Cv Mod Scaling - Pre-lookup bug

Hi, I think there is a bug in the pre-lookup mode on cv scaling. setting this to 1.0, feeding a 1v signal should shift the signal up an octave but instead the scaling is very slight. Working much more accurately in the post-lookup, I guess some maths is getting in the way somewhere.

Hello! Sorry for the delay on this. :cold_sweat: It sounds like you are getting the expected behavior. So please allow me to explain:

A table lookup takes an index (0-99) and converts it to a voltage.

So “pre-lookup” means that we are still in the index space. So adding 1 means index+1, or 1 semitone if you are using the 12ET table.

And “post-lookup” means that we are now in voltage space. So adding 1 means adding 1V or an octave if you are using the voltage to represent pitch.

Oh I see, I’d got that pretty wrong then! :slight_smile:

Would it be possible to have a different mode, where 1v is 1 octave but then quantized to the nearest index value?

Hmm. Allow me to play a little bit of devil’s advocate. Why isn’t your external voltage already quantized?

Suppose you have an lfo waveform or a sample and hold signal, it could be constrained to quantized values.

Or perhaps the keyboard or offset signal isn’t as accurate as the indexed values in the 101/102

In this case, scale-degree quantization (like it is now) is actually preferable, no? For example, if you use a triangle LFO to sweep through a C diatonic scale using 1V/oct quantization then you will hear a rhythmic pattern of long-long-short-long-long-long-short. However, with scale-degree quantization you get a regular straight rhythm.

Quantizing an already quantized voltage will give you inconsistent results.

Perhaps, I don’t totally follow but I’ll frame this another way.

If I want a -5v/5v signal to shift up and down an index by 12 steps (typically an octave) what should the gain be for prelookup mode?

The gain should be 12! :smile_cat:

Oh, doh! That makes sense. Cheers :slight_smile:

One sec! I misread your question I thought you said a -1V to 1V signal. If you want 5V to mean 1 octave up then the gain would be 12/5 = 2.4.

Anyway, I think you get the point now.

Yeah that makes more sense, so 12.0 is 12 indexs per volt.

hold on…just to get this right. Is it possible with CV MOD to call for indexs with cutom voltages ?

Lets say i have gamalan scale with six notes, how can i use the 102 as a custom quantizer for just this scale ?

Guess i confuse things here…

…rec in A1 and A2 are doing this. CV MOD just adds to the index value. correct ?
would that make sense to add in the config file the option to access the index number via CV MOD ?

i am just realy start enjoying custom scales and so far its the only module i know which can handle this.

I’m not sure I’m following you. It is possible to set the config file to ‘quantise’ after the voltage has been added (ie snap the result to an index entry). Set it to post lookup.

Is that any help?

Or are you trying to achieve something else?

1 Like

wow… ! will try this. Thanks

1 Like

excellent, works as i imaged.

1 Like

Sort of a necro bump here, but I just want to confirm that what I’m experiencing is correct because I haven’t seen it explained in the way I’m understanding. (Or, as the manual says, it’s easier to do than to explain.)

I have the config.ini file set to pre-lookup. Let’s say I have the slope X on CV-A set to 1.0 and I’m feeding the CV X input 3.206V. No other slopes anywhere to make it simple. If my starting index on CV-A is 12, my understanding is that the index on CV-A with the input, should be 15.206. My question is, should the output for CV-A be the exact value of index 15, the exact value of index 16, or an interpolated value between the two? i.e. if we’re on 12et, should I see 1.d50, 1.e00, or ~1.d60?

I think I am getting the third option. This gets back to @iPassenger’s question about using an LFO. Should it still be sliding through values (as you would expect in post-lookup/“voltage land”) or should it be snapping to the closest index value as described by Brian with the “rhythmic pattern”? Right now I have sliding voltages when I think I should be getting a sweep through a scale.

Firmware 2.24, for the record.

It sounds like you are still in ‘post-lookup’ mode. Can you show me your CONFIG.INI file?

The reason why I ask is that this area of the CONFIG.INI has changed recently so it is now track specific:

[Track 1]

; CV Slope Modifier behavior
; pre-lookup - slope modifier is added to the CV index before table lookup.
; post-lookup - slope modifier is added to the CV after the table lookup.
cv-X-slope-mod = post-lookup
cv-Y-slope-mod = post-lookup
cv-Z-slope-mod = post-lookup

[Track 2]

[Track 3]

[Track 4]

Ah ha! It appears that I have an old config.ini file! Uploading for posterity, but it appears I should just delete it and start over. (2.4 KB)

EDIT: I got the updated config.ini and it is functioning as expected. Thank you for your help!

1 Like