The second idea is to add non-intrusive phonetic subtleties to a
word's pronunciation to reflect the word's spelling, e.g. introducing
subtle phonetic differences into the pronunciations of "week" and
"weak" to tell apart their spellings. This idea can be combined with
the first idea so that "ee" and "ea" will have additional visual marks
to indicate two slightly different sounds both of which are derived
from the sound [i:]. How do we devise these slightly different new
sounds? So far the best way I have thought of is "slightly different
tones". It should be noted that such tonal variations should not be as
strong as those used in the Chinese language.
For the second idea, acoustics has a great space for adding extra
information (spelling) to a word's normal pronunciation. Physical
devices can perform very complex sounds, and human-pronouncible sound
space is also well studied.
Human languages contain a very rich set of phonemes which can be
borrowed as close variants to English phonemes so that additional
information can be encoded in an English word's pronunciation.
However, there is another approach that doesn't even require any
change to the original pronunciation: the hearer will receive the same
original pronunciation (hereinafter "received pronunciation"), but the
speaker (typically a non-native English learner) deliberately makes a
particular mouth posture as if he is going to make a particular
variant of an original phoneme of a word but actually produces the
same phoneme. Such a particular moth posture encodes a particular
spelling of the phoneme into the speaker's voice memory or motor
memory.
Actually, we can slightly modify a word's literal form in order to
remember its spelling better, such as raising/lowering a character or
character combination to make it more different from other characters/
character combinations that may have the same pronunciation.
Raising/lowering a letter, adding a special mark above/below a letter,
etc.
Or letter combinations. Some Unicode marks can connect two letters.