Hess v. Indiana, 414 U.S. 105 reaffirms Brandenburg v. Ohio, and refines that ruling w.r.t. "imminent". The holding in that case does not directly announce a concrete rule to be applied henceforth, saying
Appellant's language did not fall within any of the "narrowly limited
classes of speech" that the States may punish without violating the
First and Fourteenth Amendments, and, since the evidence showed that
the words he used were not directed to any person or group and there
was no evidence that they were intended and likely to produce imminent
disorder, application of the statute to appellant violated his rights
of free speech.
In Gooding v. Wilson, SCOTUS ruled that
The constitutional guarantees of freedom of speech forbid the States
to punish the use of words or language not within 'narrowly limited
classes of speech.'
The Hess court concluded that under no prior standard could "We'll take the fucking street later" be prohibited speech. However, under the fighting words standard of Chaplinsky v. New Hampshire, Ng's speech could easily be deemed to be "fighting words", since the speech was clearly directed at the Youtuber. In Brandenburg, the court's holding was that
Freedoms of speech and press do not permit a State to forbid advocacy
of the use of force or of law violation except where such advocacy is
directed to inciting or producing imminent lawless action and is
likely to incite or produce such action.
This is not the same as saying that a law restricting speech can only be constitutional when the forbidden action is advocating immediate use of force or lawless action, that is, Chaplinsky was not overruled.
The circumstances of the present video contrast starkly with the circumstances of Hess – there isn't even a round-about hint of urging some lawless action, the communication was directed to the Youtuber and there is nothing resembling a crowd to "incite". To the extent that the "fighting words" doctrine is still viable, Ng's conduct might be claimed to fall within the realm of "fighting words", which could at least maybe make the arrest legal.
The most recent ruling on the fighting words question is R.A.V. v. City of St. Paul, 505 U.S. 377, weakly reaffirms Chaplinsky:
A few limited categories of speech, such as obscenity, defamation, and
fighting words, may be regulated because of their constitutionally
proscribable content
but the ordinance restricting kinds of hurtful speech was found unconsitutional as substantially overbroad, for "display of a symbol which one knows or has reason to know 'arouses anger, alarm or resentment in others on the basis of race, color, creed, religion or gender'". No ruling has held that a direct person insult is constitutionally-protected speech. On the face of it, his conduct could violate the Arizona law against disorderly conduct, because it is directing "abusive or offensive language or gestures to any person present", but it is also required that it be "in a manner likely to provoke immediate physical retaliation by such person". The Youtuber's calm conduct clearly does not support the "immediate physical retaliation" element.