I hear chord progressions very easily in existing melodies. Part of it is applied knowledge of book stuff - chord scales from minor major and modal keys.
The point of this point is not to say that what you did is wrong or incorrect, and some amounts of dissonance in music is desirable - its a device to be freely used. Anyways......
As the melodies get more complicated - its harder and harder to give a free reharmonization as the melodies themselves spell out various chords.
Example - Pachelbel's Canon in D. The melody spells out strong chord tones straight off, and as each layer is added, the harmonization becomes more and more locked.
here only listen to the bass motion and the violin melody. Discard the pizzicato plus that arpeggiate the chords. It would be very difficult to reharmonize the song based on those existing parts.
Its the same thing with your piece in your first video. Assuming key of Db Major. This list is what is easily heard based on the melody alone.
Bar 1 - Db
Bar 2 - Gb
Bar 3 - Db
Bar 4 - Ab
Bar 5 - F min
Bar 6 - Bb min
Bar 7 - Ab
Bar 8 - Bb MIN
Bar 9 - Ab
Bar 10 - Db
Bar 11 - Gb
Bar 12- Ab
Bar 13 - Gb
Bar 14 -Ab
Bar 15 - Eb min
Bar 16 - Ab
The fact that you came up with a different harmonization was based on some arbitrary rules - like if I understand correctly - you took the first note in every bar to be a chord tone - which ignores that fact that in a lot of bars the melody employs a passing tone to get to a chord tone and that passing tone.
The strict musical definition of whether its an accented passing tone etc doesn't really matter - I am calling a non-chord tone a passing tone here.
Bar 5 is one good example - The melody plays an arpeggiated F minor triad but your harmonization Plays an Ab chord. Is it it wrong if you assume pan-diatonic / pan-tonal thinking? No of course not. The chords and melody come from the same scale. Its like playing Vsus over the V chord or ii over the V chord. It could work.
But does it sound plausible? Is it the smoothest and most obvious reharmonization? Would it fool a lay listener? Would it fool an edjucated listener?
How about bar 10? The melody clearly spells out a Db Major Chord (I) but the harmony stays on the Ab chord (V). That one is more of a stretch for me - you broke a steady harmonic rhythm where you were changing chords once every bar, and you played the "avoid" notes as strongly accented melody notes. The harmony is Ab (Ab, D, Eb), and the Db Melody note will either be heard as a Major 7th or b9 interval against the C (3rd of Ab) note.
That's a crunchy sound right there. Which is going to make most people's ears perk up. Is it right or is it wrong? Does it work?
I think what the other respondants said about doing things by ear is important - you keep moving through your choices until you reach something that fits. Even if you start with a theoretical basis for something, you can still qualify if it works by ear.
there is lots of music where the harmonies are rich and.......modern for lack of a better word, but I think while the methods might employ formulae or "chance" combinations, everything is always qualified by ear. If it doesn't sound smooth then something else is chosen. Some examples would be Michael Brecker's "Pilgrimage" https://youtu.be/JjghmN_-j-g?t=171
or a lot of Snarky Puppy albums.
In Pilgrimage - thats a lot of chords, and the melody is more sparse so that it can work together. In almost all cases the melody notes can work one chord forward and one backward - so there is some commonality, a thread of continuity although the harmony is usually moving in non-diatonic fashion.
Thinking out loud here