I've held off on buying a XP-80 for about 6 months now while reading and posting messages addressing the timing problems that I have heard about. I am very eager (almost desperate) at this point to buy a new synth, but I need TRUE 64 polyphony with no malfunctions in other features. I also don't think I should have to buy multiple synths to get REAL 64 polyphony when one instrument claims to be able to.

For months now I've heard many Roland users describe the timing problems they have. At the same time, just as many users say they have absolutely no timing problems, even users that say they use very large complex arrangements with no timing problems. Why is that? Of course the words large and complex are relative, but there may be more to it.

I have noticed that many of the users who don't have the timing problems say that they only use software sequencers and almost never use the onboard sequencers on the XP series. After listening to the comments of many Roland users over time, I'm wondering if the timing problem actually originates around the Roland sequencer.

Roland XP specs state that the sequencer records at a resolution of 96 ticks per quarter note and will play back at 480 ticks per quarter note, a very large difference. That's 80% fewer ticks then the average sequencer which means more notes crammed onto fewer ticks around the beat that need to be processed simultaniously if recorded on a Roland sequencer. I can't find where the playback tick specs were shown, but I remember thinking when I saw it that 96 wasn't really that much and wondering why so great a difference between the number of record and playback ticks.

All experienced midi musicians know how over quantizing, in addition to making music sound too rigid, will cause many note events to be performed at once, placing great stress on the processor. I'm wondering if the fewer ticks of Roland sequencers causes this to happen also. The 500% more note events that happen simultaniously on Rolands, bogging down the processor, could be the cause of the timing problems. Those using Roland products to sequence and those using other products to sequence might be the difference in the timing problems.

As I said I am not a Roland user yet and I am very unfamiliar with operating a Roland JV/XP, so this brings up many other questions. Is there a way to select what tick resolution a user wants to use on a Roland? Can any Roland users tell of their experience regarding the points I have mentioned and their method and equipment for sequencing, and whether they have timing problems or not.

I have also heard users in the past talk about conflicts between the software sequencer clock and the onboard clock which caused timing problems. When both clocks were being used, timing problems occured. It sounded like both clocks were fighting each other. Once the Roland was adjusted to use the software sequencer clock (computer clock), the timing problem was alleviated. Can any Roland user who has timing problems try this and see if it corrects your problem?

Hopefully, if we all put our heads together, we can fix these problems with Roland products, bill Roland for our efforts (yeah right), and get back to making music.

I have read several comments about the polyphony of the XP's. The 64-voice polyphony is not a Roland lie, it is completely correct. It is important to remember that this refers to voices, or tones, and not necessarily notes. If a patch uses 4 tones/voices, the synth can only play 16 notes simultanously, and if the patch uses 2 tones you get 32 notes polyphony. This is the way all PCM-synths and -keyboards are.

If you want a synth that has constant polyphony, your only choice is Kurzweil K2000 or K2500, but these have only 24-note and 48-note polyphony respectively. These synths use "virtual" oscillators to generate 3 of the 4 tones in the patch, i believe. Besides, the K2500 costs a fortune.... The other alternative is to buy a 256-voice synth, if you can get one (Some E-mu module, perhaps?). This should always have at least 64 NOTES of polyphony.

The suquencer resolution appears to be 96 ppqn per track. This gives enough resolution for most situations, if handled with care. The problem is, that if the sequencer sends a LOT of MIDI messages, the processor of the synth itself may not be able to generate the sounds quickly enough. This gets really imortant if you use a lot of MIDI-commands that consumes a lot of processing power, such as SYSEX or perhaps Control Changes. What I am saying, is that the sequencer may not always be the limiting factor when it comes to timing.

Besides, the synth still has maximum 64 note polyphony and 64 voices, even when playing back from the sequencer. If you use 2, 3 or 4 voice patches in your sequenced song, the synth engine itself will get overworked long before the sequencer.

I hope this may be a little informative

Stig