In Tarquesta [TQ] pronunciation an initial velar nasal ñ [ŋ] was often pronounced as a dental nasal n, especially among the Vanyar. This sound change was not reflected in spelling. Tolkien mentioned this phonetic development in both the Outline of Phonetic Development (OP1) from the 1940s and the Outline of Phonology (OP2) from the 1950s:

Note that nasal ñ when derived from ñg endured in classical PQ, and had a distinct alphabetic sign. But ñ [ŋ] in isolation was not a sound found later in Lindarin or Telerin, and the initial ñ of PQ was usually replaced by n in TQ pronunciation (and hence in the transcription of nouns such as Noldor in the histories). The Noldor, however, who possessed the sound [ŋ], and in certain circumstances used it initially, in their own language, often retained ñ (OP1: PE19/36).

The later history of initial ñ whether from old simple ñ or from ñg requires to be noted. It remained in Ñoldorin pronunciation, but became n- (by sound substitution) in the colloquial TQ of the Vanyar, though the learned among them continued to employ ñ when reading or reciting PQ. The precise nature of the change ñ > n is debatable, owing to the insufficiency of the records of actual colloquial TQ at the period immediately preceding the Exile. A probable view is that in Vanyarin TQ initial ñ > n, though a substitution (and due to distaste for free ñ), was an actual phonetic development independent of the spelling. The Vanyarin n- (for ñ) reappears among the Exiled Ñoldor, but must be regarded as parallel to the avoidance of rhotacism, or retention of V. [þ]. In the Numenorean use of Quenya initial ñ was usually pronounced as n - by a similar but independent sound-substitution. Ælfwine appears to have heard both pronunciations (OP2: PE19/76-77).

In OP1 this was a Lindarin [pre-Vanyarin] sound change. In the original draft of OP2, Tolkien reversed this to become a Ñoldorin-only sound change (PE19/77 note #44) as he did with other sound changes like how [θ] became [s] and [z] became [r]. Unlike those conceptual shifts, Tolkien reversed himself again, switching ñ- > n- back to being primarily a Vanyarin phonetic development as described above. These revisions were made in red ink, so it is hard to tell when Tolkien made them. In notes from the early 1950s or late 1960s Tolkien said ñ- > n- was Ñoldorin-only sound shift, but ñ was frequently restored after the Exile:

In this respect EQ represented the recognition of sound-changes which had begun among the Noldor before the Exile and had already caused Noldorin Quenya to diverge from the language of the Vanyar. The principal of these were: ... the change of initial [ŋ] to n ...

With respect to pronunciation a curious situation arose. Quenya as it entered Beleriand was, as spoken, a language that did not possess the sounds þ, z, or “free” ŋ (sc. not followed by k, g); but in Sindarin þ and ŋ were frequent and their distinction from s, n was essential. After their acquisition of Sindarin the Noldor were thus familiar with þ and ŋ and the more learned often reintroduced the ancient pronunciation of the Feanorian letters that represented these sounds (now pronounced s, n). But in more popular use among those who were only familiar with written Quenya in Exilic form this was not done. It was in the EQ form (without þ, ŋ) that Quenya reached other peoples such as the Numenoreans. (PE17/129).

The above is consistent with the original scenario described in OP2 (prior to the red-ink revisions): the initial sound changing ñ > n in the Ñoldorin dialect but being frequently restored after the Ñoldor encountered the Sindar. Tolkien also mentioned this sound change in The Lord of the Rings Appendix E:

NG represents ng in “finger”, except finally where it was sounded as in English “sing”. The latter sound also occurred initially in Quenya, but has been transcribed n (as in Noldo), according to the pronunciation of the Third Age (LotR/1114).

noldo (older ngoldo) one of the kindred of the Noldor, nwalme (older ngwalme) torment (LotR/1123).

Tolkien only rarely used þ and z in the representation of Quenya words, but outside the main narratives, ñ is quite common. It appeared in The Etymologies from the 1930s, where Tolkien frequently (but not universally) represented this initial sound as ñ-. The form Ñoldor appears almost as often as Noldor; for example Ñoldor is ubiquitous in OP2 and the Quendi and Eldar essay (WJ/360-417). The tengwa that originally represented this sound (g) was never lost or repurposed.

I think it is safe to say that while the pronunciation [n] was normal among men, the pronunciation [ŋ] was still common, especially among Elves, with [ŋ] > [n] being either a very late TQ change among the Ñoldor or the sound [ŋ] being restored after their encounter with the Sindar. Neo-Quenya authors likewise tend to use both n (after the spelling in The Lord of the Rings and The Silmarillion) or ñ (after the spelling in Tolkien’s more private writings). Eldamo mainly follows the first pattern, but indicates where the original spelling was ñ (g), just as it does with the original spelling of þ (3).

Conceptual Development: The sound change of initial [ŋ] > [n] also appeared in Early Qenya, as mentioned in the Qenya Phonology of the 1910s (PE12/15-16). Some examples:

• ᴱ✶ŋaistă- > ᴱQ. naista- “to oppress” (GL/37).
• ᴱ✶ŊARA > ᴱQ. nara “strife” (QL/64).

It seems this sound change was universal in the 1910s, without the survivals/restorations of the 1930s through 1960s.