Fix: use biggest subbuffer size for snapshot max-size

Instead of using the output max size divided by the total amount of
streams in the session, we find the largest subbuffer in the session's
channels. Using that value, we multiply it by the total amount of
streams which gives us the lower limit of the snapshot size. That is
enough to make sure that we can take the snapshot or not.

Once done, the max stream size possible used for the snapshot record is
the largest subbuffer size in the session. This is to make sure that
every channel can extract the same amount of data which ensure fairness
between each channel in the session.

Fixes #783

Acked-by: Julien Desfossez <jdesfossez@efficios.com>
Signed-off-by: David Goulet <dgoulet@efficios.com>