what de f**ck is that man : always positive waves...

Mehdi_T · Post by **Mehdi_T** » Wed Jul 07, 2004 11:00 pm

In a waveform, like a sample in your favorite audio app, how would you interprete :

-the wave graphics which goes under the middle line (0)

- a wave which has only positive value ? (always positive, never goes under the zero line : never, like in never, never , never) (just discovered that on a certain device, a track made 2 years ago)... and the result in a mix... well...

Anyone can answer the second part ? have you ever noticed that ? what does it mean in theory ?

[ This Message was edited by: Mehdi_T on 2004-07-08 00:24 ]

nprime · Post by **nprime** » Wed Jul 07, 2004 11:39 pm

Waveform only above line means a massive DC offset. Possibly incorrect cable used for digital data transfer.

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 12:35 am

Modular (2 or 3, I don't remember) patch through asio, direct to cubase. I remember i read something about that before in the modular discussion, so it's apparently a known issue. Thanks for the answer !

[ This Message was edited by: Mehdi_T on 2004-07-08 01:36 ]

at0m · Post by **at0m** » Thu Jul 08, 2004 12:37 am

Dc problem a known issue? It's synthesis and maths, how to make AC or DC from a patch. I use quite some patches to manipulate modulation signals, which are mainly DC. The blue wire in SFP project view means there's a 32bit data stream from one point going to the other. PWM synthesis is extremely suitable for DC generation, check on your oscilloscope

Alfonso updated our Flexosaurus patch, after I discovered it could generate DC too. He inserted DC filters which don't really solve the problem, and sometimes have their own artifacts as on subs, but it was the best way to preserve the original patch and its preset sounds.

_________________
More has been done with less.

[ This Message was edited by: at0m|c on 2004-07-08 01:46 ]

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 12:38 am

mmm, I think it's a patch which used granular synthesis in a weird way... I haven't used the modular much before, so...

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 12:39 am

one of the strange effect is that, with high gain (but under saturation) the sound simply disappear just as if it was phase-cancelled (but it's mono so...).
so i'll try to take care of this in the future (if it's possible)...
Thanks again !

[ This Message was edited by: Mehdi_T on 2004-07-08 01:40 ]

at0m · Post by **at0m** » Thu Jul 08, 2004 12:59 am

Whenever you mix, phase problems can occur in modular, also on a mono signal. In DP you could lock atoms to a chip. Phase problems can occur when the path before mix passes thru a module on another chip. Also, the Xmod&FeedBack Connector module is one sample delay. You worked with phasers and delays, and you hear the result... For example on digital delays, you could set BPM slight higher (for rushy feel) or lower (laidback) than song BPM. On top of that, the timing offset avoids delay induced phasing which sometimes appears on 'perfectly on time' delays.

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 1:00 am

I know the issues about delays. it's not as simple as locking a few atoms to a chip (which can bring some other problems, less easily fixable than phase inmho), it's not enough, and require permanent attention (re: subs) but as you say, it's caused/managed by dsp allocation (or something like that) ... i even wonder if that atrom on same dsp is worse (in my view and experience, and after serious analysis of various situation, i came to the conclusion that it's a bit useless and illusory - at least for devices which use more than one DSP ?. phase fix with micro delays is not always enough neither.
MAths ? may be, but as long as the modular is not for mathematicians (is it?) the output should be correct in terms of DC offset. It's pretty weird that it output sounds that cannot be mixed normally, like other synths &effects (it's more or less the same atoms!). But may be it is indeed for mathematicians, after all... I'm just a musician, trying to work fast, no time to get bothered with too much side issues

Of course, all what I say is conditionned by the fact that I use a generation 1 board (I've been told some issues have been resolved in Generation 2 boards ).

[ This Message was edited by: Mehdi_T on 2004-07-08 04:29 ]

[ This Message was edited by: Mehdi_T on 2004-07-08 04:33 ]

[ This Message was edited by: Mehdi_T on 2004-07-08 04:46 ]

blazesboylan · Post by **blazesboylan** » Thu Jul 08, 2004 3:47 am

With high gain you get basically a minorly fluctuating DC current going to your speakers, so of course there would be no sound. You need the speaker to move in and out (AC) before any air moves to make sound.

My suspicion about root cause -- mind you I've never seen, tried, or imagined this to be possible with Modular, but -- you've got some kind of rectifier circuit. What does the waveform look like, Mehdi? Does it have flatline sections? Or does it maybe look like the path of a bouncing ball?

If it looks more or less like a gentle sine wave then it probably has been just DC shifted (i.e. it's not a rectifier circuit after all). But that would be a lot of shift!

Cheers,

Johann

[ This Message was edited by: blazesboylan on 2004-07-08 04:49 ]

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 3:49 am

I'll try to upload a pic, but as you know (now) i only have a royal 2 mb left for spacef (I hope there is still room for a litlle gif). It has to be seen I think.
see you in 5 minutes.

[ This Message was edited by: Mehdi_T on 2004-07-08 04:50 ]

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 4:05 am

here is a pic

As you can see it is a perfect line. absolutely nothing goes under 0.

My problem is that I wanted to put that sound louder, and it disappeared. I had to lower the main outputs of cubase by a few db (12) to be able to have enough headroom to put the above sample at the correct volume in the mix. But I think it is only a workaround.
you are right, I haven't used the modular for years so there is certainly an awful lot of issues i'm not aware of (i never had this with devices for example).

Is there a simple cure for this ? like something to add in the modular (or on that sample) to get rid of that effect ?

[ This Message was edited by: Mehdi_T on 2004-07-08 05:08 ]

blazesboylan · Post by **blazesboylan** » Thu Jul 08, 2004 4:08 am

How long is the sample?

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 4:09 am

3/4 seconds approx, almost 3 measures at 139 bpm.

And it is done without tricks ; that's really what comes out of the Modular (it uses wavetable oscialltors modulated by additional wavetable oscillators and granular stuff).

[ This Message was edited by: Mehdi_T on 2004-07-08 05:17 ]

alfonso · Post by **alfonso** » Thu Jul 08, 2004 4:16 am

An example on how to get full positive signal with bipolar ones is to multiply a signal with itself, 2 positive values multiplied give a positive value, 2 neg. values give positive value as well....

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 4:19 am

the signal is not multiplied by itself. The osciallation wave, and modulation wave are different waveforms. There's no feedback circuit or things like that. Just an oscillator modulated by another one (different wave) and the modulator is not "heard" per se, it goes to the mod input of the main osc.
I know that basic math Alfonso

(i know a bit more too about those common rules

)

2 neg. values give positive value as well....

So if the wavetable osc have negative values the above sould go negative too (they are not synched/similar, so neg*pos=neg)

So what, is it normal ?

[ This Message was edited by: Mehdi_T on 2004-07-08 05:26 ]

blazesboylan · Post by **blazesboylan** » Thu Jul 08, 2004 4:25 am

Well I can't figure out how the hell you would 1) convert an audio signal of some sort into that or 2) convert that back into an audio signal of some sort.

Audio is by definition alternating current. With current and a positive voltage force you push/pull the speaker in one direction. With negative the speaker pushes / pulls in the opposite direction.

DC is not audio.

All alternating currents have a "zero" reference point. When they cross this threshold the current "alternates". This sample is pure DC.

I don't believe that sample makes any sound. Sorry, I just don't believe it.

Therefore you might have a few options, but none of them will get anything really meaningful out of that bizarre sample:

DC shift the whole thing down. This will not give you anything like the above waveform.
Assume that the sample is really high frequency (> 22.05 kHz at 44.1 etc, but from the looks of it more likely = 44.1 kHz) and use god knows what tools to basically phase-invert every other sample. This will at least give you some approximation of the waveform, and your dog might enjoy the sound.
Replace the sample.
Keep the sample as is but realize that it will only work as a strange "effect" in the mix. It will not work as a "sound" on its own. Never. It's impossible.

Sorry I'm not more helpful.

Who knows, maybe someone out there will come along with an obscure idea...

Cheers & good luck Mehdi,

Johann

blazesboylan · Post by **blazesboylan** » Thu Jul 08, 2004 4:29 am

On 2004-07-08 05:16, alfonso wrote:
An example on how to get full positive signal with bipolar ones is to multiply a signal with itself, 2 positive values multiplied give a positive value, 2 neg. values give positive value as well....

Hi Alfonso.

That would be a form of full wave rectification. But the sample does not look like any rectified audio signal I can imagine!

alfonso · Post by **alfonso** » Thu Jul 08, 2004 4:33 am

I know you know...

it's just for general information. it's a fact however that something is overlooked somewhere...

Another thought...if the two waveforms are different, but they run at the same frequency, they will have the same algebrical sign at a certain moment, as the zero point. Obviously this applies in the case the two waveforms start their cycle in the same moment at the same point....

What type of modulation is occurring between them? I'd be happy to have a look to the patch...never tired to learn something new..

blazesboylan · Post by **blazesboylan** » Thu Jul 08, 2004 4:33 am

Mehdi are the osc + mod waves same frequency?

Or are they complex waveforms, or...?

I'm enjoying this very much but I need my sleep! Hopefully y'all will have solved it by the time I wake up.

Johann

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 4:57 am

Blazesboylan : I'm glad I'm not the only one not to beleive this

Alfonso : thanks

(i haven't slept for 48 h, i guess i'm becoming a bit touchy

sorry about that..
(mp3 removed, sorry i needed the space).
But you can find the moduylar patch in the Modular Forum.

[ This Message was edited by: Mehdi_T on 2004-07-08 15:14 ]