How do you handle discernible English walla in crowd BGs (library or otherwise) in an M&E for dramatic television? I've heard some take the momentarily understandable bits and just reverse it, others pull the entire effect leaving holes that don't match the full mix. Is there a preferred, "standard" way of tackling this?
Reversing is an easy, tried-and-true method that usually yields acceptable results. Or you could copy some non-discernable english from another part of the clip and use it as fill. Both approaches are standard.
I wouldn't advise leaving holes in your track; they might sound like dropouts and flag a QC kickback.
Keep in mind that the composite BG track is its own entity and has to be evaluated as such; no one outside of the dub stage will be soloing the 'crowd' pre dub and critiquing it. Even if you have some english dialog in the BGs some other elements may mask it, ie. traffic, birds, movement, etc. Play everything together; if you still hear the english, then fill it in.
I wish there was a 'standard' way of dealing with this. It really is a case by case basis. Sometimes reversing the section works as long as it is short and doesn't sound too weird. Other times you can usually find a section of the track that you can copy and paste over the discernible dialogue. What we usually try to do is have our BG editor try to cut in 'non-discernible' backgrounds (crowds, sidewalk, office..etc) that carries the sense of people and then we'll also get a more understandable track as the main BG. That way if we have to completely remove something we still have stuff to fall back on.