I did not make this video but I have wondered if it was just my unit.
EDIT: this is by design, as per the comments by @ess and @olle in this thread.
The range of the LFO on the digitakt is roughly five times more than reasonable - to the extent it saturates the pitch modulation input. I hope this gets fixed.
Bug been there since I got the machine on 1.05, still here on 1.11Measurements done using CueMixFX FFT visualiser
The tuning is limited to +/- 24, has nothing to do with the LFO
This! Also, the values are are not relative to the destinations that they modulate. Nothing strange at all. Compare it to CV signals in a modular synth, outputs and inputs are independent of eachother. You can also add/substract modulations from more than one source like p-locks, velocity, modwheel e.t.c
Then why are the values ± 64? that’s confusing UI then. /shrug.
MIDI is 127 values. If you want a bi-polar value it’s 63 one way and 64 one way. Same for the filter cutoff that is 127 values and not in Hz.
First: I love my Elektron Boxes, and would not trade them for anything in the world. I love what you guys do and trust the decisions you guys make.
Again, I trust you guys but I too, also found this behavior odd at first.
Yes you are correct it’s 128 values 
The hi-res values are solved in the midi standard by one CC being MSB and another CC being LSB. One sets the whole numbers and the other one sets the decimal.
Which makes FM using the LFO, limited…not complaining, it’s a sampler
No, the LFO is not broken at all.
If you modulate a destination to a point where you’re out of range, it will naturally clip.
This could happen to the filter frequency too if you have set it at a standard value that is high.
This could also happen on an all-analog system*, it’s nothing weird or wrong, it’s just how modulation works and is an expected caveat.
(* For example, you have an LFO that goes from -1V to +1V, and you have an oscillator which tracks from -0.5V to +0.5V … You will need to scale the output of the LFO accordingly. Luckily, on the Digitakt it is easy, just dial back on the depth.)
Just to clarify - again, that has nothing to do with the LFO, only with the sample engine. The pitch limitation lies within the engine itself. (Which is due to using high quality sinc interpolation)
I feel this might need a further explanation too.
The standard MIDI range is indeed 0 - 127, 128 values in total.
But you can use a MIDI standard called MSB (Most Significant Bit) and LSB (Least Significant Bit) to extend the amount of values.
This works by using one message that controls for example the coarse resolution (127) and one that controls the fine resolution (.00), both forming 127.00
Effectively, this can extend the range from 128 values to 128*128 … 16384 values.
…
And as for why the LFO is then -63.00 to 64.00? This keeps the coarse range within a standard MIDI CC message (0-127), but can be fine tuned using the MSB/LSB method. It allows for both standard and extended range control.
without muddying the waters further - it ought to be Most/Least Significant Byte (i.e. 1 byte is 8bits or 7 useful bits for midi) thus combining two 7bit params gives 214 possible values (16384) - it’s also probably -63.99 (aot -63.00) or similar on the Digitakt if it has the 14 bit equivalent range
I know it’d be crazy but i wish these ranges were expressed accurately - there’s rounding errors between devices and plugins in my experience - presenting the detail in HEX over two digits would be unpopular, but accurate if you want to get precise values - as it stands you have to do this in your own editor - basically i’m not a fan of 2 decimal point approximation of a 7bit range (128) - i doubt it will ever change, but it’d be cool if the exact values could be read
such as two stacked 0-127s whilst editing (as opposed to a dial on e.g. A4), thusly
127 MSB
127 LSB
one can dream
it’s quite a nice system to use a pair of CCs (if you have CC slots spare) as not too many controllers are great for stringing NRPN messages up where the MSB and LSB parts are packaged into their own message - necessary on A4 as it has so many parameters to map
I do understand, just something to keep in mind if attempting to achieve FM synth like sounds out of a sampler.
Hi,
The manual states that the LFO depth is a high-resolution parameter, with CC LSB value, but it doesn’t say the bits! So, how many bits has the CC LSB on the Digitakt?
The CC parameters are the following:
Thanks