Patrick's new baby... TAL- J-8X

[beely]

but for sure if you set one at 6 and one at 7, they should behave differently.

Yes of course, we understand that. I think we're talking at cross purposes. Someone said they wanted to change eg the attack (like the global calibration on other synths). I suggested that, if you adjust the attack VT knob *the same on all voices*, does that give the effect you're after?

You then replied saying no, each voice would be different, and therefore my workaround would not work. I'm not in a position to test right now (I can look later), but it should be simple enough to try it and check the behaviour, and thus evaluate whether my advice *is* a workaround, or not - ie whether that knob is an offset, or a randomisation range.

(I assumed it was an offset, not a random amount, with a global random button to randomise the offsets across voices, but it could be that I am wrong.)

[zvenx]

You may be right.. I guess it all depends on what say a 6 value means vs a 10 value on the SC/VT.

But I always have assumed (well I think assumed) that it was relative to the other voices.
that is if you go in and change all of them to the same value, as a whole the synth doesn't behave differently, (than if they were all at zero), once those are all at the same value..

But I could be wrong.
(now let me go and type all of this on vi-contol just to piss off you know who )
rsp

[strangeflange]

...I don't think it's reasonable to ask Patrick to chase someone's JX-8P peculiarities. There could be a hundred guys out there, all telling him their JX-8P sounds a little bit different.

This comment right here

That's exactly what I implied. Patrick can't be chasing Youtube commenters or forum commenters, trying to please every JX-8P owner out there.

In any case, there will always be a guy who will tell him it's different and thus "not good enough because it doesn't match my unit". I merely implied that wasn't fair to Patrick.

Patrick can't be responsible for that. The goal of this project was to model Patrick's JX-8P Rev 2 and for that he is responsible to deliver it to the best of his present ability.

He already achieved a spectacular result, but the development is still ongoing and improvements will be made, no doubt.

[beely]

You may be right..

Ok I checked, and my initial assumption is correct. Easy way to check - set all voice detune knobs to minimum (which is half a semi-tone detuned), and play chords. If it's still in tune, all voices have been detuned by the same amount, so the knobs are therefore offsets.

(I checked with a few of the other voice parameters as well.)

So yes, they can be used as global calibration settings in the manner I previously described.

[zvenx]

Great..thanks for doing the test.. so it is an offset value.
rsp

[jamcat]

Here's an article on programming the JX-8P and Super JX from the October 1986 edition of Keyboard Magazine, written by Eric Persing himself.

Fun!

https://web.archive.org/web/20190427011 ... uperJX.pdf

[zvenx]

Here's an article on programming the JX-8P and Super JX from the October 1986 edition of Keyboard Magazine, written by Eric Persing himself.

Fun!

https://web.archive.org/web/20190427011 ... uperJX.pdf

What a find.. thanks for sharing.
rsp

[beely]

BTW most of those article patches are in the sysex banks you can find online - if you import the folders into the plugin, it's easy to then search for them. There were a couple that I didn't have, but someone can input them from the magazine article (just like we had to back in the day) to get those patches... (if so, please share them here!)

[Caesarg86]

JX-10/MKS-70 filters self-oscillate while the JX-8P don't. Nice addition by Roland to the Super JX.

[zvenx]

Version 1.0.5 / 09.04.2026

Possible click with voice stealing and VCA dynamic enabled fixed.
Aftertouch volume with VCA dynamic fixed. (calibration)
Slightly more env smoothing. (calibration)
First digital envelope step shortened. (calibration)
Chorus improved. (calibration)

There is an update available.
rsp

[hurricaneaudiolab]

Sweet, this is what TAL does - constant improvement, so this is great. I haven't had a chance to test the new version, but I’ve been A/B’ing the hardware vs the plugin pretty closely, and the main difference I keep landing on is the transient behavior. On the hardware, the attacks feel sharper and more defined. On the plugin, they’re slightly smeared or softened. I'm not the only one that's observed this.

So now I'm wondering about the limiter behavior in the chorus path. On the JX, the limiter is on the chorus send, not the dry signal. That means the dry signal can stay intact while the chorused signal is controlled. So perhaps TAL is not modeling this correctly, or is doing it differently. Thus the attacks are being masked or softened.

And then the main culprit for me is the chorus, where one of my working theories now is that it could be too wet. Or the phase relationship between the dry and wet signals of the chorus are causing smearing in some cases. Or maybe since it's a BBD chorus, the filtering of the highs on the chorused signal isn't happening in J8X.

The first factory patch - PIANO 1 is the perfect example. The transients are smeared on the J8X.

[simmo75]

Version 1.0.5 / 09.04.2026

Possible click with voice stealing and VCA dynamic enabled fixed.
Aftertouch volume with VCA dynamic fixed. (calibration)
Slightly more env smoothing. (calibration)
First digital envelope step shortened. (calibration)
Chorus improved. (calibration)

There is an update available.
rsp

Did you get an alert in the plugin or did you actively go to TAL to check?

[simmo75]

ignore

[zvenx]

...

Did you get an alert in the plugin or did you actively go to TAL to check?

With Tal you have to randomly check..
What was even worse this time, is it said 1.00 on the website (since corrected), but I redownloaded anyway and saw the date was newer and realised there was a new version..

rsp

[zvenx]

Sweet, this is what TAL does - constant improvement, so this is great. I haven't had a chance to test the new version, but I’ve been A/B’ing the hardware vs the plugin pretty closely, and the main difference I keep landing on is the transient behavior. On the hardware, the attacks feel sharper and more defined. On the plugin, they’re slightly smeared or softened. I'm not the only one that's observed this.

So now I'm wondering about the limiter behavior in the chorus path. On the JX, the limiter is on the chorus send, not the dry signal. That means the dry signal can stay intact while the chorused signal is controlled. So perhaps TAL is not modeling this correctly, or is doing it differently. Thus the attacks are being masked or softened.

And then the main culprit for me is the chorus, where one of my working theories now is that it could be too wet. Or the phase relationship between the dry and wet signals of the chorus are causing smearing in some cases. Or maybe since it's a BBD chorus, the filtering of the highs on the chorused signal isn't happening in J8X.

The first factory patch - PIANO 1 is the perfect example. The transients are smeared on the J8X.

Hi, I assume you did email him about the limiter chorus thing that you brought up here a few days ago.

I think I had pointed out that what we have is not a modeled jx8 chorus but an interpolation between the jx8p and jx10's chorus (well that is my understanding of it)

rsp