This is a cover of one of my favorite songs my friend and I did to play in church. The only thing I had left was a horribly mixed cassette copy and my 4 track copy of Steve’s guitar part recorded on the Tascam 244. So after finding no possible way to rescue the cassette copy, I decided to recreate all the other parts in LogicProX and then import Steve’s guitar track from the 4-track tape. Unfortunately, that 4-track tape had been recorded with the variable speed trim NOT set at zero, meaning it was just a bit too slow by what turned out to be less than 1 bpm, which caused it to be out of tempo and out of tune.
So what I did was to chop his entire guitar track into separate notes and very small phrases. But since the overall track was longer than the LogicProX track, it meant that there was a little overlap of the end of each fragment with the beginning of the next. So what I did was to make two separate audio tracks, A and B, on LogicProX and position the fragments alternately and sequentially where each should be hitting with respect to the other tracks. When I got them were they should be, a very tiny fraction at the end of the fragments on track A would overlap the beginning of the fragments on track B. Then I used control T to split the ends of A at the beginnings of B. When it was all done, there were no overlaps and no gaps. I combined the two tracks and then used control J to join all the fragments into a continuous audio track.
Now the timing was right but the pitch was not. So I exported that guitar audio track and imported it into Melodyne by Celemony and moved all the slightly flat notes into the correct pitch without changing the track’s overall duration. I reimported that into the LogicProX project onto another audio track. Now my friend’s guitar part was playing at the right tempo and pitch.
I know now there’s a feature on Logic called Time Stretch which could be used to adjust the length of the slower, longer track to the new version and another way to adjust the pitch. But I realized that the way I did it was probably better. For a single note there’s its attack, sustain, release, and decay. If the end of one phrase is is not continuous with the beginning of another phrase, it’ll still sound fine if it’s only a fraction of 1 beat longer, as long as it’s the correct pitch. And on continuous notes in a phrase, having only a tiny last part of the sustain of a previous note whacked at where the attack of the next note begins isn’t going to be obvious over such a short space of time because that sustain was going to be ending on the attack of the next note anyway. And doing it this way meant that there was no resampling done of the entire track and random bits of audio subtracted along the way, everything else then just being contracted to eliminate the microscopic gaps. Something’s being lost either way, I’m just making sure that it was at the last tiny bit of sustain before the attack of the next note and not spread out throughout everything.
Anyway, it was a fun sort of experiment that resulted in being able to keep my friend’s guitar part that was the best part of the song anyway.
I didn’t include the vocals because it’s now out of my range. I suppose, though, that I could make a scratch track where I lower the pitch of everything, record the vocal, and then use Melodyne to raise the vocal back to where it would have been before. Though maybe a combination of slightly raising the pitch of the recorded vocal while slightly lowering the pitch of the music part (excluding the drums), would give the best overall result.
And since the thieves did not get my microphones or my AudioBox D/A converter, that’s something I could at least try to do just for the sheer experimental pleasure of it.
At the time we did this, there were two versions of the song, the one by Keith Green and the cover by Phil Keaggy. I thought Keith Green’s version sounded too breathlessly devotional and Phil Keaggy’s sounded too slick. Since then there have been many versions of this song. Almost all of them but one (the one done by Russ Taff) suffer from the same thing: the artists, rather than taking the song for what it is, all seems to be trying to use it to do their own special thing.
I know that anyone’s interpretation of a song is going to be his own special thing. And I know that Martin Luther was supposed to have said something to the effect of “Why should the Devil have all the good music.” But a song has a certain character. As my friend said about composing a song, “Let the song become what it wants to be,” meaning that trying to force it to fit what you had originally conceived before starting to actually work on it can vitiate it. And I think the same is true of doing a new version of an existing song. There are things you can do that will enhance the inherent nature of that song and then there are things that are just gratuitously pimping your ride.
I mean, just listen to some of the different versions of Yesterday. They are not all equally valid interpretations of that song. Some, even by well known “artists” are absolute crap and were meant to be nothing but, “Hey, listen to ME and what I did to this song.”
Listen to the Verve Remix series. There are some, well, most, of those remixes that make the song even better than the original, even though a whole lot’s been done to it. For example Sarah Vaughn’s Whatever Lola Wants. It was written as a tango, but it was short on some essential tango elements, a little too fast in tempo, and almost entirely lacking an introduction to set the mood. The remix version fixes all of that in a really spectacular fashion and only enhances Sarah Vaughn’s voice. The same is true of the remix of Billie Holiday’s Speak Low. It really cranks up the essential sentiment of that song. And neither of those remixes is the equivalent of turning a Willie Nelson song into a polka.
08 LoveBrokeThrough(KeithGreen)(SteveConnelly, guitar)-1989.4