Making a Tremolo follow Tempo

[Scott Swartz]

I have a recording project I am working and need to make a tremolo effect follow the time of the song, ie a tap tempo application.

I needed a quick solution, so I tried the old trick of keying a gate with the tempo input (I had the drummer play the 1/8 note tempo on a wood block), and setting the attack and decay of the gate to get to a tremolo.  Its works fine tempo wise, but since the driver is a triangle wave with a flat top it doesn't sound as good as a real tremolo.

What I would like to do is feed the tremolo a freqeuncy modulating LFO signal generated from the wood block track.  It is trivial to break into the trem circuit and add the external LFO input.

The trick is how to generate the frequency modulating sine wave LFO signal, I am trying to come with a block diagram of how to do this.

How about-

1. Feed the woodblock track to a Schmitt trigger, which creates a pulse for each hit.  These pulses can drive a square wave generator, I know I have seen a circuit for this, need to look it up.  Might need to halve or double the frequency also.

2. Run the square wave into an integrator and filter or resistor diode circuit to produce a sine wave.


Or if I can make the square wave, can I simply use that to drive a function generator chip?  I did a quick scan of the XR2206 and ICL8038 datasheets and did not see a way to drive them with a wave.  They can do voltage controlled oscillation.



   

[R.G.]

You really only have three choices.
(1) a phase locked loop  approach
(2) an incredible clot of analog stuff to do a lock and hold on frequency. Something like a frequency to voltage then a sample and hold driving a voltage controlled oscillator should work, but the sub-second cycle times makes this quite involved and ugly.
(3) a digital approach

In my mind, (3) is the only one worth pursuing. Again, two ways:
(a) involved digital logic doing counting clock pulses to get the basic tempo, then down-counting the clock to run a digital output generator.
(b) an eight pin DIP PIC for about $1.50 and some programming time.

The PIC code looks something like:
Init [initialize all the innards and such]
Until forever do
read the waveform switch
lookup next output value in proper section of table
spit it out to output
enddo
on interrupt
read tap switch
if this the second one then figure out tempo
write tempo to ouput speed table
else store the switch time until we get a second one
return
end
Table of sine/triangle/square output values to look up in.

Obviously I've left out some minor details which are dependent on exactly how you arrange the switches and output setup, but those are fairly trivial. This may not be the optimum way to do this, but it's one way.The $1.75 12F629 will certainly do it, the $1.03 10F200 probably will. The circuitry around the chip is all passives, R's and C's.

Did I mention that I have a partially-finished PIC LFO chip project in the works that I can never get time to finish?

Oddly enough, I think we've finally reached the point that Don Lancaster was implying when he said that it was silly to do any discrete logic, even with a 555, with PICs around. The 10F200 is $0.55 in quantity, and among other things, it can do a drop in replacement for the 555 - I think.

[The Tone God]

I personally think it is waste of time and effort to do tap tempo with anything but a uC (pick your flavour). I had some ideas on how to do it but the thing ends up being big if not bigger then the pedal itself plus the time spent would not justify the feature and the cost for my time which I value highly.

I wrote the tap tempo code in a matter of minutes. It works great. Life is good.

Andrew

[Scott Swartz]

RG thanks for the response.

Not sure if my first post was perfectly clear, but I am not interested in extracting a particular tempo and holding the freq at that value.

I want the freq to be free to drift around with the tempo of the input wood block track.  Does that change the scenario for analog?

I am thinking if the tempo is wandering, the triangle and/or sine I extract from the square wave will not have perfect symmetry but that shouldn't matter.

[Rodgre]

In the studio, I use a few tricks to accomplish this.

Most of the time I use the noise gate trigger trick, but if you set the attack and decay right, you CAN get a sine-like waveform, instead of the choppy trem.

Nowadays, I use Protools and draw in my tremolo in time to the song.

When recording, I find it's more crucial to have the drummer slaving to the tremolo/click, than to try and do it later. It's a PITA, but it can be done. Even live, when I need my trem to be in sync, I usually have the tremolo coming from a triggered gate where the drummer is listening to a click in time to the trigger.

I also use an Electrix Mo-FX with tap tempo on the tremolo. You can keep resetting it as the song goes if you drift out of time.

In my experience, unless you're a real electronics wiz and you just love to tinker, it would seem that the woodblock/schmitt trigger/tempo sync idea is pretty involved and sort of "rube goldberg" as they say.

If you can make that work, you're probably going to make a lot of money

Roger

[R.G.]

I usually have the tremolo coming from a triggered gate where the drummer is listening to a click in time to the trigger.

Hey!!! Feed the drummer a synch pulse. Audible for alert, intelligent, obedient drummers, electrical to slow learner drummers. I think you have something here. Maybe we could hot wire the drummer's throne so that if he drifted off time the throne got progressively more electrically unpleasant, if you get my drift. 

[puretube]

optical...

[gez]

If you want cheap and cheerful, double up your click-track and divide down the pulse with a flip-flop to get a 50:50 duty cycle.

A simple RC passive integrator at the output of the FF will get you a relaxation waveform.  Amplitude will decrease as frequency increases but you could amplify the waveform (you’ll need to buffer the waveform anyway) and reset the depth with different tempos.  Alternatively you could set the thing up so that the cap charges to full supply voltage at highest frequency.  Although this gets round the problem of varying amplitude you’ll get a ‘thinning out’ of the waveform as frequency increases.

Getting more complicated, the output of a linear frequency to voltage converter could be used with an OTA to integrate the square and maintain constant amplitude.  A second OTA integrator can be used to do sine shaping (think of it more as a filter than an integrator - works well).

[puretube]

since you`re in the studio,
you`ll have a comp to eq the decrease...

[Paul Perry (Frostwave)]

Is the tempo fixed thru the song? If so, make the drummer play to a click from the tremolo. (it's easy to get a click from a tremolo......even when you don't want one!).
Even if you want the tempo to vary during the song, you could have someone carefully varyingthe trem speed, like a conductor.

[Processaurus]

How about riding the fader up and down on your track in time with the beat on your recorded track?  Too low tech?

If you're recording digitally and are using a click from the PC, or any digital metronome, you can use a tremolo plug-in and do a little math and the trem will sync perfect, for instance if the songs 120 BPM, make the trem 2 Hz or 4Hz.

[The Tone God]

Just some of the things I have done to make drummers stay in time:

- Stopping then restarting the song counting out the tempo like in school

- Chewing out the drummer on mic

- Yelling

- Spitting

- Throwing things (i.e. bottles, glasses, picks, automotive parts, slides, lollypops, instruments, underware, etc.)

- Taking away parts of their kit that they don't need leaving only the basic bits so they can learn to do it right. Bonus points: doing it while they are playing.

- Give them a loooong drum solo in every song they screw up

- Play faster to give them the hint

- Dirty looks / gestures

- Finally I just play the drums myself then let the drummer try again

Back to topic all.

Andrew

[petemoore]

  Greg, the guitarist who 'recommended' the drummer, said "We figured it out", "You gotta give him a lude and a couple shots...then he'll be fine, and play in time".
  It seemed he was in a hurry to get out of there or something, wanting to finish the songs in 1/2 the time...I said "We're on the clock see, working fast here only makes more work", "We won't have enough material to finish the night !!!"
  Certain songs sound good fast, I don't mind 'fast' necessarily, sometimes it works great, but singing and playing leads in double time...alot of the good stuff is omitted...makes the guitar sound stupid.
  "Go ... run around the block 3 times, then play". Snowball.

[bwanasonic]

I like the way tap tempo works with the Line 6 Echo Park, but have not tried their trem with tap tempo. I use a Fulltone Supa-Trem, and there is something to be said for the visual method, or *follow the bouncing LED*. I also like the Blue Ringer's trem sounds. Could a click somehow be used to generate a CV that *syncs* the carrier?

Kerry M

[Transmogrifox]

I agree with RG that a PIC is probably the easiest way to do this.  If you use a simple, small PIC then the code is easy to learn from the datasheet.  I guess you would have the overhead of purchasing or building a PIC programmer, but then that opens a whole new world of things you can do with FX and so on.

If you're just looking for a Trem that follows the drum pulses, but doesn't sound harsh, then just build an ADSR to follow the trigger output.  I know the Korg Delta used a "discrete" ADSR (in that it didn't use an ADSR IC, just op amps & FETS & passives).  I also designed one this evening that I think would work fine, but is yet untested.  To get your "sustain", drive a 1-shot with the trigger input, and make the 1-shot release time variable with a pot.  I have done this, and I can get a variable range from an inaudible blip to a couple seconds of on-time before the release (I was trying a slowgear-type volume swell circuit using an ADSR once and triggered the 1-shot with my guitar).

So when all is said and done, you adjust your attack and decay times so that they arent too harsh, then try to adjust the sustain and release times to approximate the drum sustain and release. 

If that's too hard, just put an envelope follower on the drum input with and adjustable attack and release time.

[robbiemcm]

Is it possilbe for someone with quite a bit of knowledge (say an eletronic engineer) to design and build their own digital effects? I know lots of people say digital is the devil, but it has some applications.

[Paul Perry (Frostwave)]

I also like the Blue Ringer's trem sounds. Could a click somehow be used to generate a CV that *syncs* the carrier? Kerry M

The carrier in the blue Ringer comes from a triangle generator, that is then squashed to give a sine. Now it is true that if you momentarily discharged the cap in the triangle generator you would force the triangle to restart..... but that woud result in  truncated triangle (if the Ringer was set slow) or an extra piece of triange if it ran fast. It is what we call "hard sync" in the analog synth business.

I still say, make the drummer follow the tremolo or Ringer. It is true that in synchronised music setups, as a general rule you have the least sophisticated piece of timing gear as the master clock.. but this is an exception

[StephenGiles]

How about riding the fader up and down on your track in time with the beat on your recorded track?  Too low tech?

Absolutely brilliant - problem solved!!!!!
Stephen

[Mark Hammer]

Riding the fader is a reasonable thing to do, BUT it assumes the faders are up to the task.  Sadly, the physical dimensions and "slippiness" of many faders does not make them ideally suitable.

The Line 6 Tap Tremolo works very nicely.  The LFO waveshaping capabilities are in the same sort of zone as the (recent) Pulsar, and what appears to be a number of other current production tremolos (though these all vary in terms o the specifics, getting something OTHER than a stock triangle is a typical feature)  The tapping function works just fine.

Of course if one is trying to sync with a *human* beat, that may call for something very different than a tremoo rate set by someone going "and-a-one, and-a-two...".  My own inclination is to use some dependable percussive source as the basis for a trigger pulse, and use that to trigger an envelope generator of some kind, plus VCA.  The trigger extractor circuit that Craig Anderton had in Contemporary Keboard was designed for just such a purpose, though you can use the PAiA module that Jack Orman has posted at AMZ.

[Paul Perry (Frostwave)]

Mark, the problem there, is that you are in effect "hard syncing" the envelope. If the tempo changes significantly during the performance, then you would be wanting to change the attack/delay of the envelope.
For a FIXED tempo, it would work fine.
For a fixed tempo, you could just have a computer music prog  generating a click (for your drummer to follow) plus triggering a sample, and use the sample to drive an envelope follower. Or just generate a bar of fixed sine wave at say 10KHz, tweak the trem amplitude on it with a wav editor, then repeat for the length of the song, feeding an envelope generator.
If anyone wants to make a "very low friction" fader, you could take a LDR based voltage divider & waggle your finger(or toe, or...) in front of it.