Song sequences are an integral part of the Bollywood movies. In fact, it's an integral part in almost all Indian movies. Quite often, these songs are solely responsible for the commercial success of movies. Anyways, this question always comes to my mind - while shooting these song sequences do they play the songs in background? I mean, how exactly is the lip syncing done? Some sequences are shot in the studio where it seems plausible, but a lot of sequences are filmed at outdoor locations, like a desert or a mountain or some famous monument. How is the syncing done?
Normally film crew use the song in the background. It will be a raw sound from the music Director used specifically for the shooting purpose. What we hear later is a modified version sung either by the same artist or different artist as per their wish but it has more technical modification once it reach to our ears. No matter how crowded the region is, the film crew/team will take care of public from not intervening in their shooting activity.
And this goes to the outdoor songs as well, where explicitly you need a close up of actors while singing otherwise, playing song is not required. For lip-sync, actors mumble to themselves, if actors are not native speaker and a close up of actors singing is involved in the worst case, directors ask them to say "abcd" or something close to those lyrics, and will be concealed through camera angles to give viewers an impression that they are indeed singing. Later a huge portion of issues are sorted in the editing table.
I've only ever worked on one musical scene on a Bollywood movie, so this is by no means a comprehensive answer, but agreeing with Jason in general...
For a large outdoor musical scene - this particular one shot at London's Marble Arch - something akin to a large concert PA system was rigged across the entire location, so that the music was loud and immersive enough for everyone involved in the scene to be able to react to the beat and know exactly where they were in the song, in relation to their proscribed moves during the scene.
This is so that, across many takes from several angles, all the people in the scene will be in the exact same place at the same time, from each angle - so that when the final scene is edited together, people aren't mysteriously appearing in one place, then another, depending on the camera angle.
From a lip-sync perspective too, it's far easier to get your lip-sync right if the music is really loud enough to sing along properly than if it's quiet and you have to mumble to be able to hear the track over your own voice.