I think you are right here about DE capturing somethings that FE cannot. I can't quite get the example of idioms out of my head but it seems terribly meaningful and practical as an example of where a DE approach is needed. With the translation of Scripture in particular there is an extra aura of mystery and obfuscation about the language from the distance at which we are removed from its time, language, and culture because they are dead languages (ancient Hebrew, Aramaic, and Greek - not their modern counterparts) which adds an extra dimension of difficulty to translate things. So we may be tempted to just chalk up the problem of translating idiomatic expressions to "well we just don't fully understand the language and culture at that time" because it is a dead language, however the same problem exists today between very modern languages. Since I happen to be learning German currently I can supply an interesting example or two.
A good number of idioms include uses of prepositions, which can wildly differ in their usage between languages and cannot be simply "carried over" even for cognate equivalents in another language. One example might be how in English we say "step by step" whereas in German that expression would be written "Schritt für Schritt". A literal (FE) rendering of that in English might be "step for step" which perhaps could mildly make sense in English but is grammatically/semantically wrong to say, and if we were to try to use the German equivalent for the English "by" in the German phrase it would look like "Schritt bei Schritt" which would be rendered back into English as meaning "Step with/at step" which doesn't make sense to either the German or English speaker.
Now for a more pointed example, consider the following interrogative sentence in English and the figure of speech/idiom that is in bold:
"What kind of car is that?"
In German you must phrase the question like this:
"Was für ein Auto ist das?"
So "What kind of" in English is "Was für" in German. Literally rendered into English with a strict FE method of translation that German sentence would be translated: "What for a car is that?". Try using that one the next time that you are at a store talking to people and see what kind of blank or confused stares you receive.
Their response might be something like:
"What did you just ask?"
Thinking to self: Hmm... maybe they speak broken English.
"Did you mean 'What is that car for'? Driving of course!"
The point is that you absolutely "must not" use a literal translation in some cases if you want it to make any sense when it comes to idioms that do not match between the source and target languages (if it renders it as nonsense). Now some idioms like the Bible's "lift a finger" (as in the Pharisees not lifting a finger to keep the law) may not render as nonsense if carried over literally, so it may be fine to use FE, but still one may have to elaborate on the text to explain what "lift a finger" actually means in the sentence. So my assertion is that sometimes it is "necessary" to use DE, at least in the case of idioms.
So as for a Bible translation example, Matthew 8:27 in English should be rendered:
"What kind of a man is this, that even the winds and the sea obey Him?" (NASB)
But in German should be rendered:
"Was ist das für ein Mann, daß ihm Wind und Meer gehorsam ist?" (Luther Bibel 1545)
"What kind of" and "Was für" ("What for") are not Formally Equivalent, but they are Dynamically Equivalent. And I argue that in this case Dynamically Equivalent wins the toss for what is needed. There is nothing to say that FE and DE can't be (and aren't) mixed in translations, and for as FE as the NASB is I have seen DE translations in it just like the NIV on occasion, so it also mixes FE and DE.
~Josh