Pretty sure SR was not out when I last played with tapes (early 1980's); at least, none of my decks had it.
No, it was not built in.
Roger, could you briefly describe SR? I know I could go look it up, but it'd be nice to get a capsule summary from the expert here in this thread, if you don't mind.
SR is like the other 2-way noise reduction systems that encode before the tape and decode after with the complementary process, thus restoring the signal and reducing the noise. What SR did was concentrate on achieving this with the least manipulation of the signal as possible. Noise is not audible all the time, so there's no need to "fix" it all the time. Yet a system like dbx (to take an extreme example) is dynamically adjusting the entire audio signal all the time. A-type NR divided the spectrum into 4 fixed bands to provide selectivity, and it gave 10 -15 dB NR. In order to increase the dynamic range improvement further, SR has more bands, but that means they can be more selective. Only when needed do they act.
One other interesting aspect of SR is the compression ratio. If you listen to the signal on the tape, you can hear significant differences between the various NR systems. So what? No one listens to that without decoding (Ok, except maybe B-type cassettes). But it is telling to listen to the compressed signal. If it has large changes in the signal, like when a transient occurs, that behavior has to track perfectly in the decoder in order to invert perfectly. But tape is not a perfect medium, and that affects how well the decoder tracks the encoder.
SR took the philosophy of ensuring the compressed signal was free of any obvious dynamic manipulations. It sounds compressed, but you do not hear rapid signal modulations. It's very smooth. By virtue of that, the decoder is not going to expose its operation.
There's a lot more going on besides just NR.
Here's the Dolby paper in case you want to dive in.