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

I read somewhere that each part in the JP-80 has a dedicated processor (or core).

Which probably explains why the polyphony is divided into 4 x 64, if that really is the case.

I don't think that 64 voice is a limit per part. It's the maximum possible number of voices to reserve with the (very useful) voice reserve function. But as stated earlier, there were already 72 voices achieved before notes started dropping off.
It would be interesting to see on the other side, how many voices are possible with the SN acoustic tones. For that part, one should select for example 4 pianos and put them into one live set. On the system level the velocity should be set to a constant value, 100 or 127 for example.
Now, you could try to play a chord by adding one by one note, hold the sustain pedal and see how far it goes. If there is a 64 voice limitation, you should hear notes cutting off after the 8th note.
In this example, 4 pianos use 8 voices since they are stereo. It would also be interesting to see if there is a difference whether one uses effects or not.

Ok. I just had a chance to retest isolating the lower part. Still no lag. I isolated each section to test that and I really can't detect any lag. I'm not sure what is going on for you there. Try isolating each part and see if you still hear it.

I ran your test and did notice the note drops on the way down but couldn't hear it throughout the mess on the way up. I played 10 note chords and on the way down the tone remained the same when I go to the thumb on my left hand.

There seem to be 24 partials making up the sounds on 01 A-7. I just made a full Reg item with all SN acoustic tones and could not replicate the note dropping.

On a practical note, if you layer 24 partials, you probably won't be playing more than one or two notes at a time. A-7 sounded like a mess to me after 2 notes.

and I think that's the main point, Bruce. Is this discussion even germane in the real world? How many voices are you going to use to reach this theoretical limit in a musical context? All I hear is MUSH at this point, that I would never put into a mix or venue. For any practical musical application.....the JP80 has far more than enough headroom to create all the GARBAGE you can use. Its enough rope to hang yourself with! The truth is....in a musical context...the polyphony here is beyond anything you need. If you just forget about it and make some quality music...it will never be a problem. The ONLY situation I can see where it could present itself as a problem is in multi-timbral sequencing. But then, JP80 is only 4 part multi-timbral.....and I don't think its the application most people are worried or thinking about. JP80 shines at layering and expressiveness. Over-do it...which you pretty much have to do to reach these limits....and all of that expressiveness and sonic bliss is gone anyways. My opinion of course......

RonF wrote:

and I think that's the main point, Bruce. Is this discussion even germane in the real world? How many voices are you going to use to reach this theoretical limit in a musical context? All I hear is MUSH at this point, that I would never put into a mix or venue. For any practical musical application.....the JP80 has far more than enough headroom to create all the GARBAGE you can use. Its enough rope to hang yourself with! The truth is....in a musical context...the polyphony here is beyond anything you need. If you just forget about it and make some quality music...it will never be a problem. The ONLY situation I can see where it could present itself as a problem is in multi-timbral sequencing. But then, JP80 is only 4 part multi-timbral.....and I don't think its the application most people are worried or thinking about. JP80 shines at layering and expressiveness. Over-do it...which you pretty much have to do to reach these limits....and all of that expressiveness and sonic bliss is gone anyways. My opinion of course......

Totally agree. I would never have realized this unless I ran the test. I can't imagine any application under which I would come across an issue with the Jupiter's polyphony.

The test results make me think the Jupiter might be 128 for lower and perc. and 128 for upper and solo, which would be fine with me.

RonF wrote:

and I think that's the main point, Bruce. Is this discussion even germane in the real world? How many voices are you going to use to reach this theoretical limit in a musical context? All I hear is MUSH at this point, that I would never put into a mix or venue. For any practical musical application.....the JP80 has far more than enough headroom to create all the GARBAGE you can use. Its enough rope to hang yourself with! The truth is....in a musical context...the polyphony here is beyond anything you need.

the whole point of me spending all this time testing the jupiter was not because i had some random urge to verify roland's polyphony spec - it was to investigate all the voice stealing and missed note triggers that i could clearly hear going on while i was playing factory patches on the damn thing! that's a real world musical context if you ask me. so no, the polyphony here is not necessarily beyond anything i need if i want to play some 5 note chords on some sounds. other sounds no problem, percussive ones, pianos, vibes for example but anything sustained like brass, strings or synth pads i sometimes have these problems with. have a listen to my mp3 and tell me that's not a practical musical application!


The test results make me think the Jupiter might be 128 for lower and perc. and 128 for upper and solo, which would be fine with me.[/quote]

it would probably be fine with me too bruce if that were the case! i'm not sure how you come to that conclusion though after 3 other people have reported the same note stealing going on and you yourself said you could hear notes dropping out when doing my test. i am absolutely convinced that the assumption of a 64 voice per layer limit is correct especially after having done the same test with the same result using supernatural acoustic sounds - flute, trombone, alto sax and others. although i agree that this appears strangely at odds with the manual saying that voice stealing limit only comes into play after 128 voices are played.

anyway i'm not going to spend any more time worrying about this, for all my protestations i'm loving making music with the jupiter most of the time and that's the main thing.

I took a listen to your mp3. That is the sound of classic note stealing. You can hear the release envelopes being cut off. In addition, you can clearly hear the older notes being stolen due to polyphony limits.

To be fair, every synthesizer has this problem. The most unnoticeable is Kurzweil. The Kronos is pretty good, but maybe not as good as Kurzweil. In any case, we just get, what we get. We just layer less.

I have never heard any note stealing on the JP80. I don't need to do any tests. It just doesn't happen to me. And I would say that I am more than a casual user....I program and go pretty deep with the JP80. In fact, the ability to do so is what intrigues me about the JP80, so I exploit it!

I don't doubt you are hearing what you are hearing.....I just doubt you (or we) fully understand why you are hearing it. In the meantime, I have some SICK layers I have made on the JP80, including loading up the entire board (just for fun) with detuned saw waves to make the ultimate ridiculous super saw (so to speak) and I hear no note stealing. On the contrary, its smooth as butter (if not a bit ridiculous, LOL). Again...I don't doubt your experience, which under certain tolerances (the criteria of the test you have developed) results in what you are hearing. I know you are just reporting the facts as you find them. I am just telling you my experience which apparently conflicts with your results. So....what I think we can say is....Neither one of us knows what is going on here. I'm not too worried about it though, because the JP80 has more headroom than any hardware synth I have ever programmed. Certainly more than I have ever needed....its just a non-issue...for me.

RonF wrote:

I don't need to do any tests.

I have a bit of a problem with my keyboard setup:

the stands' legs seem to be built for a slightly curved Earth surface, and based on some kind of gravitational force,

but I believe the Earth to be flat.

Anybody knows where could I find a Flat-Earth keyboard stand?

I just did this test with no problem , I heard your mp3 file , maybe try not to bang on the keys like a kindergartner .

The Jp-80 is smarter than you think, it know that what you are doing has no true musical application .
Supernatural is always trying to come out and be expressive , check your attack levels on all levels .
Did you call roland for help maybe your unit is faulty .

I only see this problem happening when you sequence all parts of the jp-80 at the same time and keep layering .

RonF wrote:

I have never heard any note stealing on the JP80. I don't need to do any tests. It just doesn't happen to me.

respectfully ron, i might suggest that you may hear the note stealing going on better if you bothered to take your head from the sand where it would appear to be firmly buried ;-) (although comparing you to a flat-earther maybe going a little far!)

RonF wrote:

I don't doubt you are hearing what you are hearing.....I just doubt you (or we) fully understand why you are hearing it.

well, you can speak for yourself! and if you listen to my mp3 you can hear it for yourself, as others here have done and acknowledged the result. and maybe you should reread my original post because "loading up the entire board with saw waves" was exactly what i did for my first test! so maybe we should just agree to disagree......

Rocness wrote:

I heard your mp3 file , maybe try not to bang on the keys like a kindergartner .

doh! thanks for the advice - i KNEW i must've been doing something wrong!

Rocness wrote:

The Jp-80 is smarter than you think, it know that what you are doing has no true musical application .

wow, scary comment on every level! although that may have been a clumsy attempt at humour??

Rocness wrote:

Did you call roland for help maybe your unit is faulty .

no, i tested another unit with exactly the same results. read the thread - other users here have had the same experience as me. 2 suggestions for you. firstly, if you have the time or the inclination, redo my test EXACTLY as i did - then play some five or 6 note chords in reasonably quick succession. you will then, i'm certain, conclude that despite your undoubtedly well-refined keyboard technique, the jupiter 80 has done a fine job of making YOU sound like a musically incompetent 3 year old too! and secondly, never be quick to presume.....

I'm going to side with the doctor on this one. I hear a weird gating effect sometimes. Maybe its just a combination of inputs at the right time freaking out the behavioral modelling? We're still beta testers. Registrations like 1A7 are somewhat useless to me because I don't exclusively work with massive synth layers. There needs to be room for other instruments too. For what I need Jupiter is an expressive machine & beyond capable.

Okay, maybe there's a break in polyphony. So what?

The mighty - and I would say market leading - strength of the JP-80 is it's built in capability to be expressive, like a true acoustic instrument.

I have just been composing a piece which is very like a second, slow movement of a piano concerto.

First, I played the piano on the OASYS, with strings from the JP-80.

Then I swapped them.

O.M.G. The piece took on a whole new dynamic dimension that was enthralling, exciting, passionate, immediate, lush and wonderfully beautiful.

Therein lies the beauty of the JP-80 (for me) - for the first time ever I have a true electronic 'instrument' that, even compared to the power of the OASYS, is a sublime work of art.

Polyphony? Not an issue.

I can replicate the note dropping but not the lag. I have tried it several times with different loaded Reg items and I just can't get the lag. I am playing 7 - 10 note chords as quickly as possible.

This is a non issue for me because I don't see any practical application where note stealing would come into play. A Reg item like a7 is pretty useless with anything more than 1 or 2 notes played at a time.

I will note that when the 700NX came out users reported strange note stealing behavior which was subsequently fixed via update. I wouldn't call this case strange, but without knowing the exact distribution of the polyphony, it might be that there is something going on at the software level that needs to be addressed. If it is causing issues for you, perhaps an email to Roland might be in order.

cello wrote:

Okay, maybe there's a break in polyphony. So what?

The mighty - and I would say market leading - strength of the JP-80 is it's built in capability to be expressive, like a true acoustic instrument.

I have just been composing a piece which is very like a second, slow movement of a piano concerto.

First, I played the piano on the OASYS, with strings from the JP-80.

Then I swapped them.

O.M.G. The piece took on a whole new dynamic dimension that was enthralling, exciting, passionate, immediate, lush and wonderfully beautiful.

Therein lies the beauty of the JP-80 (for me) - for the first time ever I have a true electronic 'instrument' that, even compared to the power of the OASYS, is a sublime work of art.

Polyphony? Not an issue.

Since I bought the Jupiter 2 months ago I have been singing the praises of its uniquely expressive capabilities. It is so different than anything else out there that you won't really understand it until you spend some time playing it. If you just walk up and hit notes to hear sounds, you will the be missing the picture entirely. It is definitely made for the keyboard player. It gives you the ability to express things in real time, which makes the creative process so much more natural and enjoyable, yielding results that reflect sonic quality and genuine musicality.