Most people think of samplers as keyboard instruments. But within that deceptively normal-looking exterior lurks a bunch of digital audio recorders - and they sure can come in handy during a mixdown. Craig Anderton explains how...

Uh-oh ! You're doing a mix and you start hearing things... like clicks and pops, and out-of-tune parts, and even a fluffed note on the piano section in the second verse. And the singer didn't quite make that high D...

In the pre-sampling days, these problems would have meant either redoing a track, or getting intimately involved with a razor blade and splicing tape. But no more. Although most people look upon samplers as sampling keyboards, this is only one way to look at these wonder boxes. In the studio, for example, you might be better off thinking of a sampler as a bunch of little digital audio recorders, controlled by switches that - just by coincidence - come configured as a piano-style keyboard. With our consciousness thus properly shifted, we're ready to 'fix it in the mix' and get into some serious cut-and-paste work.

Most of the following applications involve recording a part from tape into the sampler, manipulating it within the sampler, then recording the part back on tape in place of the flawed original. Thanks to 12- and 16-bit sampling systems (or the better 8-bit systems), the loss in fidelity that occurs during the transfer is minimal - and is certainly preferable to whatever glitch you wanted to eliminate.

STOP THAT PLOSIVE!

One of the quickest ways to spoil a great vocal take is with a popped 'P' or 'B' sound. Dropping in one word can be a big deal, since it's going to be difficult for the singer to get the same feel and match the level. Besides, what happens if sloppy engineering on the original session let through that exploding 'P' sound, yet you have to solve the problem during the mix when the singer isn't around to re-do the part?

Simple. Sample the offending word, complete with pop, into the sampler. Next, use the truncation function to cut off the pop at the beginning of the sample. Well, maybe not the entire pop; usually you want to leave in just enough to retain a bit of the 'P' sound. When the plosive has been properly tamed, send the truncated sample back into the recorder and punch-in over the original word.

What? You say it's a short word and who can do that quick a drop-in? Then transpose the sample down an octave, slow the recorder down to exactly half speed, and do the transfer. Or if the word is embedded in the middle of a phrase, sample the entire phrase and copy it to a second sample. Then, truncate the end of the first sample just before the pop, truncate the beginning of the copy just after the pop, and splice the two samples back together again and record it over the original phrase. Usually any pops are short enough that the few milliseconds lost in the splicing process won't significantly affect the overall timing.

THE TRANSPLANT

Thanks to the repetitive nature of most pop tunes, melody lines and parts often repeat in more than one place. Recently, I was mixing an album project where, in the second verse, there was an audible electrical click on one of the notes in a part. The musician who originally played the part was on tour, so she wasn't available to replay the line.

I checked back in the song and, sure enough, there was an almost identical line in the first verse. We sampled it, and recorded it over the line with the glitch. The process worked so well it's impossible to tell which is the 'transplant' and which is the original.

TESTING A SPLICE POINT

Do you get as nervous as I do when splicing up a multitrack tape? It's no fun to slash a 2-inch tape into pieces, only to find that maybe that snare stroke wasn't the right splice point after all.

The usual procedure is to test your splice by recording a mix of the multitrack on to quarter-inch tape, and then performing your 'practice' splices on the quarter-inch version. Once you find the right place, you then do the 'real' splice on the multitrack master. This is all well and good, but no longer necessary if you have a sampler with a splicing function.

Start by sampling a couple of seconds of material around each of the potential splice points. Truncate the samples to the splice points, then use the splice function to glue the truncated samples together. If it works, great - go do the splice. If not, re-truncate the samples and try again. When you find the right splice point, you can proceed to slicing and dicing the multitrack master.

FIXING THE OUT-OF-TUNE NOTE

Here is where a sampler can truly save a session. The singer has just given a stunning performance except... that last note was just flat enough to blow the take.

Or was it? Sample the note into the sampler, and drop-in over the existing bad note. Only this time, artfully move the sampler's pitch bend wheel at just the right moment to bring the note up to pitch. This is the kind of salvage job that can make a singer your friend for life.

Incidentally, you can also use this technique during the recording process to help a singer hit a note outside of his or her range. Sample the highest note the singer can sing, and when it's time to hit that ultra-high note, play it on the sampler instead by transposing the sampled sound upward.

STEADY, STEADY...

What happens if the bass drum is just a shade off tempo? You guessed it: assuming the drum has been recorded on its own tape track, you can sample the sound, then during the transfer back to tape hit the sampler key at the right time so that the bass drum falls right into the pocket.

NO-VIBRATO

During a mixdown once, I heard a vocal line that was just crying out for a really heavy, thick, even vibrato at the end of a phrase. Unfortunately, not all singers can provide this... but a sampler can!

All you need do is sample the note, and when you record it over the existing note, turn up the modulation wheel for the desired amount of vibrato. You can even do some really bizarre things like simultaneous pitch bend and vibrato - pretty bionic sounding.

BUT I'VE RUN OUT OF DDLs!

You've got one DDL chorusing the lead vocal, one on the drum overhead mic, and your very last DDL doubling a sax solo. At this point the rhythm guitarist muses about how it would be nice to have a little eighth-note slapback echo every time he hits that fancy B maj 7th chord... you know, a tail of about three or four echoes that fade out nicely into the drum fill.

Here's where velocity keyboards come in handy. Sample the rhythm guitarist's fancy chord, and play the sample on the keyboard when the echo is supposed to come in. Play it once for each echo, and play the note a little softer each time so that the echo's level diminishes properly. If you have a spare tape track, the echo can be recorded there and you won't have to think about it any more. If all your tracks are full, find someone to play the part in real time during the mix itself.

I must confess to using this technique a lot, even when I haven't run out of DDLs, since it allows for echo effects that would be just about impossible to achieve with standard DDLs (such as strange polyrhythms, echoes that grow louder and then softer, and so on).

TESTING, TESTING...

And since we're in a mix-oriented frame of mind, we want to make sure everything is aligned level-wise. Sample a 100Hz, 1kHz, and 10kHz tone into the sampler; they won't take up much memory since all you have to do is record a few cycles and loop them.

Then make up a little sequence that spits out these tones in the right order for calibration purposes. Some samplers will have a hard time dealing with a 10kHz signal, but if you sample at a fast sample rate and sample the tones at a moderate level, you should have something good enough to use.

Keyboard players who would like to endear themselves to guitarists can sample an E= 164.8Hz tone into their sampler. When transposed down an octave, this produces the right frequency for tuning the guitar's low E string. When transposed up an octave, you can tune the high E. Notes for the other guitar strings (A, D, G, B) lie within this two-octave range. You can even make up a little tuning sequence - say, ten seconds of each note needed for tuning.

AND THERE'S MORE!

While technically not a 'fix it in the mix' application, some samplers (Emax, Prophet 2000, Akai S900 et al) let you slow the sampling rate way down, thus allowing very long sampling times. This can be a real boon to songwriters. I don't know about you, but I often come up with song lyrics by playing a chorus or verse over and over again until I've got most of the words filled in. In the Dark Ages, this meant rewinding your tape a lot. Now Tascam, Fostex, and others offer 'block repeat' functions where the tape will rewind back to a set point, play forward to another set point, rewind, play, and so on. But why torture your tape deck transport? Just sample the entire verse, chorus, or whatever, and loop it.

TRUNCATE END

Samplers make great musical instruments, no doubt about that. But they also make versatile digital recorders with very useful editing capabilities. So next time you need to fix something in the mix, see if a sampler won't do the job in the quickest, simplest, and most cost-effective manner.

© 1986 Electronic Musician magazine ((Contact Details)) and used with the kind permission of the Publishers.