Okay I have the *DEFINITIVE* word on this from Roland UK tech support.

*All* JV and *all* XPs have the same problem.
Its "because there's no gate on the sample" - whatever that means.
It is a design feature
And it is nothing to do with any A/D converters
There is absolutely no solution to this.

Oh dear.

The only good thing is that any expansion cards for the XP-60
etc can be used on any of the XV range of modules.
And NO they do not have this problem.
When asked "why not" he sounded unsure about this,
but said it was because there was "more memory" in the XVs..


Oh dear.
Well, I'm *really* unsure what to do at this point. I have just bought an XP-60.
I could try and send
the XP-60 back but I want to be able to lay down basic sequences
of music without a computer. And I need high quality classical
/acoustic music instrument sounds. And I dont think anything else
will be within my budget (c.GBP 1100)

Alec

OK, flame me but I think this is all a fuss over nothing.

How big a problem is this distortion - really? Is it really and truly noticeable to anyone other than those tying ever so hard to hear it?

I love my XP-60 - so do tens of thousands of 60 - 80 owners around the world.

In my (amateurish and ill-informed) opinion this is a theoretical problem, not an actual one.

My ten bob's worth, anyway.

Martin

Yeah, really; all synths are apparently not created equal, since my XP60 has never exhibited this phenomenon. This whole thread has always puzzled me. And FWIW: I think the Roland UK techie's "no gating" explanation is a bit spurious.

I posted this message on this board about a year ago..

"I have just done some "research" on the much discussed decay noise/distortion on the XP/JV synths, and I thought I should publish my findings. I own an XP-60 that I used for these tests.
As the sound on the XP-60 decays, a slight noise/distortion may appear. Some users have blamed this on the D/A-converters, others on the synthesis engine itself.

I used a piano-patch that I know has this noise on the decay. In tone 4 of the patch, I included a sine waveform and assigned it to the lower part of the keyboard. I limited the piano to the upper part, so that I got a split. I adjusted the sine so that it was in a different part of the frequency range than both the piano and its decay noise. (A sine only contains one frequency, so that it is easy to separate from the decay noise and the piano). I turned the tone volume of the sine down really low, as I have to turn up the volume of my amplifier to hear the decay noise at all.

I then tried to play a note with the piano patch, and let it decay. At the same time I increased the volume of my amp towards max. As the decay noise/distortion appeared, I tried to tap a key on the lower part of the keyboard, assigned to the sine waveform. With brief taps on the key with the sine, I could clearly hear the decay noise disappear! It reappeared the moment I released the sine.

From this I conclude:
The decay noise is not in any way related to the synthesis engine itself. The noise/distortion appears when the TOTAL signal-level of the output drops below a certain level. I have heard users complain about hearing the decay-noise all the time while playing e.g. sequenced songs. This must be a problem of their hearing, or the placebo effect!

The decay noise does NOT appear at the end of every note played, and is not really related to decay (in the synthesizer meaning of the word) at all. It is only related to the outputs. This means that in a typical sequenced song, the decay noise will appear for the last 2-3 seconds of the song, when the total level is dying!! Other than this, the outputs will mostly perform perfectly. Playing in patch-mode (with percussive instruments), the noise may appear slightly more often.

Decay noise of this kind seems to be quite typical for digital instruments, exactly the same symptoms were discovered in a Technics KN-920 keyboard. Only worse...

It may not even be the D/A-converters that are to blame. Perhaps there is something in the analog part of the output?

Some users have claimed that this very minor fault in the XP/JV-synths make them unusable for professional purposes. Well, if it's pro enough for Jean-Michel Jarre (XP-80 on Metamorphoses), I must admit that it is more than pro enough for me!!! But perhaps we have some REAL pros on this message board? "

End of (my own) quote..

Hope this information was useful.....

Stig

dnarkosis

Are you absolutely *sure* !?
Have you tried listening to a
e.g. Piano patch on decent headphones
and played quietly and with
the volume turned right up ?

And then listen to a slow decaying note
right to the end ?

No fuzzing, no fizzing - not AT ALL ?!!

If so when & where did you buy your XP-60 ?!

Maybe they've fixed it in some XP-60s but
wont actually admit it !?!


Alec



[This message has been edited by stewart33 (edited 03-17-2001).]

stigf

Interesting. Are you saying that this effect
only happens when the *total* volume of output approaches zero ?

So that if for example I had a very very quiet note running through - a bit like a background hum - then we wouldnt experience this effect ?

Either way I'm not sure what to conclude !?!
I shall go back and test my XP-60 and see if
it's the same effect on every single key on a piano patch.

But I'm particularly interested in
dnarkosis's post saying that his machine is fine. Do we believe him ? Because if it is fine, then I can shout like bl**dy h*ll at Roland, (though if it isnt I'm in trouble!)

A

The problem is the cheap d/a convertors. Roland will do nothing about it.

The problem effects all xp60, xp80,xp30, jv 1010 and jv2080.

Listen with headphones at high volume, with effects on in quiet environment, especially on accoustic guitar and piano patches.

stewart33:

That is exactly what I am saying.. Try it for yourself..

If you had a background hum or something, that stopped the sound from dropping below a certain level, then the problem would not be there at all.. Which should, as I said, that the problem is NOT in the samples, but in the output...

And people:
This may sound rude, but let's try to listen too hard to Arvon45 in this matter. He has shouted about this problem for 2 years on ths board, and has come up with nothing constructive. I think he sold his XP years ago, and is now a Korgian...

Stig

By the way..

My XP DOES have this problem, but i find it so minor, that I don't care. Most other noise-sources are far bigger..

If this has been a problem with all XPs/JVs, then why was it discovered only a couple of years ago?!?!

Maybe because it is so minor...

Stig

Okay, I checked. Nothing. I listened to all the piano sounds in Preset A and a couple of slow-decay pads. Nothing. I then got my wife to listen; she hears high frequencies much better than I do. I asked her if she heard anything during the decay that didn't belong to the patch, like white noise or a reverb tail that seemed bogus. Nothing. Two different headphones (not $300 phones, but the same brand I've been using forever).

Man, I just don't know what to say. I'd *like* confirm this anomaly, but it's just not there. I have no reason to to jive you or jive myself, and my wife certainly couldn't care less one way or the other.

Really, I have no clue. I've also recorded my XP digitally, and the engineer didn't say anything (it was a multitimbral piece, but still I would guess something would show up).

I'm not doubting you guys are experiencing this, it just puzzles me that I am not.

I've had my XP-80 for almost a year, long before I knew about this board or this discussion. I've noticed the distorted buzz a bit, but it isn't really a problem, and here's why. The sounds on the keyboard, example, acoustic piano, aren't good enough to be used solo. I wouldn't want to do a solo piano album on the XP. Once you've got guitars, drums, bass, etc, there is no way on earth you're going to hear this problem. None!

Hey all. You know, I was sort of upset when I first about this situation. Upon turning up the volume on my sytth, I noticed this right away. But then again, I had a Roland XP50 before this. I never heard that problem there and am aware that this doesn't affect owners of XP50s.

Also, after I recognized it, that was it. It never bothered me again. I write biug, huge arrangements. There is always so much going on that I never even recognize the problem anymore.

If I had to spend the money to buy another XP60, I would do it again (God forbid - I'm a college student and do not want to have to go through saving up again). As Jimmy Jam and Terry Lewis said themselves, the Roland XP60 is one heck of a keyboard.

The Infamous EPU.

Epu:

If my theory is correct, you will never hear the distortion in an arrangement, because it will not be there, not even theoretically..

To me, this fact reduces the problem to almost nothing. I can live with a very minor distortion at the end of each song.

Stig

Til deg du visne lyng om haug
der draumar sviv,
eg gjev min song om dimd og draug
og dulde liv!

Here is a useful test:

Go to PR-A and choose the patch "House Piano."

Play more than one note and hold the notes down.

Since the patch is programmed (as you'll see if you check the EFX Parameters) with MASSIVE compression, it will highlight and reveal this "decay noise" people have been worrying about.

It's been there in my XP-80 since day one, and I've never minded it.

The DACs on the XP-80 are, according to the service manual I have, high-quality Burr-Brown converters. You can buy a service manual for your synth directly from Roland and confirm this for yourself, with a little research. So this shouldn't be a problem; if it were the DACs, they'd have to be technology two generations behind the chip in the synth to even be theoretically capable of being responsible for the "problem" people are talking about.

You'll note that if you "gate" the "House Piano" patch, by reducing the decay using the sliders, by the _tiniest amount_, i.e. enough to affect the decay of the sound, but not enough to cut it out entirely, that this will reduce the sound significantly.

This is what I think the Roland UK office is probably referring to as "gating the sample."

Playing around with the programming of this patch will give you a good picture as to what the sound is, where it's coming from, _why_ it's there (it's got actual use in this patch, to give it a gritty, lo-fi quality appropriate for house music, so it must not be an accident that it's there), and so on.

I consider this "problem" an actual intentional part of the JV/XP sound. We all know that pure, clean digital can sound sterile. Introduction of a little noise can actually provide some "air" to the sound that would otherwise not be there, if it is handled in recording appropriately.

Steve

\\Playing around with the programming of this patch will give you a good picture as to what the sound is, where it's coming from, _why_ it's there (it's got actual use in this patch, to give it a gritty, lo-fi quality appropriate for house music, so it must not be an accident that it's there), and so on.\\


This isnt the grunge decay issue.


\\I consider this "problem" an actual intentional part of the JV/XP sound. We all know that pure, clean digital can sound sterile. Introduction of a little noise can actually provide some "air" to the sound that would otherwise not be there, if it is handled in recording appropriately.\\


Unwanted noise at the end of a sample that sounds like a dying toaster is not "flavor", its a *flaw*.

Well, this is the _only_ "grunge decay" "problem" I have on my XP-80, so I guess I must be one of the lucky ones!

Steve

BTW, one last piece of information, from a BBS back in 1997, when someone asked about the XP-80 DACs. More recently the Burr-Brown PCM69au's been replaced by the PCM1800, which are 20-bit. Same logic applies, though. Here you go:

From: WeAreAs1@aol.com
Date: Mon, 6 Oct 1997 20:44:56 -0400 (EDT)
Subject: Re: Re: [jv1080] XP-80 D/A convertors

The JV2080 does not have 20-bit D to A converters. It uses Burr-Brown
PCM69au converters (the same ones used in the XP-80 and several other
models), and these have 18-bit resolution.

It is important to understand, however, that the JV/XP series all have 16-bit
*data resolution*. Running this data through an 18-bit converter does not in
any way increase the resolution of the data. The additional two bits will
simply remain unused.

There is a slight sonic improvement when running 16-bit data through 18-bit
converters. This improvement occurs at the very lowest signal amplitude
ranges, such as decaying note fade-outs and reverb tails. The additional
bits will effectively correct for possible amplitude quantization errors for
signals that are at the lowest possible amplitude or are approaching zero
amplitude. Whether this effect is actually audible in the real world is up
to question.

Anyway, don't get all worked up about 18-bit this and 20-bit that. The JV/XP
is a 16-bit sample playback machine, and no amount of additional resolution
at the converter will be able to change that.

BTW - in answer to the previous question about the conversion of the effect
proccessor data: All signals in the JV/XP are digitally mixed within the
synth's tone geberator and DSP chips before going to the D-A converters.
There is only one generation of D/A conversion.

StevenA:

Let me begin by saying that I am not an expert in digital signal processing, D/A-converters etc.

The XP-series synthesizers have 16-bit samples, as do all current PCM-synths. It is however important to remember that much of the signal-processing is done at higher resolution. I would not be surprised if the filters, envelope generators, LFOs etc work with higher resolution to reduce errors in the calculations.

We do know that the effects-processor has 20-bit both in and out, and actually delivers a 20-bit audio-signal. The internal processing in the effects-processor is most likely done at 24-bit, again to reduce the errors.

I usually compare this to modifying a picture digitally in Photoshop. You can take a low-resolution image and "map" it in to a higher resolution. Now, the picture itself does not get better from this, but if you start using Photoshop-tools such as smudge in the picture, the higher resolution becomes apparent, and the result is much better than it would be otherwise.

This technique can easily be translated into the audio-world. If you take "low-resolution" sinal, map it into higher resolution, and then process it, the output signal WILL have higher effective resolution.

My conclusion is that the XPs use their 18- or 20-bit converters to their full extent, and that this enhances the quality of the sound. The new XVs all have 24-bit converters. Why would Roland add 4-6 more bits to be unused??!?

As i said, I am not a professional in these fields, but I think this description should be fairly correct..

Stig

[This message has been edited by stigf (edited 03-28-2001).]