Default python compression level (11) causes more delay than it's worth on larger files

Using a 7314288 byte (~7MB) json file as an example:

The image is kinda blown out, so here's the raw data:

Quality | Compression Time | Size
-- | -- | --
0 | 0.011061668 | 287292
1 | 0.009209156 | 213457
2 | 0.018366814 | 184692
3 | 0.01734972 | 184569
4 | 0.023967743 | 177295
5 | 0.037658453 | 151161
6 | 0.042690992 | 150042
7 | 0.054255962 | 148143
8 | 0.072593212 | 147667
9 | 0.108716726 | 146448
10 | 1.435314894 | 130876
11 | 20.70757484 | 122797

I understand the desire to have the library put its highest compression quality out there first, but wouldn't it create an overall better impression on users if a faster default was used? Quality 8 provides nearly as good a result as 11, but it's almost 300 times faster.

Start by locating the Python binding's default compression quality and the code or tests that define it. Compare the reported quality 8 and 11 timings and output sizes, then determine what benchmark or test should establish the new default; done means the default behavior and its documentation or tests consistently reflect the chosen quality.

Written by the indexing model from the issue text.