KVR Audio

itoa · Post by **itoa** » Thu Sep 11, 2014 4:10 pm

I was trying to add tracking lpf to the signal - original saw looks very close, but 50/50 very different

. Upper pulses are curved at the opposite directions! lol ... thats why this makes me crazy

btw. Thanks, how can HPF boost fundamental? Is it subtracted? Tell me more tell me more

itoa · Post by **itoa** » Thu Sep 11, 2014 4:12 pm

Aciddose: why? how? arguments?
I compared this recording with Redominator - exactly the same.

itoa · Post by **itoa** » Thu Sep 11, 2014 4:17 pm

aciddose wrote:
The theories... I have no idea how people come up with these fantasy theories.

Really? I'm just a person in a cave that sees a shadow of an elephant. I need to come through numerous fantasy theories before I recognise him

. But seriously what is wrong with waveforms I posted?

aciddose · Post by **aciddose** » Thu Sep 11, 2014 4:22 pm

itoa wrote:Aciddose: why? how? arguments?
I compared this recording with Redominator - exactly the same.

The aJU has exactly the same wave generator as the JU-106, with some additions.

The important things though are:

The waveforms are all clocked from the same source, they are digital, not analog
Due to the clock shared between the waveforms, all edges line up perfectly give or take a very small latency (in ns, absolutely non-audible)
The entire operation is performed in a single chip, just like in the JU-106, with some supporting logic on the outside
The noise is mixed from a single source, so in any JU- model the identical noise signal should be shared between all voices
You can get 100% of the information you need to accurately model the waveforms from the drawing on the panel of the synth

Okay, so "100%" isn't entirely accurate. "99%" then.

I'm not going to go to all the trouble to hook up my scope, probe the synth and take photos of the waveforms so you'll just have to trust me on this one: The waveforms you get directly out of that chip are way more pure than anything you could produce in software. So just try your best to produce absolutely pure mathematical shapes and you're there.

itoa · Post by **itoa** » Thu Sep 11, 2014 4:29 pm

Aciddose: and pwm saw is just a filtered saw am with pulse? :]

aciddose · Post by **aciddose** » Thu Sep 11, 2014 4:38 pm

Calling it "AM" is really not accurate. Technically it's "chopped". I haven't seen the internals of the IC, but I am willing to bet it looks like a resistor carrying the ramp signal into the collection of an NPN transistor. The emitter to ground, the base has the pulse signal on it.

The pulse signal is selected to be:

square 2x
pwm 2x
square 8x
square 8x chopped by square 2x

The interesting thing is really how they achieve the pwm 2x signal. Regardless, the output itself is very pure so there is no need to worry about it unless you want to design your own IC.

Urs · Post by **Urs** » Thu Sep 11, 2014 4:40 pm

itoa wrote:I was trying to add tracking lpf to the signal - original saw looks very close, but 50/50 very different . Upper pulses are curved at the opposite directions! lol ... thats why this makes me crazy

btw. Thanks, how can HPF boost fundamental? Is it subtracted? Tell me more tell me more

Shelving Filter

itoa · Post by **itoa** » Thu Sep 11, 2014 4:42 pm

Thanks! This is one of informations I've been waiting for. Couldn't find any schemes.

Ah and the most important... Poor me, indeed the recordings have been made with HPF set to 0, that means bass boost, I didn't know about that :/. I have to figure it out yet.. seems like the bass boost is a 2 pole lpf added to the original (URS: yes shelving, but LP

the only case with fundamental boosting hpf I can imagine is resonance)

aciddose · Post by **aciddose** » Thu Sep 11, 2014 5:21 pm

Technically it isn't "shelving", but it is a lowpass mixed directly. This is usually the same circuit used in analog "shelving" filters as it is very cheap and works well. It is similar enough that you'd probably never notice the difference.

Sendy · Post by **Sendy** » Thu Sep 11, 2014 8:56 pm

No theories offered by me, just a way to make that family of waveforms with readily available tools. I didn't realize this was the dev thread, this question usually comes up in Synthesis. Time for me to gracefully bow out now

mystran · Post by **mystran** » Fri Sep 12, 2014 5:13 am

itoa wrote:- No the time is not reversed - and this is the problem :]

So why is the ringing on the wrong side? Why are Sendy's images in flipped?

Seriously, your images are time-reversed. Didn't we have this problem with some earlier images as well?

janostman · Post by **janostman** » Sat Sep 13, 2014 12:54 pm

The Juno SAW PWM is nothing more than Wrap like I have on my dsp-G1.

You use the ordinary pulse PWM or with analog, a comparator, and when the pulse changes state you wrap or change the sign of the saw wave.

Simple but effectively PWM on any waveform.

aciddose · Post by **aciddose** » Sat Sep 13, 2014 3:39 pm

That isn't how the circuit works in the JU-, not even close.

janostman · Post by **janostman** » Sat Sep 13, 2014 4:37 pm

aciddose wrote:That isn't how the circuit works in the JU-, not even close.

Maybe not but that's what it is.
It's even used in the JP-8000 and called offset.

But if you want it to be something else then go ahead.

aciddose · Post by **aciddose** » Sat Sep 13, 2014 4:57 pm

What on earth do you mean "maybe not but that's what it is", do you have any idea how the electronic circuit works?

http://en.wikipedia.org/wiki/Chopper_%28electronics%29

What you've described regarding "wrapping" is completely unrelated to anything the JU- synthesizers do. The waveforms are chopped using a single transistor. This is all done from gnd to +v, there is no multiplication (AM/ring) and no wrapping involved. Do you have any idea of the complexity that would be involved in implementing a circuit to wrap an analog voltage?

Obviously not.

Roland Alpha Juno-2 pwm sawtooth