron wave packet in a high-intensity laser field, in which case the two viewpoints predict measurable photoemission rates that differ by orders of magnitude. Under the treatment of the 'extended-charge-distribution' model, the strength of the radiated field is significantly limited by interferences between different portions of the oscillating charge density. Alternatively, no suppression of photoemission occurs under the 'point-like-emitter' model because the electron is depicted as having no spatial extent. We designed an experiment to characterize the photoemission rates of electrons accelerated in a relativistic laser focus. Free electron wave packets are produced through ionization by an intense laser pulse at the center of a large vacuum chamber. These quantum wave packets can become comparable in size to the laser wavelength through natural spreading and interactions with the sharp ponderomotive gradients of the laser focus. Electron radiation emitted out the side of the focus is collected by one-to-one imaging into a 105-micron gold-jacketed fiber, which carries the light to a single photon detector located outside the chamber. The electron radiation is red-shifted due to mild relativistic acceleration, and we use this signature to spectrally filter the outgoing light to discriminate against background. In addition, the temporal resolution of the electronics allows distinction between light that travels directly from the focus into the collection system and laser light that may scatter from the chamber wall.