[Proposal] Alien custom mix

[Chewtobacca]

Abstract: The proposal is to make a custom 2.1 mix for the theatrical cut of Alien, using the 2.0 LD PCM of the Sensurround mix as a base.

Initially, I had intended to propose a fully discrete mix based on the Sensurround AC-3 5.1 from the LD, but the recent discovery that everything above 18 Hz is cut off in the 5.1 (probably a consequence of the encoding) makes it seem as if the 2.0 PCM is the better source.

Here is a spectogram of the 5.1.

And here's one of the 2.0.


As the Sensurround mix is something of a rough draft, the basic idea is to splice in sections in which the finished mix is superior from a 2.0 PCM source.

At present, I'm not sure of two things:

1. if there is any interest;
   
2. if I'm the best person to take it on.
   

The latter will depend on the extent of the musical differences and what people want to see from a custom mix.

Ideas

• retain alternate Goldsmith score at start
  
• put back missing dialog
  
• replace sections in which the Sensurround's sound editing is rough
  
• retain alternate take of "Good morning, Mother."  I don't know about you, but I prefer it.
  
• take the LFE straight from the LD AC-3 5.1 (deep bass throughout)
  
• encode as DTS-HD MA 2.1, allowing the receiver to handle the upmix
  

Useful Information

Disclord's discussion of the Sensurround mix

Here are msycamore's notes about how the Sensurround mix differs.

http://originaltrilogy.com/post/id/600844

• Jerry Goldsmith's score on the title sequence and opening is a little bit different, pay attention when we first enter the corridors of the ship, the 35mm is scored right away.
  
• What appears to be a distant horn emanating from Nostromo when we first see the ship is absent.
  
• Dallas line "Good morning, Mother." is an alternate take.
  
• Kane's line "Roll 92 degrees port yaw." is absent.
  
• Lambert has an additional line of dialogue heard when they prepare to land on the planet, "900 meters and dropping." (subtitled on the 20th anniversary DVD)
  
• From the moment when Nostromo is approaching the planet until they enter the atmosphere, the sound editing is very rough with a few alternate cues heard. (Deep bass from the engines)
  
• Right before the Space Jockey scene, Kane's lines "Come on down here. Something different down here..." are absent.
  
• When the crew discusses Brett's death, Ash's response "Kane's son." is absent.
  

[PDB]

Well you know how I feel. I think this is a pretty good idea.

[spoRv]

Chew, don't know if it's possible, but what about using the AC-3 5.1 up to 16/17KHz, and the PCM 2.0 dematrixed for higher frequencies? Doing so, it will retain most channel separation of the discrete track, but with the high frequencies of the matrixed one.

[zoidberg]

The 18kHz roll-off is typical of 384kB/s AC3. Is the 2.0 a direct fold-down of the 6-channel or a distinct mix of it's own?

[Jetrell Fo]

I think this is a cool idea for sure.

[Chewtobacca]

The 18kHz roll-off is typical of 384kB/s AC3. Is the 2.0 a direct fold-down of the 6-channel or a distinct mix of it's own?

Yes, it's a consequence of the encoding. The PCM is the same mix. See the spectrograms.

Chew, don't know if it's possible, but what about using the AC-3 5.1 up to 16/17KHz, and the PCM 2.0 dematrixed for higher frequencies? Doing so, it will retain most channel separation of the discrete track, but with the high frequencies of the matrixed one.

That occurred to me, too, but it's highly theoretical and way beyond my ability.

[Jetrell Fo]

The PCM 2.0 looks lovely like that and I know it sounds just as lovely.

[AntcuFaalb]

That occurred to me, too, but it's highly theoretical and way beyond my ability.

Let's work on this together, then. I think I can figure it out.

Step 1 is to DPLIIify the 2.0 into 5.1.
Step 2 is to take the true 5.1 track, figure out the rolloff point, and low pass filter it a few Hz below that to a known value.
Step 3 is to take the DPLII 5.1 and high pass filter at the value from Step 2.
Step 4, I think, should now just be a simple sum.

Edit: Oh, and before any of this the two should be normalized to be as close as possible.

[Chewtobacca]

Presumably, step 1 would have to be done with a hardware decoder. I know a number of ways to handle the upmix in software, but none of them are stellar. As PDB has previously mentioned, the music is likely to be a problem, even with DPL2.

We also have to account for the fact that both 2.0 and 5.1 versions of the Sensurround have been edited to fit the BD by different people -- not that there is anything wrong with either of the syncs, just that the edit points might contain differences that could lead to distortion.

[spoRv]

I know a number of ways to handle the upmix in software, but none of them are stellar.

...IIRC I wrote it before...