filter problem

Hi mate,
just tried it and mine is all in tune, pure filter, no Osc input with Keytrack knob to coarse tune then `Shift button` Keytrack knob for fine tuning.

Cheers,
Glyn.

Hang on a minute. Are you going through each yellow part 'on button' as each sound will be a different polysound. What are you doing exactly?

I think you seem to be confused about how poly stack works.
It is the one sound played as z 4 voice poly analog.
If you turn Stack On the Voice with both the Blue Light and the Yellow Light Below it becomes the master voice all the voices when played are replications of it. So if for example i have a lead sound on Analog Part 1 and it is the only selected part and then select Poly Stack on it become the same sound with four voice polyphony and not a combination of the different sounds on all 4 parts.
If for example I switch the button - Yellow to Analog Part 2 it now becomes the sole sound with 4 voice polyphony and its synth settings apply to all four voices.
There are no separate filter settings in a Poly Stack analog voice the others all slave to the Analog part selected in Blue Above and Yellow below.
If you were to have all 4 analog parts on RED and one selected - the one selected in Blue is the available voice to edit but all other voices will layer monophonically and are totally different independent synth voices in there own right.
Filter tracking works exactly as one would expect in Poly stack mode and on each individual voice in a mono layer and or when played multi-timbrally.

It doesn't sound to me that the OP is confused about poly stack given their issue is one of key-tracking for playing the self-oscillating filter.

They said they were "in poly stack" - so all they've have to do is press one note, hold it, then press another and so on to hear the filter cutoff for each voice in turn, and that cutoff would be for the same key-tracking amount (as it's poly stack).

No matter which part they had chosen as the poly stack master, there's no mechanism for the voices to have different cut-offs for the same key-tracking amount, other than for there to be an issue with the analog voices themselves (which are self-contained entities).

I just tested mine via this method:
init patch > poly stack > LPF1 > full res > full key follow > hold note and tune with filter knob
using a melda analyzer
repeatedly play note, then hold that note and move an octave higher and repeat until all 4 voices are playing.

The filter tracking is anything but reliable:

1st note/voice: I'm getting between 136.2Hz (closer to C#) and 147.3Hz (closer to D) on each keypress.
2nd note/voice: 269.6 and 295.5
3rd note/voice: 578 and 584
4th note/voice: 1160 and 1164

it's not fluctuating wildly while the key is held, it's "setting" on key press and changing with the next key press for that voice.

Thankfully this is less of an issue with the jd-xa than it could be, as we can get a sine wave on each of the 8 analog oscillators we have to play with.

Just checked again and I see what you mean, its stepping through the filters in turn. Maybe something for a future update!

thanks for the replies guys!

i don't think i'm confusing it :b not new to analog synths

no i'm not stacking 4 mono sounds. i'm in poly stack, where voice 2-4 slave to the first one (i'm only editing voice 1) for 4 voice poly (monotimbral). as i play a single key repeatedly, the jdxa plays through all 4 voices in order from 1-4. i get a slightly different filter behaviour from the voices 1&2 and 3&4. on 3&4 the filter is a bit more open.

if i bring the filter to self oscillation (with 100% keytrack) and turn down the oscs, the tracking over the whole keyboard isn't correct and get's out of tune the higher or lower i play the keys (didn't use keytrack fine tuning though!). but what's worse is that voices 3 and 4 always have a slightly higher pitch than 1 and 2 when i play one note repeatedly. impossible to play a nice 4voice chord without sounding like an out of tune mess.

thanks OneOfManyPauls for checking this out. so it seems that it's more or less normal behaviour but i still think my unit is acting a bit extreme.

sure it's not something i need to use all the time. and yeah there's sine wave (although the sine wave has more overtones than a tri wave. thats a little odd?)

i think i will send it back once again. after all it's a brand new machine and i paid a a few bucks so i expect a flawless machine.

I don't think key follow fine will make any difference given the fluctuations within a single voice each time a key is pressed with the settings not being touched inbetween.

although the sine wave has more overtones than a tri wave. thats a little odd?

yeah, the jd-xa analog sine wave is not a pure sine wave by any stretch of the imagination - it's kind of a slight morph between a sine and square wave.

I kind of like the quirkiness of the jd-xa though - keeps things interesting.

OneOfManyPauls wrote:

although the sine wave has more overtones than a tri wave. thats a little odd?

yeah, the jd-xa analog sine wave is not a pure sine wave by any stretch of the imagination - it's kind of a slight morph between a sine and square wave.

yeah no biggie. filtering helps. but what i really miss is the woody tone of the square from the old rolands :(

Just tested this myself.
Had to make sure there were zero filter mods going on elsewhere ie: LFo1 or 2, Via Velocity, Mod LFO and/or Wheel 1/2 and After touch.
The key tracking seems pretty solid of about a three octave range and only starts to drift in very low or very high registers on my unit. The drift isn't totally hideous either but i could play to the character on offer here as a strength providing some unique tones I couldn't get otherwise.
General i only use self oscillating filters on Mono parts and/or percussive tracks.
I still find it useful in this context though.

sorry for the bad quality. just recorded with my iphone from the speakers.
poly stack, no osc, low cutoff, full res, 100% key tracking. pressing c repeatedly.

same settings, playing always the same 4 note chord (cegb) repeatedly.

by the way this is all just about lpf1. i tested lpf2 today and it behaves like expected. tracks fine over the whole keyboard. more or less (had to fine tune key tracking +12). and the difference between the single voices is minimal. and lpf3...well...is weird like expected hehe

I listened to your repeated playing of one note and that's exactly how mine sounds in poly stack mode - 1 note as per the tuning for part 1, followed by 3 notes approx a semitone higher as it presumably cycles the voices.

With just part 1 in non-poly mode, it tracks pretty solid.

I copied the same part across all 4 analog parts and played them in turn in mono - part 1 is tracking a semi-tone lower than the rest for LPF1. As with your findings, it's fine for LPF2.

Weird fruit.

OneOfManyPauls wrote:

I listened to your repeated playing of one note and that's exactly how mine sounds in poly stack mode - 1 note as per the tuning for part 1, followed by 3 notes approx a semitone higher as it presumably cycles the voices.

With just part 1 in non-poly mode, it tracks pretty solid.

I copied the same part across all 4 analog parts and played them in turn in mono - part 1 is tracking a semi-tone lower than the rest for LPF1. As with your findings, it's fine for LPF2.

Weird fruit.

roland could make an update to implement a calibration routine (don't think so, but you never know)

anyways thanks for taking the time!