The perpetual Encoders versus Pots’n‘Knobs Thread

As the question “Why don’t you uses Encoder instead of Pots/Buttons” pops up frequently here and there and every time the discussion converges to the same arguments i think its a good idea to have everything neatly sorted out here. You can then stop the discussion by referencing as everything that can be said will be said before, here.

Not lets start…… everybody is invited to express his personal believe, technical bits and fundamental world views as well as concrete examples…. every 100th poster gets a free original Victory Candy for paying shipping :wink:

For multipurpose menudiving & parameter changes: encoder. For dedicated functions, ie: cutoff & resonance, pots. For a good steak, cast iron skillet. Be sure to get it as hot as possible first. Sear both sides and finish in the oven @ 400f. Finish with a nice wine and shallot reduction butter.

Both can work great, but encoders are harder and more expensive to get right. The Killamix Mini and the Elektron gear are good examples.

Pots are probably a bit nicer for live tweaking of the same parameter during a performance. I’ve found that encoders work just as well for slower to medium speed sweeps as long as the synth or sound engine in use has good parameter interpolation.

I prefer encoders for programming and pots for performance, but only by a small margin.

I don’t really care as long as there is something to turn.

This thread came up because I asked if it might be possible to use encoders with Frank´s programmer where the Shruti XT uses +/- knobs.

@piscione I’m with you on this one :slight_smile:

@loopino

To answer you question: to have an acceptable user experience you need to scan a pot at about 15Hz or more.

An Encoder has 24 detents per revolution, skipping from one detent to another produces 2 pulses on the L and R line you need to track if you want to know in which direction the encoder its turned
If you are fast you can turn it in 2/10th of a second around 270° so it will produce

24 detents * 2 pulses/detent * 270°/360° / 0,2 s = 180 pulses/second you need to scan. On 2 Lines. If you want to exchange all the XT buttons by Encoders you will need do scan 6 * 180 = 1080 scans per second for the 6 encoders on 12 Lines, right.

Perfectly doable with a 20 MhZ ATMega, i just would suggest you skip calculating these Oscidingens (who wants these wobbly signals anyways??) because they eat up all the computing power thats need to scan the lines :wink:

My Basic question for all the Encoder Lovers:

What exactly annoys you with Pots? Is it mainly the fact that sometimes the Marking doesn’t complement the Value?

I see the advantage if you have a single Encoder to navigate thru Menus and set Values like on the Shruthi (and here i insist on detents) but if you have a dedicated Knob per Parameter an Encoder only makes sense to me if you

  • have enough computing power to make it feel better as an pot, so the resolution must rise when tempo falls, and i want to be able to sweep the whole range if i turn it fast in 270°
  • encoders that substitute Pots must not have Detents
  • for selecting a value from a list the encoder must have detents

All of these things you can find in the MicroWave XT, albeit the acceleration feature that allows fast end to end sweeps isn’t working to my taste… and i don’t care on other synths that the knob doesn’t match the value if i switch programs. I mostly program by ear, not by the indicator on a piece of plastic :wink:

@fcd72 The jumps annoy and distract me. Also, depending on the kind of music you often do want to see the current value and make precise adjustments.

@thijs


Works great on black Pot Knobs :wink:

And how do you see the value on a non marked Knob?
Personally i need the instant response when moving a knob, even if the value jumps, these snap features annoy me badly, every time it think something is broken and i see my selves powering up the soldering iron tracking down a pesky failure……

@fcd72 Value snapping on pots sucks, I agree. It’s the jumps I don’t like.

The point about seeing the current value has more to do with how parameter values are presented and applies to encoders as well as pots. Sadly, this is more often lacking when pots are used vs. when encoders are used.

For adjusting semi-discrete parameters such as beat-locked LFO speed, sample slice, note number, etc. I prefer encoders since in these cases jumps tend to be even more distracting.

One thing I would love to see on a pots-based interface would be two LEDs one on each side of each knob. The left one would be lit when the current value is lower than what the pot is set to, the right one would be lit when it is higher, and no LED would be lit when the pot position matches the current parameter value.

i don’t really like the jumping with pots, but i like feeling the boundaries of the range.
led ring encoders are a good idea for visual feedback, but i still perfer the tactile feedback of pots.
motor faders are another thing. i only have bcf2000 that uses them but there they work very well for me to even get rid of the jumping when switching settings/presets

i guess in a perfect world verything would have one motorized pot or fader per function
…and turn switches for lists (like synced speeds)

I think in all cases have encoders is better than pots, if you’ve ever used a Nord Lead 3 clavia you know that.
The problem is that the encoders are much more difficult to manage because it requires a much more frequent refresh.
For Shruthi, the problem is not whether it would be better with encoders, but knowing the cost of encoder versions, encoder price, adding an ATmega to manage its and development time for Olivier.
Maybe a Shruthi for 600 €?

I don´t think in all cases encoders are better than pots.

I like to have a physical end to turning without having to look so I know when I have reached Max value while looking at something else.

I am not necessarily making an argument that a pot is better, just that each has it´s place and use.

Most encoders suck because you have turn them multiple times to get some decent value change… Maybe encoders with a resolution of 256 steps per revolution exist but I haven’t seen them.

For the jumping with pots there is a great solution in ableton live for midi controllers. It’s called scale. It scales the remaining value per parameter to the remaining ‘turn’ available on the pot. It feels really natural and there is no jumping. It always makes your pot moments behave like expected.

I prefer encoders!

When i switch from a preset to another and have 20 pots that are in a wrong position “visually” and 20 possible jumps! it’s a bit frustrating,disturbing!
if you want to start from a preset/sound to change it a bit to try to find something even better, it’s just not possible with pots!

I also like to see visually at which percentage of the variation range i am for a parameter! And for this encoders need to have some visual feedback added to be as good as pots!

I just finished during these christmas holidays a project (started more than one year ago) that is quite in the topic! it’s a midi controller with 56 encoders.

I will try to post some pictures soon!

@ shiftr with a 24 pulses/turn encoder you have 96 steps possible/ turn. Not so bad !

In some ways, alot of this seems to come down to how you make sounds. If you use your eyes, the preference seems to be encoders but if your use your ears, pots are just fine.

I have been looking into these encoders with led rings around them.

They seem really nice, the feedback (visual, granted) is super cool, and can be contextual to the parameter being tweaked.

It would be great to have some type of “force feedback” on an encoder, that way you would get info without having to actually stare down on it.

To be honest, i like to have all set in one 270° turn. Machines have been built for decades without that encoder stuff and i can easily live without them, i learned to hate them on live gigs…