mulztiseq test results flaws and feature request

[Stubs42]

Hi, Was very happy to see you are working on v2. had to get the pre release immediately to check it out. Looks like the cpu consumption is much better than in v1 now. Thanks for addressing this.

The multiseq looks very promising but i found the following flaws:

When switching to volt mode it ranges between -10 and +10 volt and holds the output for the whole step. Double click sets it to 0 volt. After using the right clock menu to initialize the channel the range changes to be 0v to 10v and the voltage is output only for a short time like a trigger, dropping to 0 in between adjacent steps. After changing the mode to an other mode and changing back to volt mode its -10 to 10v again. I think that's a bug.

As a feature request I would propose two volt modes for 0-10V and -10 to 10v. Double clicking a step should initialize to the same value (0v) the step is initialized when using the right click menu to initialize the channel/pattern. It would also be cool if the default of a step in uni mode would toggle between 0 and 10v so double clicking on 0v will put it to 10v and double clicking on 10v would put it to 0 volt.

I know there i no space left on the panel but having a poly output for all 16 channels would be a cool add on too. At least in my patches there is a merge nearly after each voltseg.

update 1: Just checked with voltseq, also voltseq is initializing to different values in pattern mode. double click initializes to pattern 1, right click initialize channel initializes to pattern 32. Should be the same in my opinion (pattern 1) in both cases I think.

update 2: the rounding of values in pattern mode and for the cv input for patterns still seem to be different. This results in situations at the edge between patterns that the pattern number displayed in voltseq is different then the pattern beeing played on an other volt/trigseq the value is feed in for pattern. This should be fixed in my opinion. (My phrase module has a right click menu item to fix that 'Trowa pattern offset', which shouldn't be necessary)

Regards Dieter

[chichian]

Hi Dieter,

Thank you for testing it!

OK, yeah. I see that VOLT is starting at -10 V then initializing the step initializes it to 0V. I will have to go through and make it consistent. We would prefer it starts at 0 V also, but I don't remember if there was a reason why it started at -10V. (TBH, we made multiSeq module a while ago--back in v1 days--so I don't remember everything clearly.)

After using the right clock menu to initialize the channel the range changes to be 0v to 10v and the voltage is output only for a short time like a trigger, dropping to 0 in between adjacent steps.

Hmm, I am not seeing it ever do a triggering behavior though and the range for me is still -10 V to +10V. Wait, OK, I see it when I chose to initialize the entire pattern (not just the current edit channel). Yeah, "initialize the entire pattern" is resetting the channel to TRIG mode but it's not changing the knob and the screen display. Resetting it to TRIG mode is the intended behavior though (since all channels start at TRIG mode), but it should change the knob and the display on the screen... I will look into that. Definitely buggy...

I know there i no space left on the panel but having a poly output for all 16 channels would be a cool add on too. At least in my patches there is a merge nearly after each voltseg.

I think j4s0n wanted to add poly output too but yeah, he'll have to redesign the panel to fit that... Maybe in a later version.

update 1: Just checked with voltseq, also voltseq is initializing to different values in pattern mode. double click initializes to pattern 1, right click initialize channel initializes to pattern 32. Should be the same in my opinion (pattern 1) in both cases I think.

Basically voltSeq and multiSeq should be initialized to 0 V on the back end. What "0 V" means is different depending on the mode. For PATT mode, 0 V has been PATT 32. Ideally, it would initialize to PATT 1 (-10 V); I remember I had tried to make the "initialize" value different based on mode before (probably back in pre-v1) and it seemed more complicated than it seemed like it was worth. But we can revisit doing that when there is time.

update 2: the rounding of values in pattern mode and for the cv input for patterns still seem to be different. This results in situations at the edge between patterns that the pattern number displayed in voltseq is different then the pattern beeing played on an other volt/trigseq the value is feed in for pattern. This should be fixed in my opinion. (My phrase module has a right click menu item to fix that 'Trowa pattern offset', which shouldn't be necessary)

Hmm. I look into it. I thought they were all using the same code for V->pattern and pattern->V, but maybe not.