Potentially you could use Unicode encoded in UTF-32, but since there are many fewer Unicode code points than 2^32, there are a lot of valid 32-bit patterns that

Potentially you could use Unicode encoded in UTF-32, but since there are many fewer Unicode code points than 2^32, there are a lot of valid 32-bit patterns that don't map to a valid UTF-32 encoding of a Unicode code point. As a result, most Bitcoin keys would contain invalid characters, so that scheme doesn't work for all keys, but it does get you down to 8 characters for the keys it does work for. There are more than 2^16 Unicode code points, so you can definitely use UTF-16, which will get you down to 16 characters. One thing to note is that you're not actually getting any compression here: each of these character strings is the result of the same 256-bit pattern. To find out more about how much actual compression you could get of the data, look up Kolmogorov complexity.[0] For most Bitcoin keys/random 256-bit patterns, they aren't going to be significantly compressible, if at all. [0] https://en.wikipedia.org/wiki/Kolmogorov_complexity https://en.wikipedia.org/wiki/Kolmogorov_complexity