Ultimately this raises a question of balancing two fundamental rights: The right of the accused to a fair trial within due process of law & the right to freedom of expression (or in the case of press publications as in most of the cases cited by this answer, the very related freedom of the press). Both the US & Canada safeguard these rights in their constitutions (Amendments I, V, XIV & Charter ss. 2(b), 11(d) respectively) so any differences would be in their approaches to balancing these rights.
United States
I've found Some Aspects of the Law of Contempt of Court in Canada, England, and the United States (Jacob S. Ziegel, 1960) to be highly informative on the historical development of contempt of court and provided a good baseline understanding. In reference to Bridges v. California 314 U.S. 252 (1941), it states
Although the Court recognized that "free speech and fair trials are two of the most cherished policies of our civilization" and that "it would be a trying task to choose between them", it held nevertheless that an abridgment on the freedom of the press was only permitted by the Constitution when there was "a clear and present danger" that the offending publication would actually interfere with a fair and impartial trial. The majority opinion rejected the "reasonable tendency" test which had been applied by the lower courts, and still is the yardstick by which such publications are measured in England and Canada.
I'd also like to quote the following paragraph from Bridges as it raises two points I'd like to comment on:
What finally emerges from the "clear and present danger" cases is a working principle that the substantive evil must be extremely serious, and the degree of imminence extremely high, before utterances can be punished. Those cases do not purport to mark the furthermost constitutional boundaries of protected expression, nor do we here. They do no more than recognize a minimum compulsion of the Bill of Rights. For the First Amendment does not speak equivocally. It prohibits any law "abridging the freedom of speech, or of the press." It must be taken as a command of the broadest scope that explicit language, read in the context of a liberty-loving society, will allow.
The first point is that due to its broad language, the court gives a certain priority to the First Amendment.
The second point involves the age of my citations: the "clear and present danger" test is from Schenck v. United States, largely considered to be overruled by the more stringent Brandenburg v. Ohio 395 U.S. 444 (1969) "immediate lawless action" test.
This does raise the question of whether Bridges can then be considered good law. I unfortunately can't find any good references discussing Bridges in a post-Brandenburg context, however, I would argue it is still valid. While the Schenck test has been overruled, the Bridges citation I gave in fact foreshadows the Brandenburg test by using the language of "imminence" and stating the Schenck test was not necessarily the limit of the First Amendment, but just a minimum bar. I would then argue that Bridges can quite easily be considered good law by simply substituting the Schenck test with the Brandenburg one without really affecting the court's ratio decidendi.
Canada
Ziegel's article is even more out-of-date when it comes to Canada as it is pre-Charter. For a modern Canadian overview of out-of-court contempt of court I've found A Comment on "No Comment": The Sub Judice Rule and the Accountability of Public Officials in the 21st Century (Lorne Sossin & Valerie Crystal, 2013). The article cites Dagenais v. Canadian Broadcasting Corp. [1994] 3 SCR 835 as directly tackling the balance of these rights. In it the court acknowledged the pre-Charter common law of giving deference to the accused's right to a fair trial, but noting that the Charter now requires a balancing of these rights
The traditional common law rule governing publication bans -- that there be a real and substantial risk of interference with the right to a fair trial -- emphasized the right to a fair trial over the free expression interests of those affected by the ban and, in the context of post-Charter Canadian society, does not provide sufficient protection for freedom of expression. When two protected rights come into conflict, Charter principles require a balance to be achieved that fully respects the importance of both rights. A hierarchical approach to rights must be avoided, both when interpreting the Charter and when developing the common law. The common law rule governing publication bans must thus be reformulated in a manner that reflects the principles of the Charter and, in particular, the equal status given by the Charter to ss. 2(b) and 11(d).
Ultimately, the court concluded that a balance of these rights was the following test:
A publication ban should only be ordered when:
(a) Such a ban is necessary in order to prevent a real and substantial risk to the fairness of the trial, because reasonably available alternative measures will not prevent the risk; and
(b) The salutary effects of the publication ban outweigh the deleterious effects to the free expression of those affected by the ban.
While Dagenais more clearly explains the balancing of these two rights, the article also cites R. v. Vermette [1988] 1 SCR 985 which is a much more similar situation to Maxine Waters commenting on an ongoing trial. The article summarizes as follows:
While the trial was ongoing, the premier of Quebec, René Lévesque, denounced one of the defence witnesses in the National Assembly, despite warnings by the Speaker that such comments would be prejudicial to the accused.
The trial decision even contains the following comment from the judge:
When speaking in the National Assembly, however, the Premier enjoyed the privilege of Parliamentary immunity. Had it not been for this immunity, he could
have been cited for contempt of court.
Comparison
At a constitutional level, it can be seen that the balancing of these two rights are not the same in the US and Canada. The US gives much greater weight to freedom of speech, while Canada allows restriction of speech if doing so more greatly benefits another Charter right and there is no reasonable alternative.
On a practical level, there are also differences in how jury trials are run, maybe even as a direct result of the constitutional differences. In Dagenais, the court quotes the following:
Generally speaking, however, the approach taken in the United States seems to be to allow for the widest possible latitude in media reporting of events transpiring prior to and during the course of the trial of an accused person. This is counterbalanced, in the interests of ensuring an impartial and unbiased jury, in a number of ways including, during the jury selection process, by an often searching examination into the attitudes, biases and even the personal and financial affairs of potential jurors and, after the jury selection process has been completed, by the sequestration of the members of the jury while the trial is in progress to reduce the risk of their exposure to the media and other publicity generated by it.
In Canada, by contrast, the process of jury selection is neither as prolonged nor as exhaustive as a general rule; indeed the kind of questioning and probing into the affairs of potential jurors that is sometimes seen in the United States would be unlikely to be permitted under our system. Moreover, in Canada the sequestration of jurors throughout a trial occurs only exceptionally. The strong bias of our system is to prevent the dissemination before the conclusion of the trial of media publicity that might be prejudicial to the accused's fair trial.
The US & Canada simply follow different constitutional approaches here and have slightly different priorities when it comes to balancing rights.
