Hello! I was looking at Jethro Tull's quad mix of Aqualung, as it is presented on Blu-Ray in the 40th Anniversary Collector's Edition and on DVD in the 40th Anniversary Adapted Edition, and I noticed a few errors in both releases: the back left channel of the mix is in the entirely incorrect polarity, the LFE channel is out of phase (and sometimes in the wrong polarity), and the back two channels of "Hymn 43" were accidentally swapped.
I meticulously tried to remedy these errors to the best of my ability.
The Back-Left Channel's Polarity:
The entire polarity of the back left channel's audio needs to be inverted for the mix to be correct. Having an element in the wrong polarity can mess up the stereo (or in this case, surround sound) image. Hear these 24-second clips to see what I mean:
Mod Note: Links to Copyright Material not allowed.
Notice how in the fixed back channels, you can make out the bass, flute, and vocals in the center, whereas in the original, it sounds as if those elements are hard panned to both the left and the right.
The LFE Being Out of Phase, Sometimes in the Wrong Polarity:
Across all the songs (except perhaps in "Slipstream"), the LFE is at least a tenth of a second out of phase with the rest of the mix, and across 7 songs, it appears to be in the incorrect polarity as well. In "Cross-Eyed Mary" and "Mother Goose" specifically, the LFE is over a whole second out of phase from the rest of the mix.
Heavy use of a correlometer and a low-pass filter helped me figure out exactly where the LFE needed to be in the mix to properly be in phase. The low-pass filter made it so I could eliminate the high frequency elements from the mix so I could properly compare it to the LFE channel.
First, I took a channel that I knew had low-frequency audio that would also be present in the LFE channel (in this case, the front right channel):
Next, I applied a low-pass filter to both channels. (In this case, I used a 70Hz low pass filter with a 48dB roll-off.) The reason I apply a low-pass filter to the already low-passed LFE audio is because the low-pass filter changes the phase of the signal, i.e. it makes the entire signal later than it should be.
You can see below that the LFE channel is so out of time with the rest of the mix in this case that you can already tell when seeing the signal zoomed out.
Next is the trial-and-error process of shifting the low-passed LFE over to get it to a place where it looks in phase with the rest of the mix, then checking with a correlometer to see if it actually IS in phase. Only pay attention to the low-frequency elements on the correlometer, as only those frequencies matter in regards to the LFE.
However many samples your LFE was shifted to get it into phase with that other low-passed channel, THAT is how many samples you need to shift the LFE by. In this case, the LFE needs to be shifted 267,088 samples later than it was originally.
Open the file again, and with that number of samples in mind (267,088), I generate that number of samples of silence to the ORIGINAL LFE (or, if the LFE needs to be shifted to the left, delete that number of samples from the start of the LFE)...
And viola, your LFE is now in phase!
As the LFE seemed to be out of phase by variable amounts across the mix, I note all of my specific edits at the bottom of this post!
The Swapped Back Channels of "Hymn 43":
I believe the back two channels of "Hymn 43" should be swapped because of the panning of the guitar and piano in the front channels. At 41 seconds, you can make out a piano panned to one side in the front channels, and yet that element is panned to the opposite side in the back channels. The same is the case with a guitar present at 23 seconds. As no other tracks deviate from the panning scheme in this manner, I think it's fair to say that these two back channels of "Hymn 43" ought to be swapped to create the intended mix.
Every Single Edit I Made, Including Specific LFE shifts:
Keep in mind that the sample values are for the 96kHz Blu-Ray release. If you own the DVD, divide the number of samples by 2 (as that release is 48kHz)
Before shifting the LFE...
*I call it "Almost" original because I swapped the back two channels, as per what I noted earlier.
I meticulously tried to remedy these errors to the best of my ability.
The Back-Left Channel's Polarity:
The entire polarity of the back left channel's audio needs to be inverted for the mix to be correct. Having an element in the wrong polarity can mess up the stereo (or in this case, surround sound) image. Hear these 24-second clips to see what I mean:
Mod Note: Links to Copyright Material not allowed.
Notice how in the fixed back channels, you can make out the bass, flute, and vocals in the center, whereas in the original, it sounds as if those elements are hard panned to both the left and the right.
The LFE Being Out of Phase, Sometimes in the Wrong Polarity:
Across all the songs (except perhaps in "Slipstream"), the LFE is at least a tenth of a second out of phase with the rest of the mix, and across 7 songs, it appears to be in the incorrect polarity as well. In "Cross-Eyed Mary" and "Mother Goose" specifically, the LFE is over a whole second out of phase from the rest of the mix.
Heavy use of a correlometer and a low-pass filter helped me figure out exactly where the LFE needed to be in the mix to properly be in phase. The low-pass filter made it so I could eliminate the high frequency elements from the mix so I could properly compare it to the LFE channel.
First, I took a channel that I knew had low-frequency audio that would also be present in the LFE channel (in this case, the front right channel):
Next, I applied a low-pass filter to both channels. (In this case, I used a 70Hz low pass filter with a 48dB roll-off.) The reason I apply a low-pass filter to the already low-passed LFE audio is because the low-pass filter changes the phase of the signal, i.e. it makes the entire signal later than it should be.
You can see below that the LFE channel is so out of time with the rest of the mix in this case that you can already tell when seeing the signal zoomed out.
Next is the trial-and-error process of shifting the low-passed LFE over to get it to a place where it looks in phase with the rest of the mix, then checking with a correlometer to see if it actually IS in phase. Only pay attention to the low-frequency elements on the correlometer, as only those frequencies matter in regards to the LFE.
However many samples your LFE was shifted to get it into phase with that other low-passed channel, THAT is how many samples you need to shift the LFE by. In this case, the LFE needs to be shifted 267,088 samples later than it was originally.
Open the file again, and with that number of samples in mind (267,088), I generate that number of samples of silence to the ORIGINAL LFE (or, if the LFE needs to be shifted to the left, delete that number of samples from the start of the LFE)...
And viola, your LFE is now in phase!
As the LFE seemed to be out of phase by variable amounts across the mix, I note all of my specific edits at the bottom of this post!
The Swapped Back Channels of "Hymn 43":
I believe the back two channels of "Hymn 43" should be swapped because of the panning of the guitar and piano in the front channels. At 41 seconds, you can make out a piano panned to one side in the front channels, and yet that element is panned to the opposite side in the back channels. The same is the case with a guitar present at 23 seconds. As no other tracks deviate from the panning scheme in this manner, I think it's fair to say that these two back channels of "Hymn 43" ought to be swapped to create the intended mix.
Every Single Edit I Made, Including Specific LFE shifts:
Keep in mind that the sample values are for the 96kHz Blu-Ray release. If you own the DVD, divide the number of samples by 2 (as that release is 48kHz)
Before shifting the LFE...
- Invert 5th channel across the whole mix
- Cut 1.3 seconds from start of "Wind Up" and put it at the end of "Locomotive Breath"
- ("Wind Up" starts with the fade-out of "Locomotive Breath", so I thought it would be better to append that bit to the previous track.)
- THEN move "front left" channel 1 sample to the right, move "back left" channel 11 samples to the right, and move "back right" channel 10 samples to the right (at 96kHz)
- (This minor sample shift may not make much of a difference, but I did it anyway.)
- Move LFE 1,230 samples to the left
- Move LFE 267,088 samples to the right and invert polarity (2 sec and 75,088 samples)
- Move LFE 2044 samples to the left and invert polarity
- Move LFE 125,004 samples to the right (1 sec and 29,004 samples)
- Move LFE 2,076 samples to the left and invert polarity
- Move LFE 2,132 samples to the left and invert polarity
- Move LFE 1,238 samples to the left
- Move LFE 2,056 samples to the left and invert polarity, swap the back two channels
- (LFE is too quiet to make any difference)
- Move LFE 2,168 samples to the left and invert polarity
- Move LFE 1,992 samples to the left and invert polarity
*I call it "Almost" original because I swapped the back two channels, as per what I noted earlier.
