Mehdi Weird PureDC sound

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 6:16 am

<a name="planetz-file"></a><a href="http://www.planetz.com/Pulsar/files/mod ... .zip"><img src="/forums/images/file_icon.gif" border="0" alt=" File"> File</a> <a name="planetz-tag"></a>Type: Synth <a name="planetz-tag"></a>Pulsar Version: SFP 3.1 <a name="planetz-tag"></a>Requires: Modular 3 _____________________________________ WARNING ! lower your volume before playing this patch : unamicable high frequencies may disturb you (but it's not that harmful under normal conditions, just don't pump up the volume and read below before use)

Hi there !
This a patch uploaded for scientific purposes, because I would like to know what happens and why.
It is linked to the thread here :
http://www.planetz.com/forums/viewtopic ... forum=5&20
It looks like it doesn't match DC theory explained on the thread, because Only the parts which can be considered as "Pure DC" produce a sound that can be heard and which is interesting. The rest is un-hearable and requires sample cut after recording.
Well, have fun, but be careful

PS : if you can't get a sound with direct midi input, try with a long sustained note from your sequencer ( i think i have a midi keyboard issue too

.
Also, the midi channel is 1 , not omni, I didn't pay attention, sorry.

[ This Message was edited by: Mehdi_T on 2004-07-08 07:36 ] <a name="planetz-fileimage"></a><IMG SRC="http://www.planetz.com/Pulsar/files/mod ... Csound.jpg" BORDER="0">

Michu · Post by **Michu** » Thu Jul 08, 2004 8:11 am

Mehdi
as far as i can see oscillators are not connected to any frequency source, ie you use granular gater on slowly changing DC.
Now, oscillator sounding shite because not connected to frequency source is a known issue

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 8:30 am

it sounds great

(well, depends on the oscillation)

it's not about how it sounds at all... it's about understanding why the wave doesn't go under zero (rectified? as warp69 said).

But you are right there a missing connection to the Sin Freq In... less fun when it's connected (pretty horrible, but on the whole oscillation, imho, no more kookoo in the deep forest).

Doesn't matter anyway, the mix is finished

and I remember why i stopped using the modular...

(i'm exagerating)

[ This Message was edited by: Mehdi_T on 2004-07-08 15:27 ]

Michu · Post by **Michu** » Thu Jul 08, 2004 9:07 am

hmm,
try watching output of oscillator not connected to frequency signal with oscilloscope.
Modular is great tool as long as you understand what you are doing

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 9:13 am

i don't

this is only one ofmany patches, and it does the sound i want (don(t think there 's another way to get that sound, which is great michu, when you know how to use it

(you still didn't get the point of the post read better please)

[ This Message was edited by: Mehdi_T on 2004-07-08 10:18 ]

alfonso · Post by **alfonso** » Thu Jul 08, 2004 9:27 am

As Mehdi said, be very careful!
DC can destroy your speakers if high enough.

I checked also the Granuliner2 patch from the stock M III Creamware patches (I know who built it

together with a bit less the half of those) and on a couple of presets it produces indeed a noticeable dc offset.

I inserted a dc blocker module between the Vol Attenuator on the down right side and the Phaser module. It makes it's job well, leaving the sound as it is but recentering the zero crossings.

I strongly recommend You add the DC Blocker too. It doesn't add any parameter to the preset list, so you will not have the x% valid preset file issue. Add it on both Mehdi's patch and Granuliner 2 patch.

I would add it post the poly module in ALL Modular patches. It can save your speakers, I confess that as a Patch maker I overlooked this issue in the past.

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 9:53 am

I inserted a dc blocker module between the Vol Attenuator on the down right side and the Phaser module. It makes it's job well, leaving the sound as it is but recentering the zero crossings.

I strongly recommend You add the DC Blocker too. It doesn't add any parameter to the preset list, so you will not have the x% valid preset file issue. Add it on both Mehdi's patch and Granuliner 2 patch.

I would add it post the poly module in ALL Modular patches. It can save your speakers, I confess that as a Patch maker I overlooked this issue in the past.

You're great ! thanks

Michu · Post by **Michu** » Thu Jul 08, 2004 11:36 am

mehdi
think of dc offset as of sound of frequency equal to 0 Hz or close to it
any HPF with cutoff quite low will get rid of it.

Mehdi_T · Post by **Mehdi_T** » Thu Jul 08, 2004 11:43 am

Thanks Michu

(you are great too

)

[ This Message was edited by: Mehdi_T on 2004-07-08 15:10 ]

blazesboylan · Post by **blazesboylan** » Thu Jul 08, 2004 10:06 pm

OK I originally came to this thread only to try this patch out but first...

On 2004-07-08 12:36, Michu wrote:
think of dc offset as of sound of frequency equal to 0 Hz or close to it
any HPF with cutoff quite low will get rid of it.

What?!? DC offset and frequency have nothing to do with each other. DC offset affects the amplitude. It will make the "push" of the speaker stronger than the "pull", or vice-versa. The frequency stays untouched.

An example sine wave with +ive DC offset:

Code: Select all

    ___          ___
   /           /            / 1.275 V
  /           /            /
  ___________/____________/__  0 V
          __/         ___/    -0.275 V

Nevertheless Mehdi's bizarre waveform looks like no signal that DC offset or rectification would produce. Full wave rectification of a sine wave (withOUT DC offset) would look like:

Code: Select all


    ___    ___    ___    ___   0.775 V
   /     /     /     /   
  /_____/_____/_____/_____ 0 V

                               - 0.775 V

Half wave rectification would look like:

Code: Select all


    ___           ___          0.775 V
   /            /   
  /____________/____________ 0 V

                               - 0.775 V

Mehdi's waveform is something out of the Twilight Zone!

Cheers,

Johann

Mehdi_T · Post by **Mehdi_T** » Fri Jul 09, 2004 1:40 am

hehe

(thanks for clarifying too)

blazesboylan · Post by **blazesboylan** » Fri Jul 09, 2004 2:47 pm

Mehdi I couldn't reproduce this. I don't know if I'm doing something dumb, but... Please let me know if I missed a step or 2:

Hooked up a MIDI keyboard to your Modular patch
Modular to ASIO out
Recorded 2 mono tracks (1 left, 1 right) in sequencer (happens to be Nuendo)
Hit a key on the Midi Keyboard and held it for several seconds

I also recorded the Modular Out with a DC filter, but the difference for the first part of the sample was negligible (if anything). (Later on, after 5-10 seconds, there seems to be a substantial difference, but still not 100% DC).

Here's what I got: from top to bottom

Before DC filter - Left channel
Before DC filter - right channel
After DC filter - left channel
After DC filter - right channel

Here are a couple of zoom-ins of the same wave form (it looks like a square wave):

Zoom 1:

Zoom 2:

Obviously these look nothing like your original waveform!

Did I miss a few steps or something? Or maybe I hit the wrong note? ("A3" I think...)

I was tired last night, so I wouldn't be surprised if I messed somethin' up...

(BTW You probably won't hear back from me 'til Sunday but I am intrigued by this one, so I'll be back!

)

- Johann

[ This Message was edited by: blazesboylan on 2004-07-09 15:50 ]

[ This Message was edited by: blazesboylan on 2004-07-09 15:51 ]

Mehdi_T · Post by **Mehdi_T** » Sat Jul 10, 2004 3:36 pm

Hi,
what is your recording app ?
Here, SX (1.6), modular direct to asio gives you something similar to my pic (on my system).

I play G3 or A3 (should sound the same anyway).

With the DC filter of Modular, I loose the interesting part of the sound. I mean musically. I removed sub/infra basses with several HP, (all cut under 400/500 hz, approx 8 HPs just to be sure

just to avod any speaker movememts in basses-it sounds very good in the mix now, clean and very clear, no bass pollution like before, and i still have the whole useful spectrum to play with).

I've also noticed that cubase vu meters do *not* follow the shape of the drawn sample, but follow a curve that is similar to your shape (right, or bottom shape (as I can clearly see now : in cubase, the vu meters are much lower on the second part (your shape) whereas it should be louder (my shape)....
Recorded in stereo track, i have identical shapes on L and R channels.

I will not be here often this w-e neither...

Mehdi

[ This Message was edited by: Mehdi_T on 2004-07-10 17:37 ]

blazesboylan · Post by **blazesboylan** » Sat Jul 10, 2004 5:50 pm

Nuendo 1.6.

(Actually in the process of testing your Mod the other night I discovered that I was recording in Nuendo at 24bit but using 16bit ASIO in SFP. Duh. Will have to figure out if there's a way to get 64 channels of 24bit ASIO in SFP, doesn't look like it. May be too slow anyway.)

(Edit: Found that I need to use the ASIO 2-32 64 source and dest. Manual says it's for 32 OR 24 bit.)

This is very odd. I'll try again tonight and take sanpshots of my SFP setup etc. I half wonder if there's some kind of a rendering bug in Cubase?!? Maybe if you put a mark at the beginning of the sample, completely remove it from the project, and re-insert it at the marked position? Does the waveform still look the same?

TTFN,

Johann

[ This Message was edited by: blazesboylan on 2004-07-10 21:00 ]

Mehdi_T · Post by **Mehdi_T** » Sat Jul 10, 2004 8:09 pm

HI

I use 44.1/16 bit, asio 1 or 2 (same results).
From a metaphysical point of view, I would be disappointed if it is only a graphic bug

Reloading/resampling/bouncing : do not change the graphic shape in SX (by resampling, ie output to scope and record back into cubase) I can get an additional, very short peak under zero at the begining, but nothing that makes it so different from the above graph (my shape)...
i'll try resampling in another app tomorrow.
Mehdi

[ This Message was edited by: Mehdi_T on 2004-07-10 21:25 ]

at0m · Post by **at0m** » Sat Jul 10, 2004 8:53 pm

To refresh the image of the Cubase/Nuendo audio file, you can simply delete the .peak file from the project folder's Images folder.

Also, make sure your horizontal zoom doesnt skip any x-crossings. Then there is different settings in the sequencer that could show a rectified wave, a bipolar one, smoothened etc...

Mehdi_T · Post by **Mehdi_T** » Sun Jul 11, 2004 3:31 pm

i will test everything, but be patient... (not too much time in fact for the moment)...
One thing which is strange in Blazeboylan pics, is that the patch normally outputs Mono (or, dual mono, the same on both channels anyway, no pan effect or stereo differences, i will check anyway...).

[ This Message was edited by: Mehdi_T on 2004-07-11 16:42 ]

blazesboylan · Post by **blazesboylan** » Mon Jul 12, 2004 3:45 am

Didn't get a chance to do anything today, but just some quick extra info -- I used the Modular out 1 and 2. Does that sound right, Mehdi? Or which outputs did you use from the Modular?

Johann

Mehdi_T · Post by **Mehdi_T** » Mon Jul 12, 2004 1:56 pm

either both or Left only (it should be the same, unless effects are activated). that's what i tried last week at least.

Michu · Post by **Michu** » Wed Jul 14, 2004 10:10 am

What?!? DC offset and frequency have nothing to do with each other. DC offset affects the amplitude. It will make the "push" of the speaker stronger than the "pull", or vice-versa. The frequency stays untouched.

heh,
set sine osc to 0 Hz, what do you get?
now route DC offset thru hpf (creamware ones have lowest possible cutoff of 20 Hz so they filter out DC even in lowest setting).
It's gone...
take care guys