Jump to content

How to extract unique stuff out of Linkin Park songs - walkthrough


Recommended Posts

Okay Astat with this Knowledge you will be Unstoppable.


You know when you look up Tutorials on how to make an acapella and they are all like "If you have the instrumental version of the song you can line it up with the original version and invert the wavelengths to cancel out the instrumental to just leave vocals"


Well it is essentially that method.

but it doesn't always have to be instrumental against original song.

Now if you were to invert a song against itself everything would cancel out and leave complete silence because everything is the same, all frequencies are canceling out the exact same frequencies.


But what do you think would happen if lets say you completely line up the first chorus of a song against the second chorus of the same song?

Would they completely cancel each-other out? they do sound exactly the same right? all chorus in a song are just copies of each-other? Well no.


Imagine mixing a song, having the multi-tracks, and for the first chorus you have every stem at a certain volume, after you have finished mixing the first verse you get up to the second chorus, now do you mix this chorus the EXACT same way as you did the first one? well 99% of the time, the person mixing will tend to mix it just a tiny bit differently, They might add in extra instruments, they might make certain stems abit louder compared to the first one, or they might make the vocals louder.


Now lets say they keep everything the same except the volume of the vocals in the chorus,

So the only difference between Chorus 1 and chorus 2 are that the vocals are louder in the second one.


Now when the song is rendered and you have the file, what do you think would happen when you line up chorus one and chorus two and invert one against the other?


Well since they were mixed differently they will not cancel each other out completely, everything that is the same between the 2 will cancel out and everything that is different will not,

So in this case it is the vocals that are different since they differ in volume, therefore making the wavelengths different.

everything will cancel out except the vocals, Leaving you with an acapella of the chorus.



that method is how i got that result for the SuperXero file, the reason it sounded the way it did is because i lined up the choruses on-top of each other and inverted the phase of the wavelength against one another, so everything that you don't hear and sounded like was missing from the chorus in that was because in the mix Mike or whoever mixed it kept certain stems the same volume, and everything you do hear in it was the stems that he has changed.


But that other extraction i did with Mike's whisper vocals in the Crawling demo, How did i do that?

well this was really fun for me since there are so many alternate mixes of it that have been changed very slightly

i lined up the demo from the 8 trackdemo against the 9 trackdemo

and everything that was the same canceled out, and everything Mike changed showed.

his whispers were perfectly isolated because he changed the volume of that whisper stem in one of the demos but kept the instrumental the same at that part

also his "Rain Drops" verse was slightly different in a few of the demo's

so i managed to get a clear isolation of his verse, and also good isolation of his background vocals in that verse



Other cool things you can extract by doing this is layers of vocals that were added into a later mixed demo.

I phased one of the "Part of Me" demo's against another version and i got left with a weird vocal who i dont know



That voice is there because in the 9 track it is mixed with just chester singing that part

and in the 8track it is mixed with someone else singing, so phased together chester's layer cancels out and gets left with the other layer.


This can be also done to make acapella's

i phased both the 2track version of by-myself against the hybridtheory.com version and because the verses were re-recorded but the instrumental was the same, the beat canceled out and left both the vocal takes isolated


the cancellation was not perfect because the quality between the 2 files were very different from each-other, so always do it with the best quality versions.



Can be used to make instrumentals aswell, like for Carousel i phased 2 demos together and since mike's vocals during the verses were kept untouched but he changed the instrumental his verses canceled out and left the beat.

You can hear backups from chester because they were either changed or added in to one of them.

i always find it funny just before the chorus it sounds like chester is saying "YOU'RE TO WEAK TO PHASE ME" and then his voice does not get phased out :D :D like he predicted i would try this


So yea, this phasing method can basically be used to result in anything, it's all luck, it all depends how the 2 components differ in mixing
Remember if you are phasing choruses together always right click one of them and "Convert to Unique copy" this is just so you can click the invert effect and it will only invert the unique one, without converting to unique copy they will both invert and it wont work since one needs to be opposite another.
Also to my surprise this does not work on Linkin park studio albums. whoever masters their songs knows about this trick and purposely makes it so if you do line up wavelengths they wont phase successfully, this can be done by literally editing the file even in the slightest way so everything becomes uneven.
But linkin park is the only albums i have come across that are actually protected by this
i have only tried it with hybrid theory so i don't know if the others are able to or not.
So basically if a song or a chorus has a counterpart that is just mixed differently u can do this, including clean versions of their songs. when whoever is censoring the songs for the censored album version, they might accidentally change something slightly, or even purposely change some things around thinking that nobody would ever find out ;)
So inverting the clean version against the normal version will result in whatever they changed.
it usually results it getting clean vocal samples of them swearing since that's what they change.
but sometimes clean versions of songs are alternate mixes aswell.
And some files you try will be the same and will just leave you with a wet isolated sound, alot of the demo's on the 6/7/8 tracks are the same, like The Untitled is the exact same version.
This can be done on pretty much every audio program software.
remember when aligning the tracks they must be perfectly aligned, so zoom in completely
and it may be very hard to start lining the waves perfectly and will feel like trying to find a needle in a haystack
but after looking at them for long enough you will start to understand them.
So for anyone wanting to try this, i will link you with 2 files to download, one will be of one of the crawling demo's from the 9 track and the other will be the demo from 8 track, i have already aligned these 2 files so all you have to do is put one ontop of the other one and press invert on 1 of the files, and if you hear Mike's whispers you have done it right, Just so you kind of get the jist of what do to, You will have to do the allinging work by yourself if you want to do it with any others
If i have forgotten to say anything here, i will edit this later on
but hope you guys get the jist of it :) I turned 17 this year and looking forward in creating alot when it comes to music
If you got any questions or you have extracted some cool stuff, post em' here
Edited by Oliaggy
Fixing spelling and removing pointless links.
Link to comment
Share on other sites

From an engineering standpoint, there's no reason to "protect" a mix from something like this on an album, as you claim. The reason this probably works more reliably on early demos is because of the simple fact that they were much simpler mixes. Let's break down a likely tracking sheet for In the End (Hybrid Theory album version) for example:




1. Kick in
2. Kick out

3. Snare top

4. Snare bottom

5. Tom 1

6. Tom 2

7. Floor tom

8. Hi-hat

9. Ride

10. Overhead L

11. Overhead R

12. Room L

13. Room R


14. Dist. Guitar 1a

15. Dist. Guitar 1b

16. Dist. Guitar 1 room mic

17. Dist. Guitar 2a

18. Dist. Guitar 2b

19. Dist. Guitar 2 room mic

20. Dist. Guitar 3a

21. Dist. Guitar 3b

22. Dist. Guitar 3 room mic

23. Dist. Guitar 4a

24. Dist. Guitar 4b

25. Dist. Guitar 4 room mic


26. Dist. Guitar 5a (harmonics)
27. Dist. Guitar 5b (harmonics)

28. Dist. Guitar 5c (harmonics) - room mic

29. Dist. Guitar 6a (feedback 1)

30. Dist. Guitar 6b (feedback 1)
31. Dist. Guitar 6c (feedback 1) - room mic

32. Dist. Guitar 7a (feedback 2)
33. Dist. Guitar 7b (feedback 2)
34. Dist. Guitar 7c (feedback 2) - room mic

35. Clean Guitar 1a (verse harmonics)

36. Clean Guitar 1b (verse harmonics)
37. Clean Guitar 1c (verse harmonics)

38. Clean Guitar aux track (delay + reverb)



39. Bass (Mic A)
40. Bass (Mic B)
41. Bass (D.I.)
42. Bass aux track (blend of D.I. and mic'd tracks)

43. Piano loop (dry)

44. Piano loop aux track (added reverb)

45. Piano outro (dry)
46. Piano outro aux track (added reverb)
47. Strings 1

48. Strings 2



49. Samples 1 (scratch-y loop 1)

50. Samples 2 (scratch-y loop 2)

51. Samples 3 (beeping noises in verses)
52. Sampled kick

53. Sampled snare

54. Sampled hi-hat closed

55. Sampled hi-hat open


56. Mike main vocal

57. Mike vocal doubles

58. Mike vocal glitch effects

59. Mike vocal other effects (delay + reverb)

60. Chester main vocal

61. Chester vocal doubles

62. Chester vocal effects (delay + reverb + pitch correction, etc.)


...And this is honestly a VERY conservatve estimate. There are probably way more tracks for the sampled percussion parts, each portion of a vocal that was comped together from multiple takes will likely exist on a different track, I didn't include any aux tracks for the distorted guitars/bass/drums, all of the guitar/bass/drum tracks assume that those parts were recorded in a single take which almost never happens...there are realistically probably 80 to 100 tracks of audio in that song. I doubt very much that there's an instance where ANYTHING will invert perfectly over itself in two different parts of the song just based on how much information is being crammed into two channels of audio.

Compare that to the demo version, where you're probably dealing with a bunch of stuff that's been combined down to a 24 or 32-track mix based on limited studio resources, plus the fact that the demos are never brickwalled the way the album mixes are, and suddenly each element of the song has a lot more room to stand out on its own. If something's copy-pasted in a different section of the song, you have a much higher chance of being able to cancel it out via inversion because there's not nearly as much stuff taking up space in the mix. There's also a lot less reverb and such on the demos so you're working with much dryer signals on individual tracks, so they're easier to isolate without access to stems.

TL;DR - The fewer tracks in a mix, the higher likelihood of your technique for isolating elements of said mix working.

Edited by Astat
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...