There is no uniform "j" sound. So saying "the J sound is not new" is wrong to begin with. The letter preceded the sound, actually. It was used in conjunction with vowel placement in handwritten texts to show that there was an "i" before another vowel because otherwise people might not see the "i" (especially in scripts where all the vowels were very small to save space). It was written in languages, like Latin and Italian, that have no distinct sound corresponding to "j" and which have no "j sound" as you call it. However, in French, the places where the "j" showed up often, due to the way French slurs most everything, elided the sounds in such a way that the sound that we now hear in French from "j" occurred naturally. This sound change happened distinct from the letter, but it happened where the letter showed up because the letter was so written when i preceded a vowel (which was also when the elision took place).

This change crossed the Channel and ended up in English. However, some other languages which still retain the use of "j" as separate from "i" nevertheless pronounce it as "y", its original value.

In other words, the symbol was used before the sound developed, and the sound developed for independent but related reasons.

The closer analogy to what you're arguing would be the distinction between u and v, where scribes were trying to show that the old Latin Vv had, by the late Imperial Period, ceased being pronounced as /u/ or, in certain contexts, /w/, but rather as /u/ and /v/. Even then, however, the two forms of the same letter (u and v) were used in ways different to our modern distinction - v was only written at the beginning of a word (much like many js), while u was written in any other position.

However, these were writing conventions that then ended up getting intertwined with shifts in language. The "ng" was just some guy's unilateral attempt at creating an unnecessary new letter. That hasn't worked at any point in history that I'm aware of.