256 voice polyphony?? not as far as i can tell!

I don't understand the lagging you are talking about in registrations or live sets. It makes me wonder whether you are on an old OS. I have definitely never experienced this. Can you identify specific ones?

Bruce Lychee wrote:

I don't understand the lagging you are talking about in registrations or live sets. It makes me wonder whiter you are on an old OS. I have definitely never experienced this. Can you identify specific ones?

I'm so glad you posted that Bruce - thought I was going mad. I've just done 'the test'. 3 random PCM partials on each tone (didn't tweak any of them). 4 such tones to each of the lower and upper registrations, all in one live set. Then add Solo sound and perc layer - using PCMs.

Then play 1 note, then 2, 3, 5, 6 - and then a 7th - no difference - nor with the 8th or the 9th! Just keeps going... ?????!!!!!

Is it maybe a specific PCM that's involved for some folks?

I used PCM 1 for everything.

Bruce Lychee wrote:

I don't understand the lagging you are talking about in registrations or live sets. It makes me wonder whether you are on an old OS. I have definitely never experienced this. Can you identify specific ones?

ok bruce, registrations 01 A7 brass hold & sweep - 01 C6 brass money and 02 A1 jp brass attack are good examples i've come across in the last 5 mins or so of lagging AND note stealing in the same patch. try playing 5, 6 or 7 note chords in reasonably quick succession, with or without the sustain pedal and you'll see what i mean. i am on version 1.02, by the way, probably the same as everyone else.

doctorrock wrote:

Bruce Lychee wrote:

I don't understand the lagging you are talking about in registrations or live sets. It makes me wonder whether you are on an old OS. I have definitely never experienced this. Can you identify specific ones?

ok bruce, registrations 01 A7 brass hold & sweep - 01 C6 brass money and 02 A1 jp brass attack are good examples i've come across in the last 5 mins or so of lagging AND note stealing in the same patch. try playing 5, 6 or 7 note chords in reasonably quick succession, with or without the sustain pedal and you'll see what i mean. i am on version 1.02, by the way, probably the same as everyone else.

There has to be something wrong with your Jupiter. I checked those reg. items and there is no lag at all on those registrations. I tested it with rapid 2 octave scales. I also played 10 note chords with each note played in succession and could not detect any notes dropping out at all. I'm really curious about what is going on with your Jupiter because your version number is up to date.

I would get that checked out.

Edit: As a goof I just played toccata and fugue in d minor with those sounds. While the sweeps, pads and delays made things interesting, I didn't detect any lag, which would definitely be revealed by those hammer sections.

Bruce Lychee wrote:

doctorrock wrote:

Bruce Lychee wrote:

I don't understand the lagging you are talking about in registrations or live sets. It makes me wonder whether you are on an old OS. I have definitely never experienced this. Can you identify specific ones?

ok bruce, registrations 01 A7 brass hold & sweep - 01 C6 brass money and 02 A1 jp brass attack are good examples i've come across in the last 5 mins or so of lagging AND note stealing in the same patch. try playing 5, 6 or 7 note chords in reasonably quick succession, with or without the sustain pedal and you'll see what i mean. i am on version 1.02, by the way, probably the same as everyone else.

There has to be something wrong with your Jupiter. I checked those reg. items and there is no lag at all on those registrations. I tested it with rapid 2 octave scales. I also played 10 note chords with each note played in succession and could not detect any notes dropping out at all. I'm really curious about what is going on with your Jupiter because your version number is up to date.

I would get that checked out.

Edit: As a goof I just played toccata and fugue in d minor with those sounds. While the sweeps, pads and delays made things interesting, I didn't detect any lag, which would definitely be revealed by those hammer sections.

this is REALLY bugging me so i have attached am mp3 to show what is happening with my jupiter 80. here's what your (hopefully!) hearing. this is registration 01 A7 brass hold and sweep with all layers switched on but volume levels set at zero for all except the lower layer so in other words all layers are playing but you only hear live set 1042 saw brazz. first thing i do to illustrate voice stealing is play an ascending c maj7 chord starting on note c2 then adding e2, g2, b3, c3, e3, g3, b4 and c4 in sequence. it's pretty easy to hear notes dropping out near the bottom by the time of the b4 anyway but to prove it i then release notes sequentially from the c4 descending. by the time i'm back to c3 all you can hear is that note and the bottom c2 which i then release. all that is then remaining is one voice of the c3 although i'm still holding down the keys e2, g2 and b3 which are completely silent, their voices having been stolen. next you can hear me playing some 5 and 6 note chords in quick succession with very audible note stealing occurring along with (you'll have to trust me on this!) considerable lag, i.e. notes triggering late.

i would really appreciate somebody following exactly what i've done and trying to reproduce these symptoms as my gut feeling is that here's nothing wrong with my particular jupiter - rather that it's a problem with the voice architecture. of course i would be very happy to be proved wrong!

cheers
doctorrock

My Jupiter does the same thing. When I was whacking away at 1A7 last night I got the same gating effect. This is a huge 10 layer patch.

I've spent way too much time "testing" the Jupiter. My belief is Jupiter is simply fixed at 64 poly per section.

Ok. I will try this later today, as it is a little different than what I did. I did not try to isolate the lower set audio.

Devnor wrote:

My Jupiter does the same thing. When I was whacking away at 1A7 last night I got the same gating effect. This is a huge 10 layer patch.

I've spent way too much time "testing" the Jupiter. My belief is Jupiter is simply fixed at 64 poly per section.

i too have spent way too much time on this, devnor and i'm glad (in a way!) that i'm not the only one getting these results! i agree that it's pretty obvious each layer or part has a 64 voice poly limit, and this is the major problem with the voice architecture because it makes the 256 voice figure roland are quoting theoretical at best! indeed in my attempts to try and max the polyphony i only achieved around 170 voices as previously described. now that was merely an experiment and in no way constituted what i and, i would imagine, other users would call a musically meaningful scenario. what i WOULD call meaningful, however, is building a juicy fat live set (or indeed using a roland factory set/reg) and being able to play at least 5 and 6 note chords CLEANLY on it! and because of the 64 voice limit, this is impossible to do in some cases! in fact reg 01 A7 is a pretty good example of what i'm saying here because the solo part is restricted to playing only from c2 downwards, so for any time you're playing chords above c2 there is a whole 64 notes of polyphony that are sorely needed to fill out the chords but of course cannot be used because they are allocated to the solo part that is playing nothing at all!

i don't know how easy it would be for roland to rethink the voice architecture and do away with the 64 voice per layer limit but it would surely be a more practical not to mention musical way of doing it - voices would then be used as and when they're needed, which is, in my case at least, pretty often!

am i the only one who thinks this sucks hugely? i'm pretty much agreeing with an earlier poster on this thread who said that the synth is not working as advertised - and that is not something to lightly ignore.

cheers
doctorrock

Bruce Lychee wrote:

Ok. I will try this later today, as it is a little different than what I did. I did not try to isolate the lower set audio.

thanks bruce, i appreciate you doing this. by the way, playing 2 note octave runs isn't really a good test because any voice stealing and lag only happens when paying chords of 5 notes or more.

doctorrock wrote:

am i the only one who thinks this sucks hugely? i'm pretty much agreeing with an earlier poster on this thread who said that the synth is not working as advertised - and that is not something to lightly ignore.

cheers
doctorrock

My suggestion is to return Jupiter, buy a Kronos. You can find more info about Kronos here:

www.korgforums.com

@ doctorrock - did exactly what you describe and... yes, got the same result as you; stealing all over the place. Don't know where this puts us owners, but at least you know it's not just you!

@ Devnor - LOL!!!

The JP parameter guide states that the voice reserve parameter for the liveset layer only comes into play when over 128 voices are played simultaneously . Can't really understand how this makes sense, if it's a fixed 4*64 voice architecture?

doctorrock wrote:

Bruce Lychee wrote:

Ok. I will try this later today, as it is a little different than what I did. I did not try to isolate the lower set audio.

thanks bruce, i appreciate you doing this. by the way, playing 2 note octave runs isn't really a good test because any voice stealing and lag only happens when paying chords of 5 notes or more.

I was just testing lag with that. I did try a ten note chord and did not notice note stealing, however, I had all sets turned on so it's kind of difficult to hear what is going on. I will try with lower live set isolated.

I'm trying it....but I'm not hearing it!!!!