This is a big question. There are entire books that address it, but let's try a summary. Typically, you create an RF signal by first considering what kind of modulation you want and what frequency you will transmit on. These two factors will tell you what frequency stability you need, which often leads to the decision to have crystal control. Crystal oscillators can be designed to operate up to a few hundred MHz but when the frequency gets too high, the crystal becomes impossible to manufacture. There are two ways that we commonly get crystal stability at very high frequencies. The most popular method now is to use a phase locked loop synthesizer. The other way is to use frequency multipliers. Which one is the best approach depends on what kind of modulation you will have. If your frequency is relatively low, like below 100 MHz, then it is feasible to set your crystal oscillator frequency at the same as your transmitter output frequency, which eliminates the need for synthesizers or frequency multipliers.
So far, we've talked about how to generate a carrier frequency. That isn't much use unless you can modulate it with information. The techniques for generating modulation are very different depending on the modulation type. For analog FM, you can modulate your crystal oscillator or your frequency synthesizer directly. For phase modulation, you often use a mixer or upconverter. For multicarrier (like OFDM), you would probably use a DSP with A/D converter to generate the signal. So, it depends. What kind of modulation do you want?
In the real world, the transmit frequency choice is driven by government regulations. The government determines what ranges of transmit frequencies are to be used for specific services, so we have to consult the rules to find out what frequency our transmitter will be allowed to transmit on. This begs the question...why is your transmit frequency going to be 50 MHz or 500 MHz?