The Smiley Smile Message Board

Ok we all know that the 45 version vocals have gone the way of the dodo, but I was pondering this last night: what about the vocals for the nearly released version that first saw release on the 1981 1983 Rarities album? It may not have the verse vocal sections, but the rest of the song had pretty much completed vocals. One could theoretically make a stereo GV utilizing the alternate vox. Anyone know if those vocals were ever stumbled across in the BRI archives?

Interesting, but why would they leave it off the Smile box? that was the time to use it. Or maybe they want to milk that song more for the 50th anniversary box. Any who could ask some Capitol exec?

I have to imagine those vocals are gone as well - they only exist as part of a mono mixdown test, right? The three times that track has appeared (Rarities, GV 40th Anniversary Single CD, TSS), it's been in mono; I assume separate elements do not exist.

That version on 'Rarities' doesn't sound 'nearly released' to me, it sounds like a rough demo mix (for instance, there are no vocals on the verses). That version is from a radio station aircheck or something.  I'm sure the multi-track for that one is long gone, perhaps even recorded over before the single was finished.  (And just to avoid confusion, 'Rarities' came out in '83).

The actual 'Good Vibrations' multi-track is missing -- this would be the 8-track(?) master, which would likely have the instrumental track locked in on one track and the other 7 for the vocals.  Although I'm not sure how it was spliced together -- i.e., if the instrumental track was edited, then dubbed to the multitrack; or if the final mix itself was spliced.

Mark Linnet's 1988 rough mix seems to reveal what is left of the vocals on some songs. If that's indeed true, then all that likely still exists of the vocal is...

"I hear the sound of a" in the first verse
The "Good, good, good vibrations" vocals in the chorus
A single-tracked Carl vocal for the 2nd verse, but missing the falsetto "When I look"
A single-tracked Mike vocal for "I don't know where, but she sends me there"
A bit of Brian's "Oh my my" vocal in the bridge, but not the first "Oh my my" part
Part of the bridge "Gotta keep those..." vocals, but it seemingly cuts out, and the "Ahhh!" climax isn't there
The key change "Good, good, good vibrations" vocals, but not the "Nah nah nah nah nah" at the end

The doubled early Tony Asher lyric vocal seems to still exist sans the "And she's already workin' on my brain" vocal, but given the mix on the box set, the entire vocal may still exist. A mono version of the vocal can easily be extracted using phase cancelling on one of the mixes featuring the early vocal on The Smile Sessions.

So yeah. Very fragmented, as you can see. But then that's assuming I was correct in stating the rough mixes reflect what's left of these songs. "Wonderful" and "Cabinessence" seem to confirm this, as the former is missing the lead vocal (and maybe more?) and "Cabinessence" is missing the lead vocal, too, both of which have been confirmed to no longer exist or at least can't be found.

You're right it was 1983. I was confusing it with the Australian LP with the similar name which _did_ come out in 1981. According to the radio special "The Best Summers of Our Lives" from 1976 where this version first made its public debut, the "rarities" version was scheduled for release but was withdrawn by Brian at the 11th hour. All that was missing was a lead vox on the chorus. He instead opted for the hit version. Judging by the PS tracks, I would bet that the final instrumental track took up one track on the 8 track multi. The remaining vox would have been recorded on the remaining 7. More than likely a completely different tape to the hit recording. Who knows??

What's curious is that if you listen carefully, you can hear a verse bleed through....the Asher lyrics...I can clearly hear "She's already..."

How cool would it be if a tape showed up with Carl singing the Asher lyrics?

Never gonna happen, of course :P

I've said it before, but the way multitracks were treated in the 60's and early 70's really frustrates me. Even though no one could have known that 40 years down the line we could sync all these bounce down tapes to one cohesive multitrack, surely they must have thought they these should be preserved. It genuinely makes me a little sad when I read about stuff which was scrapped, destroyed or just tossed in the trash.

There are some tracks, for example 'California Dreamin' by The Mamas and The Papas which I would love to hear remixed, that stereo mix sucks, and there is so much buried in it which is barely audible, such as the second piano in the 'California Dreamin, On such a winter's day' part (listen carefully, you will hear what I mean, it's FANTASTIC!, you can hear it slightly clearer on the mix with Barry McGuire's vocals) and the guitar that plays in the solo. And some where the mix is fine, such as GV, I have no problem with Mono, where they just shouldn't have been tossed... ah well  :-\

It's a shame, yes but ...

The concept of a 're-mix' was pretty strange in the '60s. They were always working toward the final mix ... in fact, if you look at interviews Brian will say things like 'final dubdown', i.e. ... it really wasn't thought of as a 'mix' per se, since most actual mixing took place in various stages during tracking. These terms like 'mix' and 'master' were used pretty differently back then. 'mastering' of course being the actual cutting of the laquer ... making the changes that were required to get a good cut ... the goal was to keep the music as close as possible to the tape within the limitations of vinyl.

All energy was focused on getting that mono 1/4" just right.  Once you had that, that was it.  in fact, some producers didn't even use multi-track recording.  Mickie Most recorded mono to mono, even as late as 1970.

Personally, I kind of like that fact that some tracks can't be touched.  "Good Vibrations" stands as the best recording of all-time in my mind.

Actually....it's happened. SEVERAL times. That's CARL singing the "working on my brain" lyrics. NOT Brian.

Can someone knowledgeable in vocal extractions explain how one can isolate the mono vocals from the GV single mix?

Wait...wha...what?

yeh, I've always heard it as Carl.

This is a long and probably poorly-written explanation by a good sound engineer's standards, but it might give you a general idea. Anyone who knows better, feel free to contribute or correct anything.

===

Let's start with a quick overview of frequencies. Frequency corresponds to pitch: the higher the frequency, the higher the pitch. We can hear a range from 20Hz to 20,000Hz; any possible sound we can hear is made up of combinations of any of these frequencies.

Digging deeper: any musical instrument/voice has a fundamental frequency (the lowest), along with a varying number of upper frequencies. For example: a bass guitar might resonate most strongly from 50Hz-100Hz, but then it might also resonate quietly at, say, 1000Hz. If you play a bass guitar through a pair of laptop speakers, it will sound really thin, but it's not like the bass guitar completely goes away. You can still hear its upper frequencies and the sound of the strings being plucked or picked. The point being: most musical instruments resonate across a wide range of frequencies.

Now, on to vocal extraction.

Vocal extraction is usually done through a technique called spectral editing. A spectral editor analyzes an audio file, looks for relationships between frequencies, and based on that information, tries to figure out what might be an instrument/voice's fundamental frequency and upper frequencies. From there, it can draw out where entire instruments/voices are, which we can then extract or suppress from the mix.

Here's the first four bars of the chorus from "Good Vibrations" as a regular audio file:

(http://dl.dropbox.com/u/20961939/Hosted/gv-normal.png)

And here's the same four bars as analyzed by the spectral editor sonicWORX:

(http://dl.dropbox.com/u/20961939/Hosted/gv-spectral.jpg)

(time is on the x-axis, frequencies on the y-axis.)

As you can see, it's a totally different way of looking at the song. Normally, an audio waveform just shows you dynamics, but in a spectral view you can actually see certain instruments very clearly. The theremin is that squiggly line; the bass and Mike Love's vocal are the descending lines at the bottom; the cello is the sustained low line; the guitar(s?) are the shorter tightly-paired lines higher up; and those short bursts of white are the percussion.

Now, with an older, muddier recording like "Good Vibrations," some frequency groups become harder for the analysis to pick out. For instance, if you isolate just that descending bass line, you'll get something that sounds like Mike Love fused with a cello and a standup bass. That's why those lines are a little harder to discern. On the other hand, it's pretty easy to extract/suppress the theremin -- it's mostly just a single sine wave, which is comparatively easy to analyze.

Is it possible to analyze/extract under the mud? Maybe, but it would be extremely time-consuming, and you'd need very good analysis tools. A few years ago, spectral editing was used to make multitracks out of some early two-track tapes for The Beatles: Rock Band. But that's just two instruments per track -- "Good Vibrations" would be trying to dig through, like, ten instruments.

With tools like these, we can get closer and closer to the vocal. When I tried my hand at extraction last year, I got some vocals out that were pretty clear. But "close" isn't enough right now. As it stands today, it's probably not worth the time and expense to do an extraction that would meet the standards of a commercially-releasable stereo mix. But hey, never say never!

Thanks for the info.  If you have the mono instrumental backing track, can't you also isolate the vocals using phase cancellation?  I've read many discussions concerning this method, but I don't understand how it works.  I've heard some pretty good attempts through the years of combining the mono vocals with the stereo backing track, and I'm wondering if that's how they were accomplished.

Once again, others with better knowledge, please contribute/correct accordingly, but:

If you take a waveform and play it back simultaneously with a phase-inverted version of itself (something that can easily be done in an audio editor), the waveforms cancel each other out. So in theory, if you had the instrumental track and mono master running at the exact same speed, the backing track would be cancelled out and you'd be left with vocals.

The problem with syncing analog sources is that no tape can run at the exact same speed twice. You can get very close -- to the point where normal ears can't discern any musical difference -- but even then you'd probably still have to do a serious amount of digital re-syncing. Beat-by-beat, if not more precise than that. I also suspect that the generational differences between the two tapes -- 1st-generation backing track tapes vs. who-knows-what-generation that was used for the mono master -- means you probably wouldn't cancel out everything anyway, and would still need to turn to spectral editing to take out some remaining artifacts.

The most recent syncs I've heard usually just sync up the stereo backing track with the not-phase-cancelled mono master. Some are pretty good, but nearly all of them fall apart at the bridge and interlude, where the effects of phasing are most egregious.

Thanks again.  Has the mono single instrumental track ever been officially (or "unofficially") released?

Yes, on the Good Vibrations and Smile Sessions box sets.

Your best source material would be lossless versions of the Smile Sessions tracks, folded down to mono. If you have the right plugin, you can do this non-destructively, while adjusting how much of the left and right channels goes into the fold down. With just commerically-released material, that's probably the best way to do it: you can get really close to the balance of the mono master's backing track mix, and you avoid phasing issues with the stereo track while you do your retiming.

If I remember correctly, it's in bits and pieces on different tracks, not assembled as the complete instrumental.


For phase cancellation to be successful, doesn't the instrumental track have to be identical to the one on the single version?  I would assume you could create a fold-down very close to the original, but never exactly the same.

I don't know about TSS, but on GV, yes, the backing track is there, but it's a different edit...I think there's an extra "gooooood, goooood, gooood, good vibrations" iteration at some point that doesn't align properly...(and seriously, it seems to me that the same "sessions" that were on the GV box were also on the Smiley Smile/Wild Honey twofer, TSS, and Hawthorne. Think it'll be there AGAIN on the next archival release?)

There's both. The instrumental track has the chorus cello overdub, which I believe isn't on any of the session tracks. If I recall, the version on the GV box is not a vintage mix, so it probably wouldn't do you any better.


Yeah, see above comment on tape generation loss. Even if you got it close as possible, you would almost certainly still have some leftover artifacts.