To achieve spot-illumination, a light source is focused onto a small aperture (pinhole) that is then focused into the sample. When the aperture is small enough, the illumination spot is limited only by diffraction and not by the geometrical parameters of the light source and the aperture. Ordinary light sources are extended, and it is not possible to focus them to a diffraction limited spot. Therefore this arrangement is necessary (for traditional light sources), although the transparency is very low.
Lasers as light sources have a very high degree of collimation (the light is "very parallel" in good lasers). Therefore, the laser light can be focused by a single lens to a diffraction limited spot without applying a pinhole. Most confocal microscopes therefore have no illumination pinhole. The quality of the spot depends on the beam-quality of the laser. If the quality is insufficient, a pinhole can be inserted. Lasers are commonly coupled by optical fibers to confocal microscopes. These fibers themselves also act as pinhole.
It is the focusability and the high energy density of lasers that makes them the ideal light sources for confocal microscopes. Coherence of the laser light is not a required feature for confocal performance. It is a challenge for optical designers instead, as it causes spurious interference patterns, that need careful design strategies.
Also, the fact that conventional lasers emit only a single color (laser-"line"), is not per se beneficial, but generates need for complicated multi-laser arrangements when multi-fluorescence imaging and measurements are required. The white light laser source solves the color-question in particular very elegantly.
Most detectors have a comparably large sensitive area (PMTs typically a few square-centimeters). Confocal optics requires spot-shaped sensing. Therefore, the spot-detection has to be performed by inserting a small aperture (pinhole) into the beam. The light from the sample is focused onto this pinhole and the transmitted light is collected and recorded.
A detection pinhole is mandatory because the diffraction pattern depends on NA and wavelength. It is therefore necessary to adapt the pinhole size when these parameters are changed.
Modern true confocal scanning microscopes automatically change the pinhole-diameter appropriately, when the objective lens is changed (which usually is accompanied with a change of the numerical aperture NA). The pinhole usually is therefore designed as a bi- or multilamellar iris.
What actually is an appropriate size of the pinhole depends not only on wavelength and numerical aperture, but also on the internal magnification of the optical elements in the microscope.
A direct comparison of the pinhole-diameters in differently designed microscopes is consequently not just inadvisable, but essentially incorrect. If the pinhole diameter is not set to optimal value, the system will not perform good optical sectioning (i.e. transmit defocus blur) or the intensity is cut off unnecessarily without gaining further optical section quality (causes unnecessarily noisy images).
The confocal beam path in a true confocal scanning system is just the combination of spot-illumination and spot-detection. This combination acts as an optical knife. Only photons that stem from the focal plane are able to pass to the sensor. Consequently, all photons from elsewhere are filtered out. This "spatial filter" accounts for optical sectioning.
As at a given time, only a single spot is imaged "confocally", a scanning device is required that moves the spot in a raster pattern over the object field. Usually, optical mirrors mounted on scan-motors are used to perform the scanning procedure. The bottleneck is the time that is required to scan a full frame of (typically) 1,024 lines. Improvements have been achieved by introduction of fast scanners (resonant mode scanners) for scanning 8,000 lines and more per second.
Good optical sectioning is only achieved in incident-light microscopy. This is one of the reasons, why fluorescence microscopy was booming in the last 20 years (other reasons are the invention of immune staining, DNA-hybridization, fluorescent biosensors, quantum-dots and fluorescent proteins).