Verified answer

ONE photon is always the product of ONE atom making energy an energy transition. The atom loses energy that transforms into the photon. From Planck's famous equation we get the frequency of the photon:

E=h*f where f is the frequency.

So if we know the type of atom and the types of energy transitions it has we can deduce how many photons have been emitted by an object by simply dividing the total energy by the energy of a single photon. However, rarely do scientist want to produce exactly "one" photon.. Scientists are interested in interactions between a bunch of photons and a bunch of matter in order that the effects be measurable. The fewest photons that I've heard were deliberately produced was in an experiment to determine whether or not photons exist at all. The experiment proved that the threshold for the human eye to detect a flash of light is about 48-50 photons. They made small variations to a light source and found that in that region the human eye either saw the light or not, so they said that photons were indeed real (because sometimes there were only 47 photons, sometimes 51 - it's a statistical process like water droplets forming on a faucet). These experiments were conducted by Brumberg and Vavilov and I'll give you two websites that could be interesting to you.

The best way to see one hf=energy (I won't call it a photon because I disprove it) is with gamma-rays from spontaneous emission. There are many atoms doing that, but the detectors in that part of the spectrum have what they call high energy resolution. It is confusing; it is really frequency resolution. What you need to do is compare to chance. If you look for overlapping detection pulses they will overlap by the chance equation: Rc= 2R^2 T, where T is the time resolution that you use for overlapping pulses. Do a pulse height spectrum of an isotope like Cadmium-109 and look at the sum peak. The sum peak is in the spectrum where the pulses are twice as high as the characteristic pulse height of Cd-109. So you get to see one pulse at a time on the oscilloscope, and you get to see the chance overlap rate from the equation, and you get to see the experimental rate that the pulses go twice high. What you find is that the sum-peak happens at about twice chance. That violates the photon model and quantum mechanics. Therefore the photon model fails. This scheme lets you see everything. It resolves the wave particle paradox and disproves the problematic photon model. The E-hf is quantized at emission, but thereafter spreads classically. You generate and sense single hf, see chance defied, and understand the nature of light by looking at the pulse height gamma-ray spectrum of Cd-109. See unquantum.net for details. There are many more experiments there to clarify my case. Parametric down conversion is often cited as a way to generate single hf, but it does not; it generates thousands and you will not get to do the analysis that reveals what light is, that I described. Another way often cited is quantum dots. The dots will emit a single hf, but you will again not be able to perform the pulse height analysis described above, to see through the illusion of the photon model.. . .

You can use a filter to dim the light down to as few photons as you want even one.

You can tell how many photons you have by taking an energy spectrum with a spectrometer.

Photons are discrete packets of energy. E=hf --> so if you know the frequency f and measure the energy E you can determine how many photons you have.

PS the only way to generate a photon is to accelerate charge. All photons are generated this way.

All I've ever generated was methane. Sorry