What is the test for intervention by a first court of appeal in the findings of fact of a trial judge?

Similarly, Lamer J. in Beaudoin-Daigneault v. Richard, 1984 CanLII 15 (SCC), [1984] 1 S.C.R. 2 at 8-9, 37 R.F.L. (2d) 225, 51 N.R. 288, states: So far as the rules for intervention by a first court of appeal in the findings of fact of a trial judge are concerned, in my opinion, whatever the uncertainty of the rule regarding some of our decisions in recent years … the rule is clear with regard to findings based on the credibility of witnesses: an appellate court should not intervene unless it is certain that its difference of opinion with the trial judge is the result of an error by the latter. As he had the benefit of seeing and hearing the witnesses, such certainty will only be possible if the appellate court can identify the reason for this difference of opinion, in order to be certain that it results from an error and not from his privileged position as the trier of fact. If the appellate court cannot thus identify the critical error it must refrain from intervening, unless of course the finding of fact cannot be attributed to this advantage enjoyed by the trial judge, because nothing could have justified the judge's conclusion whatever he saw or heard; this latter category will be identified by the unreasonableness of the trial judge's finding … (citations omitted) (emphasis mine)