I figured out that Chinese syllables can be broken down into 4 segments:
Medial: i, u or yu which can function either as a consonant or a vowel.
Final: vowel (optionally turns into a diphthong or has got n/ng)
There are 11 initial consonant pairs (In one pair, I included a silent consonant in case if medial is used as a initial) Initials can appear independently only for unpalatalized sibilants and a silent consonant which might be used for er sound.
For medials, there are 4 of them if a silent one is counted. They can appear independently as initials, in the between of initial and final, and as a final vowel.
For finals, I counted them as ‘a’ based (a, ai, an, ang, au) and ‘e’ based (e, ei, en, (e)ng, ou).
There is also ‘o’ final, which is used only for ‘lo’ and hard ‘i’ final, so I am going to count them both as a silent final which is used also for medial finals.
If all initials, medials and finals are empty, it is going to be a er sound.
If the initial is an unpalatalized sibilant, the medial has to be empty and the final has to be empty too.
There are also 5 tones, but there is nothing to talk about them.
So all information I have to remember:
which one of 11 consonant a syllable does have?
is the consonant from the pair a type 1 or type 2?
which one from 4 medials does the syllable have?
has the syllable got ‘a’ based final, ‘e’ based final or empty final?
has the syllable got -, -i, -n, -ng or -u as a final consonant?
which one of 5 tones does the syllable have?
I generally got the idea from The Hanzi Movie method, but I wanted to create my own system. Hence I think the image of the initials in my memory palace can be based on the people or characters. The final consonant can be based on locations and the tones on the position in these locations. If there is an empty final vowel, it has got a special room. Whether is it ‘a’ or ‘e’ and type 1 or 2 (of the initial) can be determined by weather. It can be either sunny/windy (a1/e1) or rainy/snowy (a2/e2). The thing I don’t know is how should I manage medials. I thought about adding color to the image. For example red for i, blue for u, green for yu and normal for -. But this does not feel like a good idea for me, is there a better way put that info into a palace?
Since this system has to be for remembering character pronunciation, a person from the image needs to have got an item or items that can represent the radicals and the meaning. This isn’t that hard to do. The problem is connecting two or more characters to create a sentence. Is creating a long corridor with photos from palaces placed next to each other a good idea?