DIY Vocoder projects?

freewave · Joined: Aug 15, 2007 Posts: 8 Location: Ontario Canada

I do belive there was a fix mentioned on page four of this thread. I think the work on that pdf. was owsome and I want to try the project myself. I hope someone will get into this thread and breath some life into this interesting project.

Regards

Broadwave · Posted: Thu Apr 23, 2015 10:28 am Post subject:

I've almost completed the Okita build, just some hardware wiring and I should be good to go.

I'm afraid I think I may have found another error on the lower right of goodrevdoc's PCB... Cut the trace as shown below (I think this may be why the Op-Amp was getting hot)

The output of pin 1 (IC29) shouldn't be linked to the Instrument gain pot via the jumper... Maybe this is also causing the indicator problems - I'll find out tomorrow, when I can start testing the boards.

Cut trace.jpg

Description:

Filesize:

60.92 KB

Viewed:

31182 Time(s)

IMG_0289.jpg

Description:

Filesize:

1.91 MB

Viewed:

913 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Broadwave · Posted: Sat Apr 25, 2015 11:34 am Post subject:

OK, so breaking the track (above) fixes the indicator problem and no hot Op-Amps.

BUT... I've also found another major error on goodrevdoc's component layout - all the 4.7uf and 10uf tant caps around the NE570/571 are the wrong way round.

I'm just about to replace them, so I'll see if that fixes things.

Broadwave · Posted: Sun Apr 26, 2015 11:41 pm Post subject:

Built, tested & finally working Smile

I had to revert to Okita's original schematic, along with the additional 470k resistor modification on pins 1 & 16 to ground of the companders - sorry, but there's just too many errors on goodrevdoc's version Sad

Intelligibility is ok-ish, maybe not quite as good as the SVC-350 which it's based on, but still pretty usable, although I can detect a definite formant shift... maybe I should have used better quality capacitors Embarassed

You need to be careful of both input levels, otherwise it will distort! I found that it helps to roll off the bass end on both input signals. I've got an old chorus pedal somewhere, which I'll install at some point, but the good news is that substituting NE570's with NE571's works fine, and you can get them from Futurlec for a reasonable price.

For this quick test I used a Behringer C1 condenser mic, via my DI Box and Studiologic Sledge (Basic sawtooth, 3 oscillators, 1 octave apart, no detuning or filtering). Recorded directly from the Multimix 16 USB2 with a touch of reverb.

1 Okita Front.jpg

Description:

Filesize:

1.72 MB

Viewed:

579 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

2 Okita Top.jpg

Description:

Filesize:

1.85 MB

Viewed:

649 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

3 Okita Bottom.jpg

Description:

Filesize:

1.73 MB

Viewed:

664 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Just random mumblings.wav

Description:

Download (listen)

Filename:

Just random mumblings.wav

Filesize:

6.42 MB

Downloaded:

1131 Time(s)

mosc · Posted: Wed Apr 29, 2015 1:46 pm Post subject:

Congratulations on a beautiful build. It sounds quite good in your demo. I use the Nord Modular G2 for vocodeing. I like the unintelligible quality of vocoders. Sometimes, though, I want the voice to be understandable.

A little trick you can use is to mix in the direct voice in with the output of the vocoder. I process the direct voice with a high-pass filter as I have found that all you need to add for intelligibility are the sibilances, just the high frequencies. This is ironic (I love irony) because most audio engineering writings are all about how to reduce sibilance. Smile

I have two extra controls on the vocoder patch: one control is for a cross fader between the direct voice and the vocoder output, and one control is for the frequency of the high-pass filter on the direct signal. You thus get a lot more control.

zephyrin · Posted: Thu May 28, 2015 6:05 am Post subject:

Hello

I was starting to built this vocoder too.
I wanted to know if someonne found the 0.05 uf and 0.005 uf capacitor ?

I used for the other value , mylar capacitor with 5% tolrerance.

For 0.005 % i found styroflex with 1% tolerance.

for 0.05 uf i used 2 mylar capacitor (22 nf and 27 nf ) .

Broadwave · Posted: Thu May 28, 2015 8:27 pm Post subject:

I wouldn't worry too much about those caps, the standard 0.047 & 0.0047 values work fine - it's only a *tiny* shift in the frequency response, and I doubt if you would even notice it Wink

zephyrin · Posted: Fri May 29, 2015 5:59 am Post subject:

thank you AndyR1960
I have some questions about the wiring.

I have used tha japanese PDF for help me.

I think there are a misstake for the phone pot.
i see wire numer 15 and 14 on PDF but i think is number 13 and 14

i have also a question about the volume pot with stereo option what is the number of the wires ? i just see number 5 !

For the cut trace i see on the japanese PDF that there are a J connexion do you know how to connect it ?

For the tantale cap i compare with the japanese PDF and it's seem the same for me

I want also to built the ETI (powertran, ELRAD) vocoder, i m starting to rebuilt the typon for the big PCB ...

pcb eti total-2.jpg

Description:

Filesize:

2.1 MB

Viewed:

911 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

zephyrin · Posted: Thu Jul 23, 2015 2:27 am Post subject:

there is an other trap with akita vocoder.
I wanted to used the original Transistor but i don't see that the design has been changed for using the europeans transistor. So if you want to used the japanese transistor like me you need tu swap Base and collector Pin and also put the transistor in the other way.

I have also some trouble with the signal indicator and i have try to change the 4558 by some TL072 and it seem to be better ..

Paradigm X · Posted: Mon Jul 27, 2015 1:10 am Post subject:

Wow, Andy, that looks incredible! Amazing work.

Are you considering making any pcbs to sell?

Cheers
Ben

zephyrin · Posted: Mon Jul 27, 2015 8:30 am Post subject:

I have finished my Okita vocoder this is some pictures of the machine

Okita0.png

Description:

Filesize:

796.41 KB

Viewed:

398 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

okita1.jpg

Description:

Filesize:

118.46 KB

Viewed:

368 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

okita2.jpg

Description:

Filesize:

146.69 KB

Viewed:

411 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

okita3.jpeg

Description:

Filesize:

467.22 KB

Viewed:

531 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Paradigm X · Posted: Tue Sep 15, 2015 9:33 am Post subject:

wow, now that is a beautiful build as well. Amazing work.

isak · Posted: Tue Sep 15, 2015 10:05 am Post subject:

Andy, its amazing!
Shocked

electrospeaker · Joined: Jun 30, 2006 Posts: 49 Location: Norway

LFLab · Posted: Sat Sep 19, 2015 12:03 am Post subject:

Well, there's a schematic in the app note, it's called the "one knob squeezer" iirc
Pretty simple compressor.

electrospeaker · Joined: Jun 30, 2006 Posts: 49 Location: Norway

Yes indeed.
But you'd still need some insight in order to incorporate it with the existing input amp circuitry.

zephyrin · Posted: Mon Sep 21, 2015 12:53 pm Post subject:

Unfortunatly my okita vocoder have some defect with the instrument input when i m playing and changing the note there some audio crunch.

I have also built the MFOS vocoder

Some pictures of the vocoder

http://img.audiofanzine.com/image.php?lang=fr&identifier=image&size=normal&module=user&userPhoto_id=267920
http://img.audiofanzine.com/image.php?lang=fr&identifier=image&size=normal&module=user&userPhoto_id=285129

http://img.audiofanzine.com/image.php?lang=fr&identifier=image&size=normal&module=user&userPhoto_id=267926

http://img.audiofanzine.com/image.php?lang=fr&identifier=image&size=normal&module=user&userPhoto_id=267924

http://img.audiofanzine.com/image.php?lang=fr&identifier=image&size=normal&module=user&userPhoto_id=267925

Some extract
internal oscillator
https://soundcloud.com/user-90411072/vocoder-mfos-ininterne

Juno alpha1
https://soundcloud.com/user-90411072/vocoder-mfos-feel-so-close-juno

https://soundcloud.com/user-90411072/vocoder-mfos-blame-juno

https://soundcloud.com/user-90411072/vocoder-mfos-blame2

electrospeaker · Joined: Jun 30, 2006 Posts: 49 Location: Norway

Wow! Having built two vocoders already and on your way to build the ETI you must really like vocoders Wink

As mentioned earlier I'd really like an ensemble effect (similar to the Roland VP-330 vocoder and other Roland devices) at the output stage of my ETI vocoder, and having download a PDF of the Okita vocoder I noticed that it mentions an "Option stereo chorus print circuit board space" (page 61 in the PDF) but there's no schematic or PCB layout for this (as far as I can see).

Does anyone know anything more about this specific section?

zephyrin · Posted: Mon Sep 21, 2015 3:19 pm Post subject:

In fact i have also start the elektor vocoder.

i found this early, perhaps we can add this in okita vocoder:
here there are someone who clone the juno 106 chorus

http://forum.anafrog.com/phpBB/viewtopic.php?f=36&t=14125

and here the clone of chorus of juno 60

http://forum.anafrog.com/phpBB/viewtopic.php?f=36&t=13956&hilit=chorus+clone

electrospeaker · Joined: Jun 30, 2006 Posts: 49 Location: Norway

Thanks -I hadn't heard about those Roland chorus clones before.

I believe the "ensemble" effect I loved when I had my Roland VP-330 is basically a "chorus on steroids" type effect with several BBD chips all doing their own thing, hence the wider, more lush and "fat" sound without the pitch modulation of a basic chorus unit, right?

Having discussed the subject before and understanding how hard it is to clone the VP-330 ensemble (or other similar Roland keyboards with a similar ensemble) or another DIY ensemble effect which is relatively inexpensive and fits inside the vocoder enclosure I'm wondering if it might just be easier to rip out the PCB of some kind of enseble effect pedal.
I've been suggested Behringer's CC-300 which is supposed to be a clone of the (rather expensive) Boss DC-2 Dimension-C (which in turn is a cut-down pedal version of the (even more expensive) Roland SDD-320 Dimension-D.

Alas I hear it's a poor substitute for the Boss DC-2, apparently being digital (some people claim it's analog but then again that might just be the input stage). Besides I hear it's tricky to make it work with a line signal (i.e. the vocoder), so I'll have to look elsewhere.
There's another ensemble called the Boss CE-5 Chorus Ensemble which I see is based on BBD chips (earlier versions at least) and doesn't seem to sell at such inflated prices as the DC-2.
How does it compare to the DC-2 Dimension C (and more importantly, the Roland VP-330 ensemble), and will it accept (or can easily be modified for) line-level signals?

CE5.jpg

Description:

Filesize:

20.28 KB

Viewed:

28645 Time(s)

DC2.jpg

Description:

Filesize:

10.71 KB

Viewed:

28645 Time(s)

be64c9ae_6c72_11e2_a5f1_001e673d50f2_bf28ea9b_8cd9_11e3_90ee_001e673d50f2.jpg

Description:

Filesize:

127.14 KB

Viewed:

310 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

sdd320.JPG

Description:

Filesize:

25.03 KB

Viewed:

348 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Paradigm X · Posted: Wed Sep 23, 2015 4:14 am Post subject:

electrospeaker · Joined: Jun 30, 2006 Posts: 49 Location: Norway

Paradigm X · Posted: Mon Sep 28, 2015 3:50 am Post subject:

I dont know about costs, but Oakley's uses only easy to get parts, nothing old and rare. The ensemble is back in stock now fwiw.

As its a fully supported PCB i would imagine its much easier than scott's. Ive built a few oakley bits and they all worked first time, and sound great.

Getting pots and ICs from Tayda is a good way to save costs.

hesed · Joined: Mar 24, 2010 Posts: 6 Location: Dresden

Hi to all,

i am following this since years and after debugging here and there i found a way that solve distortion and signal/peak detector problems.

On the image attached you can find all the details.

Yellow: 500k to 100k already mentioned in this topic.

Orange: cut trace already mentioned in this topic.

Red: it should be 10k referring to okita schematics.

Green: from 47k to 10k reduce the gain sensibility so there is more space before reaching distortion

Blue: put the 1M resistor away and replace it with a 100ohm trimpot,
at the beginning i was quite surprised of that but now the signal led is working perfectly.

I still have 2 problems, maybe somebody can help me:

1: the 4053 is getting really hot (already fried one), at the moment i replace it with 3 spdt to get signal thru.

2: on the voice board the 356 has a strange behaviour, if i put it away all bands work perfectly if i put it in band 2,4,6,8,10 are "eating" the signal away, like they were inverted in phase.

mainIO-board-debug.jpg

Description:

Filesize:

516.19 KB

Viewed:

445 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

zephyrin · Posted: Sun Oct 25, 2015 6:48 am Post subject: