[jkn]

I don't think this is right. The energy of a photon is h multiplied by the frequency, but the frequency itself can be anything. Even if the process creating the photon is inherently quantized (e.g. bound electron with quantized energy levels), we will measure slightly different frequencies (and so different energies), due for example to the Doppler effect[1]. Still with the Doppler effect, you can give a photon an arbitrary frequency (energy) by changing the velocity of the reference frame where you make the measurement (that is, until someone shows that the velocity of the reference frame is quantized... not something that is presently known).

[1] https://en.wikipedia.org/wiki/Doppler_broadening

[URSpider94]

Yes, that's correct. This does not mean that the energy levels of light (more generally, EM radiation) are quantized -- photons can have any energy level. What it means is that, for light of a given frequency, you can deposit energy only in units of h(nu).

The quantization of light shows up in how it interacts with particles -- even unbound particles like free electrons, which also do not have quantized energy levels. Specifically, if light were NOT quantized, you could get the same effect with more intense light that you get with more energetic light. Instead, experiments show again and again that longer-wavelength light at high intensity gives a totally different effect from short-wavelength light at low intensity. Postulating that light consists of particles (photons) with E = h(nu) explains this difference.