Radio: From crystals to transistors to digital signal processing to software defined radio and beyond

I've recently been reading about software defined radio and am very tempted to buy either the Pegasus or Q1SR.

Here are some musings. Note: I'm not an electronics expert!

Crystal radio
The first generation of radio. You needed a crystal (or diode). Lacked any amplification and thus you needed a good antenna. Had terrible selectivity (rejection of adjacent stations).

Second generation. Tubes came along and let you amplify signals allowing for greater sensitivity. Probably some time around then people figured out how to make half-decent filters. Tubes were bulky and energy intensive. Some of the good tube radios weighted 50-100 pounds.

Third generation. Transistors and other smaller components allowed for portable (or smaller desktop) radios at affordable prices. Radio may have reached its peak during this period, before tv was so popular.

Digital Frequency
The move from analog to digital frequency readout for radio starting around 1980 was a huge change that made radio easier to use. This is more true for shortwave than am/fm, as shortwave has many more channels. You also saw the development of surface-mount components and integrated circuits which made for very small and powerful radios.

Digital Signal Processing
Only just now are we seeing digital signal processing be used for mainstream radios. The Sony XDRF1HD (the best fm tuner that you can buy and it is only $80) demonstrates the power of DSP. Cheap chinese shortwave radios are just beginning to use DSP (like the Tecsun PL-310 which has bandwidths of 1, 2, 3, 4, and 6khz - unlike the traditional cheap portable which only had 1 bandwidth). DSP has radio signals converted into digital (0s and 1s) and then processed the heck out of them (whereas analog radios process the signal using analog components).

Manufactures are going to start building powerful computer chips to run radios, and in 10-20 years they might be mass producing radios for $50-$100 that are as good as anything costing $1000 today and are a tenth of the weight (one pound or less).

Software defined radios are also part of the same trend (though may never be mass produced if they require a computer to work).

The Future - Digital Radio
The next step is the replacement of analog radio signals with digital ones. We see this happening on TV, FM, AM and shortwave (though shortwave adoption is incredibly slow).

Beyond Digital Radio
You might have radios that record signals for a day or a week and then you can choose whether to listen to it live, or from a previous broadcast. These radios will need to have a ton of memory (100GB+) which could happen in 10 years if SD memory cards keep improving.

Ultimately you could have two-way radio, where your radio could request programming. But I'm not sure if that is going to happen as we're eventually going to have trouble finding radio spectrum (which will be crowded by wireless internet). And I'm not sure if two-way radio is really radio. It's more like wireless internet.