The supersaw oscillator

[Noteman]

Hey,

I would like to hear how the JD-XA supersaw compares to the original and IMHO still unmatched king of the sound - the JP8000. If anyone could send me a clip of just one held note (say C6) with all the FX off and a fair amount of detune (about 3/4 of the way), that would be great and hugely appriciated. Does anyone know if the digital filters include the old JP8000 digital filter? Just wondering if this could be a replacement for the old but still irreplaceable tech...

[HaveIt]

I sold my JP-8000 some time ago and regretted it ever since. I do like the JD-XA SuperSaw very much though. Would be great to hear high quality sound demoes of the JP-8000 and reproduce them in the JD-XA.

B

[goldphinga]

Dont forget you can use the analogue section to do this.

Use all 8 oscillators set to saw, detuned a little, there you go!

[Noteman]

This might sound strange, but part of the reason why the JP8000 supersaw is such an awesome sound is because it has crazy digital aliasing. It really makes the top come alive.

I was able to check the system-1 supersaw. That one has only 5 saws somewhat evenly detuned (the JP8 had 7 and logarithmic detune) and less (although some) aliasing. It seems like the sound has no high pass filter either as was the case on the JP8K. If anyone could post the JDXA supersaw, just a long note all effects off, fair bit of detune, that would be helpful!

I am not in the studio at the moment, so can't send a demo. I will once I get there though...

[OneOfManyPauls]

If anyone could post the JDXA supersaw, just a long note all effects off, fair bit of detune, that would be helpful!

Here's a zip with 2 wav files containing two held notes in each. Both are single supersaw partials - the first with 0 detune, the second with detune set to 90 (from a possible 127).

[OneOfManyPauls]

Here's a full 3 partial part with the 3 supersaws detunes set to 80,90 and 100.

[Noteman]

Thank you very, very much OneOfManyPauls... Very nice!

It actually COULD BE the JP8k supersaw, no less! I will have to do a direct comparison. There is a high pass filter as with the JP, and... a lot of sizzly aliasing. The saw wave count seems to be 7 or 8... it's a bit hard to tell (higher detune helps tell the waves apart).

But I also discovered something very weird. The second example you sent actually nulls to very high degree with the first (1partial_90detune and 3_partials). Are you sure that you didn't accidentally post the many partials example two times? If that's not the case, then the wave is SAMPLED, which is no good : /. However, something seems weird, because I don't think it should null to the degree it does even if it was indeed sampled with the difference in the settings. I realize that the clips are different length though... weird.

Here is the nulled version followed by the original. https://clyp.it/0tmch4v4

If you want me to confirm send me a file where the same note is played twice or more in a row. You should also be able to hear a machine gun effect while playing many notes in a row if it's a sample.

It's very sad if it turns out that Roland has sampled their own legendary digital creation, in a VA synth no less.

Thanks again.

[OneOfManyPauls]

They're definitely different files - here's a 1partial 90 detune again- same note 3 times repeat.

[OneOfManyPauls]

just for fun, here's an 8 detuned analog oscillator patch (bypassed filter)
these are the settings for the fine detunings of each oscillator: -30 -17 -7 0 +5 +15 +23 +33

[Noteman]

They're definitely different files - here's a 1partial 90 detune again- same note 3 times repeat.

Great to hear again...

I looked at the file and sure enough, the notes null. The waves are thus samples taken from the JP8000. A real shame Roland would demean their own created legends like this, in a flag-ship synth, no less : (. Oh well, I think I will pass until they release a real VA version of the digital section, playing back samples does not cut it. So weird that they aced the analog part, but dropped the ball a bit on the digital part - which they already were supposedly experts in. Would have been amazing to use the analog part in conjunction with real classic Roland VA... (I assume the other waves are samples as well).

Anyway, thanks for helping out, oh and here is the null, original as last note: https://clyp.it/fqo5pcnf

[OneOfManyPauls]

I always took it that all the digital osc waveforms were just looping samples rather than VA oscillators.

[Noteman]

I always took it that all the digital osc waveforms were just looping samples rather than VA oscillators.

Yea, you were smart... However, I thought that in the manual the "oscillator" waves would actually be computed waves, since the numbering in the parameter guide starts after them. Should have known better with all the references to super-natural sampling engine. Still can't quite fathom why Roland would do this though : /...

At minimum, they could have added random starting point to the samples so that they would at least sound like a real synth being played, instead of a machine gun being shot (for certain passages). Oh well, never can have it all.

[HaveIt]

Are the digital supersaws samples? Just stop a minute before shouting YES.

I cannot detect any discernable looping artifacts. Can anyone? Is it more accurate to say the supersaws could be single cycle loops which are summed and detuned exactly like the JP-8000. Is there really any difference? Or in fact could the JP-8000 be doing exactly the same things. It is no less VA if it did. But it's a cheap shot to cry 'samples' and reject something as inferior. Looping is only bad if there is detectable looping or aliasing at high frequencies.

It is true that the supersaw results in the identical wave calculation, which can cause cancellation - there is no point is stacking more than one. Although detuning and analog feel should work, IMO it doesn't sound bigger or phatter than one supersaw.

B

[trrstrl]

The core synth sounds (sine, saw, square) are supposedly modeled. I couldn't find if that means the alternate versions (super-saw, sine-B, etc.) are also modeled but it seems likely. I tried comparing the V-Synth super-saw to the one on the JD-XA. The one on the V-Synth sounded a bit more organic and beefier in the low end. On the high end the V-Synth was less natural than the JD-XA. It seems likely that it is just a different type of modeling being used. To me the V-Synth sounds different from the JP8000 even though they were supposedly the same synth engine. The Jp8000 has a more organic and in your face type of sound that I like better than the JD-XA's Supernatural synth. The Supernatural synth engine might more accurately emulate analog in some ways though. If you are after that specific super-saw sound you might have to have a jp8000(8080).

[HaveIt]

I was disappointed with the V-synth super saw havng had a JP-8000, but the JD-XA has that 'fizz' of the original. Seems you can't assume a supersaw from Roland is the same on every keyboard.