New Monday #74
Happy Monday -
It looks like I am swinging towards techie for the next few weeks.
But here's a thing to listen to, just to get the day going.
https://youtu.be/S3EKEwMXmXY?
Fishmans
From Japan, Fishmans started out playing dub (reggae). Early stuff is fairly typical and predictable, but there's something about it. But they swerved, and later stuff is what you're hearing. Dreamy, textural, genre unspecific improvisations with a dub rhythmic sense and bizarre, hermaphroditic vocals courtesy of frontman Shinji Sato. Sato died in 1999. The band has carried on, not what it used to be, but still interesting.
This song, 'Long Season', is also the entirety of the album Long Seasons. Wonderfully melodic, strange stuff.
Tape
30 years ago I produced and engineered a great band from Long Island. They had a deal in-pocket and it all blew up. A pity. They had everything going for them, from writing to chops to a look.
The guitarist found the original 24 track 2" masters reels of Ampex 456. He had them baked to re-stick the magnetic coating to the backings, and had everything transferred to digital. So, now I have all these songs as multitracks and I'm remixing them and having a ball in general.
(Side note: did you know if your masking tape is tearing funny it's because the glue on the edges has hardened. If you microwave it for 10 or 20 seconds, it softens the glue and makes the tape usable. This is roughly the same thing that happens to reel-to-reel tape.)
One thing that struck me immediately is how different things sound recorded to tape compared to recording into a DAW. A huge difference. Let's get a bit in the geek weeds about tape.
How it works as fast as I can explain it
Last week I wrote a bit about transformers. Analog tape is basically a transformer split across time. Current flows through the tape deck heads—which actually look a bit like transformers—and induce a magnetic current, called flux, in the tape. Tape is actually made up of millions of little magnetic particles suspended in glue. The flux moves them around, because the glue isn't that solid, and they arrange themselves into a shape that matches (analogous to) the signal that was fed into the heads. To play it back, the tape slithers past the heads, the flux on the tape induces a current in the head that matches the pattern of the particles on the tape and voilá: you have a tape recording.
See how it's kind of like a transformer split in half and across time?
(Side note: if you think about what is going on to make a tape recorder work, it is so unbelievably complex that it's a miracle anyone ever thought of the idea. Had digital recording been developed first, no one would have bothered with analog recording.)
Tape recording has many of the same issues as a transformer. Tape recording has inertia problems.
Tape has Inertia
The magnetic particles on the tape don't really want to move. They're very comfortable sitting there, wrapped in glue (called "binder"). It takes a lot of power to get them to move, and this power comes from a bias signal, which is a very strong signal that overcomes the inertia of the particles such that the actual audio signal can move them. Think of bias, in this case, as a muscled-out goon that threatens the particles with bodily harm unless they do exactly what the boss (the audio signal) tells them to do.
So, the particles get hit with bias and audio, and move into place to match the audio waveform, but they're a little bit slow about it—they're reluctant. There's inertia both coming and going: they don't want to start and once started, they don't want to stop.
Think about this: you have a fast transient coming in, and the particles are slow to react to it. What will happen to that transient?
It gets clipped, right? Very subtly, in many cases inaudibly, but if the tape can't respond to the transient quickly enough, the transient isn't accurately recorded. The transient gets rounded off a little bit.
Once the current stops, the tape keeps moving for a bit. That doesn't seem to be an accurate representation of the waveform, does it? It isn't.
Tape is a limiter with an instantaneous attack and an almost instantaneous release.
A tape recording is a really good, but inherently slightly inaccurate, picture of the waveform. This inaccuracy is partly responsible for what people describe as "tape warmth" or smoothness. Things definitely sound smoother when you roll off transients—and remember that most of the high-frequency information on a recording is in the transients. The inertia of the tape distorts the entire waveform, not just the transient. It's always a little slow reacting to the ups and downs
But unlike using a clipper or a very fast limiter to nip off transients, and setting a threshold so that it only happens to loud transients, tape has no threshold. It is doing this "inertia clipping" all the time.
Think of tape as a limiter with an instantaneous attack and an infinitely low threshold. Everyone gets part of their head sanded off, even the children and the pets.
Now, what happens as you increase the gain and slam even more signal into the tape?
Tape compression and saturation
Well, at some point, all the particles on the tape are moved and there are no more to move, and no place to move to, so they basically can't align themselves to the incoming waveform. Now things REALLY start to clip. And when things really start to clip, a lot of harmonic distortion is generated.
This is tape compression/saturation. Remember, compression, saturation and distortion are all degrees of the same thing: non-linear reproduction of the wave form. At first it is subtle: the transients get rounded off. As you shove more signal into the tape, the particles saturate (they can't move anymore) and things distort and start to sound fuzzy.
Drums on Tape
Recording drums on tape was tricky, principally because decks used VU meters for metering and VU meters are slow to respond to peaks. I did a ton of experimenting, having a drummer play and then pushing the signal into the tape at various levels, and found that I had to run things about 10dB down on the meter to get the peaks recorded sounding uncompressed to on tape.
In other words, for drums, -10db on the VU meter was at about 0VU if the meter was fast enough.
Increasing the power would crush things up a little bit. If I wanted a fatter snare, I could push that harder into tape. But if I pushed the overheads into tape, instead of the cymbals having a "tish" sort of sound, they'd have a "pfweish" because the tape would roll off the transient a lot. It would also add a bunch of high-end harmonic distortion.
Your Homework
Here's a thing to do: go to our website and grab an Echoleffe Tape Delay demo.
Get a digital recording of a snare—from Easy Drummer or Slate Drums or whatever. Put it on its own channel and bounce it so it's an audio track.
Now, put the Echoleffe Tape Delay on the insert of that channel.
Click the Tape Saturation Mode button on the front panel. This turns off the delay functions and now you have an excellent tape simulator.
Click the Korneff Audio nameplate to get around to the back. Turn up that gain knob until that snare is squashed a bit. You'll hit a point where it really sounds great—huge, fat and seemingly stretched out across time. Bounce that too, so now you can compare the original with the tape compressed.
Here's what it looks like when I did this. This is with Logic; it renders ugly waveforms. At a glance, it is easy to see the compression. And also that the clipping is asymmetrical. That's a whole other story.
But if you zoom in:
Look at where the cursor is. Do you notice the peaks don't line up, that the waveforms don't line up, and that the "tape" version that went through the ETD is happening a tiny bit later across time.
Yes, tape inertia actually lags the waveform enough that it is visible.
Tape changes the groove??!! How insane is that?
Play with the Echoleffe just as a tape compressor and saturator this week. What does it do to vocals? Piano? Can you get it to respond emotionally? By this I mean I was always trying to set vocals levels such that when the singer got loud, the tape crunched a little, which gave the loud bit a little extra umph, grit and pain.
More homework next week!
Warm regards,
Luke
PS - if you like getting lessons, let me know. If you hate getting lessons let me know. If you have questions, ask them.
