Tap tempo tremolo usefulness

[bioroids]

...or any modulation pedal that uses Tap Tempo to set its modulation speed.

I'm working on it (with AVRs) and it seems to work fine. But I find very hard to make the LFO to stay in sync for more than say, 10 or 20 cycles.
And I'm not sure if its because the programming is not sufficiently exact, or it's only a matter of imprecise taping.

I never had a Tap Tempo pedal, so I don't know how much exactitude this things have. I'm pretty sure it's humanly impossible to do a perfect taping, so the sync will be lost after a certain amount of cycles. But how many cycles you'd say it's a too poor synchronization, provided the taping was reasonable?

In my design, you tap three times to set the tempo, so it uses the two intervals averaged (I don't have an additional switch so the taping is made with the very bypass switch), maybe that's the initial flaw that prevents this to work better?

Well any considerations about this are welcomed. I intent to post the schem/code but first I have to have a version working reasonably well!

Regards!

Miguel

[R.G.]

Miquel, you have discovered why I think tap tempo is not all that useful on anything with a persistent modulation.

I've posted this before, but it's pleasing to see that someone else is finding it out too.

Tap tempo delays work fine. The time delay is not persistent, in that it always starts exactly in time with your note, and even with a number of repeats, it dies out. There is not time for it to drift out of sync.

For tremolos, phasers, etc. there is a persistent timing. When you tap your tempo, you expect that it will remain in sync forever. However, that is never true. Any error builds up and eventually it drifts out of sync.

There are a number of sources of error. There is the error of timing on when the switch makes and stops bouncing, which can amount to an uncertainty of 50mS or so per interval. It is *not possible* to time the switch much closer than that because switch bouncing makes it impossible to know the switch state unless you take 20-50mS to decide that the switch has changed. That doesn't sound like much, but if there's a 50mS uncertainty, it builds up to a half-second uncertainty in ten cycles.

There is the error of delay quantization, as well. A microcontroller can't determine *exactly* how long a cycle is, even if it is averaged over a long, long time. That's because the uC can only put out a time that is quantized to one part in 256 or one part in 1024. The error builds up over cycles.

Then there's the inaccuracy of the human foot stomping the switch. There's some wavering there, even with averaging. Two cycles is not enough for accuracy, but three or more is too many for a human not to feel slave to the machine.

That's needed - and what er... may be under design, somewhere, in a galaxy far, far away 8-) - is a tap delay that does the obvious, it resyncs itself to features in the signal. What a "feature" is and how this works is the subject of a patent application.

Personally, I think that unless you put the tap sensor under your drummer's snare drum, you're never going to be happy with tap tempo tremolo. It's always going to drift away, that being the nature of the beast.

The other option is to start it, and have the band sync to the tremolo - humans are far better at following a tempo that way than machines are, yet.

8-)

[earthtonesaudio]

what er... may be under design, somewhere, in a galaxy far, far away 8-) - is a tap delay that does the obvious, it resyncs itself to features in the signal. What a "feature" is and how this works is the subject of a patent application.

Ask and ye shall receive.
http://www.backline-eng.com/FX-Tracker.html

[R.G.]

Hey, looks like they got there first. 

I don't see the tap trem resync there, but it may be buried under the rest of it.

[Mark Hammer]

There strikes me as being one fundamental aspect that defines the usefulnes of tap tempo anything.  And that is the accuracy of the tempo input itself.  The accuracy of that input matters more in some circumstances (like rhythmic tremolo, or slapback echo) than others (long echo delays, and slow phaser sweeps).

At least part of the issue is the use of the human foot as the "input device".  The foot is great at getting the gist of things, but is not the most precision appendage of the several that humans possess.  Of course, if you used your finger, you'd have greater precision, but you would likely a) want to change from a stompswitch to a soft-touch momentary pushbutton and , b) you'd need to get the pedal up to where you could actually press that button with any convenience.

Another part of the issue is the extent to which circumstances allow you to plan your tapping so as to yield precision.  When the inter-beat interval is brief enough, you can rely on your internal clock to know when to tap again.  When the interval is much briefer than your musclulature can accommodate, you need a means of translating one tap into multiple cycles (which some pedals do nicely via digital conversion of single taps into triplets, etc.; see the two Empress pedals here for a really nice illustration of that feature: http://www.empresseffects.com/ ).  When the interval is long enough for your feet or hands to accommodate, but longer than your internal clock allows you to guess-timate accurately, you need either some sort of conversion algorithm that works opposite to the too-fast scenario (e.g., treat the next 8 taps as equivalent to one half-cycle), or else you need some assistance in where you are in the overall cycle so as to coordinate your next tap.

This latter challenge is something we discussed here several years ago during the final beta-testing of the Line 6  Tone Core pedals.  I found the preliminary Liqui-Flange tap-tempo feature to be relatively useless because a simple flashing LED that pulsed briefly at each end of the sweep cycle provided much less feedback than was needed to align one's tapping.  Several of us here discussed some display options.  What L6 eventually used was a clever adaptation of their existing hardware/chassis; they simply divided up a sweep cycle into gradually dimming and brightening steps such that you could see the LED change in a direction.  Not perfect, but it helped immensely, and did not require any changes to the pedal other than in firmware.

If you have any sense of rhythm, the solution provided by that FX-tracker device is perfect.  I assume it detects supra-threshold signal-related events and treats them as if they were something like synth trigger-pulses, measuring the time in between them, the same way tap tempo normally measures the interval between button presses.

If you can harness your guitar/bass/whatever to a trigger extractor, you should be able to do the same thing in mostly analog form.  That is, you use something like this: http://www.synthdiy.com/files/2004/PAIA_2720-11_Env_Follower.gif, and you harness it to another circuit that "counts" a fixed number of consecutive pulses upon request, and averages out the time between them.  So, a bit of chicka-chicka picking and you have a tempo established.

[bioroids]

I think you're right R.G. I read some of your earlier posts before getting into this, but I fully understand that now
I have a re-sync function on my pedal (it re-syncs everytime you leave bypass) that has provided very usefull on my "field tests".

I'll recheck again my debouncing code to minimize it's influence. I think a good approach would be to start the tempo count when you first detect the switch changing (bouncing or not) and use the debouncing just to confirm the switch has changed, or something like that).

About the quantization, that's a problem too. I think I got a good resolution for the speed (using 16bit variables), but that's something to take into consideration too.

Still, this thing is doomed from the beginning because of the accumulating error factor. Even a 1ms error will grow to 50ms in 50cycles. Say the expected period is 500ms, that means that in 25 seconds you are 10%off sync. If the expected period is 100ms, then you are 50% if 5 seconds!

A trick I'm using is to divide by four the tapped period. So you tap in whole notes and it gives you quarters, that reduces error in a factor of four I think. The downside is that you can't tap slow speeds (who wants them in a tremolo anyway? It would be different in a phaser though).

So for a really usefull tap tempo in modulation pedals, a facility to re-sync easily is mandatory.

Ask and ye shall receive.
http://www.backline-eng.com/FX-Tracker.html

That pedal looks very interesting! but I think it can be quite tricky to use the "sync from played notes". It will force you to play in time all the time! That's not how humans play, I think. And what happens if you want to play triplets or some stuff like that? How does the pedal know what notes take into consideration?

In any case I sure like to try it, maybe the guys at Backline Engineering have resolved those issues somehow.

Regards!

Miguel

[gez]

Bo Didley had the right idea.  Start the riff and scr*w everyone else (they can find the tempo and join in)!  

[iaresee]

Bo Didley had the right idea.  Start the riff and scr*w everyone else (they can find the tempo and join in)! 

That's actually not far off with what I was going to say: tap is useful depending on what you're using it for. If you're trying to tap to an existing rhythm or a click you've got the accuracy and drift problem. Heck, you've got this with a dial setting too. The only way to conquer this is to have a midi-sync'ed trem or a trem that shows you the precise tempo it is running at.

But if you're actually providing the rhythm guide with the tremolo part the tap is useful in that you don't have to bend over, remember a relative position on a dial, and dial it in. You just tap, play and everyone else should be following the tremolo pulse for their tempo, not the other way around. In this use case it's not hard to keep it all lined up.

Take the classic use of a tremolo: to cover Johnny Marr's intro riff on How Soon Is Now? (at least, that's always been the fuel for my tremolo desire). Since the song begins with this riff you can tap in the tempo and keep it all in sync by having the rest of the band play to your tremolo rate, not by adjusting your tremolo rate to match the band. YMMV depending on your rhythm sections proficiency with holding a steady beat of course.

[DougH]

Bo Didley had the right idea.  Start the riff and scr*w everyone else (they can find the tempo and join in)! 

That's what I used to do when I had a syncopated delay part before I had a tap tempo delay. I would start first, then the band would sync to me.

R.G.'s right- it works for delay because you are resync-ing the electronic device with every note you play. For continuous modulation you don't get that feedback correction and eventually the accumulated error will throw the sync off.

This is a pretty typical synchronization problem, really, and it comes up in recording too. Try taking two separate tracks that you think are the same tempo but were not actually sync-ed to a common clock (drum track, etc) when they were originally recorded. Usually, by the time you get to the end, the accumulated error will make them way off.

[snap]

All you need is a click-out metronomic signal for the drummers monitor?

[iaresee]

All you need is a click-out metronomic signal for the drummers monitor?

An excellent feature addition to a tap tempo modulation pedal for sure! If it put out a midi time signal you could build a flashing LED box for your drummer to look at. Akin to this little project.

[bioroids]

What about a tap button, something like a video game button that you can attach to your guitar and tap/resync with your fingers. It should have a wireless connection to the pedal so you can move freely.

Or maybe it can super impose a modulated frecuency outside the audible range that travels to the guitar cord and is extracted by the pedal. That sounds cool, but it has the limitation that you cannot run other pedals before the tempo'ed in the signal chain...

regards!

Miguel

[Mark Hammer]

I should think it would be quite feasible to run a pair of wires to a parallel soft-touch push-botton located in a small plastic chassis, waist height atop a mic stand.  It will most assuredly void your warranty, but if you are confident, you just have to install a mini phone jack on the pedal so you can plug in the extension tap tempo button.  The primary footswitch would still work.

[gez]

Personally, I don't understand why more modulation effects aren't MIDI controlled.  There are some that act as controllers, but few (any?) that act as slaves that can be manipulated by a drum machine/DAW operating on a laptop/whatever. 

In a live situation, it makes more sense to me to have everyone singing from the same hymn sheet; in other words, a click track to act as a master.  Many drummers use them anyway, so why not use it to control modulation "tempo"?

One of these days I'll dust down my MIDI textbook...

[mattpocket]

In my opinion, tap tempo is great. If I'm using delay/trem in 5 songs in our set, I dont want to have to bend down to change the delay time between songs. That just doesnt look cool. It's definately a tool worth having.

Matt

[R.G.]

In my opinion, tap tempo is great. If I'm using delay/trem in 5 songs in our set, I dont want to have to bend down to change the delay time between songs. That just doesnt look cool. It's definately a tool worth having.

It is - if you don't expect it to be keeping sync over long periods of time.

Over a whole song, the tremolo timing will "drift" compared to the song's tempo. If you aren't trying to keep those in sync, that's great and you'll be very happy- if all you want is to not have to bend over to set it.

[Morocotopo]

I always thought that mod FX´s that have just a variable speed sine as the mod signal are, well, not very musical, because that static signal is not a interesting thing, I mean, it´s BORING! Music is based on parameter change over time.
Also, if the speed of the mod signal is not synched to the music, it´s BOTHERING because it clashes.
So I usually set the speed very slow or very fast, so that the mind can´t relate it to the music´s pulse and it´s not perceived as another pulse.
Digital audio FX provides more options than analog, but manufacturers don´t seem to have taken (much) advantage of this. Tradition rearing it´s head?
A few FX use env controlled parameters, but i haven´t heard many of those (except auto wahs). i tried to build the Polyphase but never got it to work...
What does this have to do with Miguel´s post?
not really sure...

Morocotopo

[mattpocket]

It is - if you don't expect it to be keeping sync over long periods of time.

Over a whole song, the tremolo timing will "drift" compared to the song's tempo. If you aren't trying to keep those in sync, that's great and you'll be very happy- if all you want is to not have to bend over to set it.

Well, exactly. It suits my needs, and I dont see the issue with having to tap it a couple of times during each song to resync it.

[StephenGiles]

In my opinion, tap tempo is great. If I'm using delay/trem in 5 songs in our set, I dont want to have to bend down to change the delay time between songs. That just doesnt look cool. It's definately a tool worth having.

Matt

It doesn't look cool? really - if you want to look cool, take your shirt off!

Seems to be a lot of fartarsing about to me.

[mattpocket]

Believe me, taking my shirt off would not be a cool thing.