There are a couple of multichannel records being released soon that have me thinking about the way 5.1 is mixed... John Lennon's Imagine and The Beatles White Album. The team producing Imagine have put out some press about their techniques and it seems wrong headed to me. They are doing surround mixes that they call "raw surround". The idea is that each channel on the master tape is assigned to a specific channel on the blu-ray. Lennon's vocals in the center. Stereo drum spread in the left and right mains, along with keyboards left and Lennon's guitar right. Bass left rear Guitar fills right rear. That seems to me like it would sound awful. All of the early quad mixes where instruments are isolated in specific channels sound really thin. And putting the bass in one side of the rear when most people use bass management with a sub up front seems like it would sound like the bass was smeared across the room with the low part up front and the high part to one side in the rear. They say that this raw mix technique makes it sound like it would if you were in the studio standing right in the middle of the band. I'm not so sure. It sounds like lazy mixing to me.
The White Album is a totally different approach. Like Sgt Pepper, it's being mixed for Atmos (even though it isn't being released on blu-ray that way for some reason). The individual elements of the mix are placed in three dimensional space, so if something is in the middle of the room, it's coming out of all five speakers and meshing in the middle as a phantom center. A lot of people complained that Sgt Pepper's mix sounded like stereo. They called it a very conservative mix. Most of those complaints are because the rear channels were mastered -6dB too low. They never admitted it, so people were listening without hearing the stuff in the center of the sound field, because the fronts and rears were meshing a few feet in front of the rear wall. But I've read some people say that even if they boost the level of the rears to the proper level, it still seems like a flat mix. I was wondering how this could be, and I've come up with a theory.
In the early days of quad, sound elements were localized in specific speakers. I call this "the four corners approach". Most movies and a lot of multichannel music is recorded by pairing sets of speakers... for instance front left and rear left... and then balancing the levels to place the sound somewhere along the left wall. I call this "the four walls approach". They might pair the left and right mains, along with the center channel to define the front soundstage; or pair the two rears to place something along the back wall. But rarely do they pair diagonally across the room. I think the reason for this is that the wide variety of room acoustics make it difficult to predict if pairs would mesh properly that way. It's more foolproof to place along walls.
But Atmos uses the "sound field approach". There's no pairing of 2 channels. All of the speakers contain bits of the sound field in varying levels so the sound localizes in the middle of the room. It seems to me that this approach only works if all of your speakers are full range and the response and levels are precisely tuned to the room. If there is any problem with the room's acoustics, the sound field dissolves and it's just bland directionless sound coming at you from all directions.
When I play Sgt Pepper, For the Benefit of Mr Kite has a section with swirling calliope music that has a sound that I can only describe as being like a pitcher of water being stirred in a circle. The sound doesn't just go around the walls of the room. It seems to swirl out from the center of the room like a rotating pinwheel. And when I play Good Morning, the fox hunt starts in the right front channel and crosses diagonally across the center of the room ending up in the left rear channel. As it crosses, the phantom center is solid. No dipping in the middle at all. However if someone's rear channel speakers were small, or if there was something blocking them, I can see how it would sound like the fox hunt starts out in the front right and fades out as it fades up behind you in the left rear.
Immersiveness seems to depend on tight control over the phantom center of the room, but most people's systems aren't capable of doing that without added speakers. A typical system can easily do "four corners" or "four walls" because that only involves the edges of the room. But if you have full range power in all channels and the room acoustics to allow a phantom center to be formed you can do object based sound fields. That might be why some people think the Sgt Pepper mix sounds flat and non defined, while others think it is immersive and well defined.
