| Author |
Message |
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
 Posted: Fri Jun 04, 2010 12:02 pm Post subject:
|
  |
|
Oh, I missed a week of discussion here, and was just about to say -
"Hey, don't you need resistor pins to do signal mixing in the patch panel? I can only see shorting pins and diode pins on the Ghielmetti site!"
| jhaible wrote: |
The Ghielmetti from my link above is around 150 Euros, for 20x20.
...
I think for a vocoder the small size (3mm between rows and columns) should be enough. I don't expect wild on-the-fly crosspatching as with a VCS3. Also, shorting pins should be ok here - you want to interchange sources for destinations, not mix them. (Or do you? ... well, maybe ...) |
Hmmm, I would definitely like to be able to mix them, if I had a pin matrix in the vocoder...
| jhaible wrote: | | BTW: Are there any interesting demos for cross-patched vocoder channels? I admit that I never saw such a big benefit in it. The formant shifting that is always quoted as a main application can also be done with a pitch shifter in the Analyzer channel, can't it? |
Yes, I guess it can. You're probably right in the assumption that arbitrary cross-patching/mixing is an "exotic" feature with a usability that can be questioned.  If formant shift is what you would primarily aim for with a matrix, then I would almost prefer a simple rotary switch to select how many stages up/down the shift would be - and skip the matrix, and save €150! A matrix with 20 pins to move is totally "unplayable" in this respect.
It would be really good to hear the opinions from someone using a Syntovox/EMS vocoder with patch panel!
(As a side note, even if zthee mentioned the risk of feature creep when it comes to the patchability, I'm thinking a bit about the possibility of another "remote-controlled matrix" connection with LOTS of more functionality than a physical pin matrix... nasty me. I'd better not tell.) 
| zthee wrote: | | jhaible wrote: | | zthee wrote: | | How does mixing work with a patch matrix? |
Basically, you just use resistor pins instead of shorting pins.
If you want real independent mixing without any crosstalk, you'd need a virtual GND summing node input for each channel. I will *not* include this (not even the VCS3 has it), but you can add it on a piece of veroboard. It's just 40 opamps (10 quads like TL074).
But if you're using resistors of about 4.7k, you should get decent results with just a little crosstalk, without these opamps. |
Sounds easy enough!
Thanks! |
Note: The mentioned 4.7k solution was to get rid of crosstalk, not to enable mixing.
To enable mixing, you need resistor pins.
It felt like you misinterpreted it, but I don't know. |
|
|
Back to top
|
|
 |
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
| Posted: Fri Jun 04, 2010 12:14 pm Post subject:
|
|
|
| jhaible wrote: | Layout for Silence Bridging / Filter Bank finished.
...
3) The Filterbank function will fade in when there is no speech. So there is always a coloured Excitation signal. While you speak, the colour is set by the speech. When you don't speak, the colour is set by the 20 Filterbank sliders. This allows a seamless blend of ariticulated and non-articulated sound.
You can set the threshold - how many dBs down the speech signal is, before the filterbank is engaged -, and the attack and release time of the filterbank fade in/out, with 3 potentiometers. |
Oh - is that really how it's done?
I always expected the silence bridging feature to just "keep the levels a pre-set bit above zero if the voice signal goes under it". What I mean is that the output vocoder level of a certain stage is the MAX value of its analyzer value and its filterbank-pre-set value. (This would be implemented with diodes/resistors to form some simple MAX function of the two CV signals in each stage.)
But you mean that it actually fades in, with a certain attack time - which means that
1) for long attack times, it can be more silent before the fade-in happens, and
2) for low thresholds, the silence bridging can be louder than the speech-induced vocoded signal?
I'm puzzled, my world is all slippery! |
|
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007
Posts: 2014
Location: Germany
Audio files: 24
|
| Posted: Fri Jun 04, 2010 2:04 pm Post subject:
|
|
|
| /mr wrote: | Oh, I missed a week of discussion here, and was just about to say -
"Hey, don't you need resistor pins to do signal mixing in the patch panel? I can only see shorting pins and diode pins on the Ghielmetti site!"
|
For teh Synthi Clone, I bought Diode Pins, opened them, and replaced the diode with a 2k7 resistor.
Then I learned you can custom-order them that way from Ghielmetti.
Not on their web site - just call and ask.
JH.
_________________
"I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007
Posts: 2014
Location: Germany
Audio files: 24
|
| Posted: Fri Jun 04, 2010 2:09 pm Post subject:
|
|
|
| /mr wrote: | | jhaible wrote: | Layout for Silence Bridging / Filter Bank finished.
...
3) The Filterbank function will fade in when there is no speech. So there is always a coloured Excitation signal. While you speak, the colour is set by the speech. When you don't speak, the colour is set by the 20 Filterbank sliders. This allows a seamless blend of ariticulated and non-articulated sound.
You can set the threshold - how many dBs down the speech signal is, before the filterbank is engaged -, and the attack and release time of the filterbank fade in/out, with 3 potentiometers. |
Oh - is that really how it's done?
I always expected the silence bridging feature to just "keep the levels a pre-set bit above zero if the voice signal goes under it". What I mean is that the output vocoder level of a certain stage is the MAX value of its analyzer value and its filterbank-pre-set value. (This would be implemented with diodes/resistors to form some simple MAX function of the two CV signals in each stage.)
But you mean that it actually fades in, with a certain attack time - which means that
1) for long attack times, it can be more silent before the fade-in happens, and
2) for low thresholds, the silence bridging can be louder than the speech-induced vocoded signal?
I'm puzzled, my world is all slippery! |
Everyone seems to implement it slightly different.
EMS crossfades with the dry signal, for instance.
I guess your method of using the max(analyzer, filterbank_pot) for each channel will spoil the articulation.
Sure, not every attack or release setting makes sense. You can put a trimpot on the board to optimize to your taste. Or solder wires and connect a front panel pot.
louder: There will be an extra pot to adjust the relative volume, in addition to the individual 20 filterbank pots. So you can adjust it with a single knob.
Good setting of the right input level in the analyzer channel is mandatory, anyway. The built-in optoelectronic compressor will help.
JH.
_________________
"I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
|
Back to top
|
|
|
Funky40
Joined: Sep 24, 2005
Posts: 875
Location: Swiss
Audio files: 1
G2 patch files: 5
|
| Posted: Fri Jun 04, 2010 3:14 pm Post subject:
|
|
|
| /mr wrote: |
| jhaible wrote: | | BTW: Are there any interesting demos for cross-patched vocoder channels? I admit that I never saw such a big benefit in it. The formant shifting that is always quoted as a main application can also be done with a pitch shifter in the Analyzer channel, can't it? |
Yes, I guess it can. You're probably right in the assumption that arbitrary cross-patching/mixing is an "exotic" feature with a usability that can be questioned. |
hmmm........i never patched anything with a Voice with my Doepfer vocoder.
I use the Vocoder for Groove stuff only.
..............i hardly like to have a Vocoder with mixing stages on very every input
the point is the added livingness you can patch. ....just patch a "living" modulation or even further a living distribution
it's unbelievable how groovy stuff you can patch on a vocoder.....( and the doepfer vocoder is really crappy. but it just not matters ) |
|
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
| Posted: Fri Jun 04, 2010 4:43 pm Post subject:
|
|
|
| jhaible wrote: | For teh Synthi Clone, I bought Diode Pins, opened them, and replaced the diode with a 2k7 resistor.
Then I learned you can custom-order them that way from Ghielmetti.
Not on their web site - just call and ask. |
Oh.
Anyway, the diode pins (and I guess, also the resistor pins) are quite expensive at over 4€ each, making them cost more than half of what the matrix costs (159€). |
|
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
| Posted: Fri Jun 04, 2010 4:54 pm Post subject:
|
|
|
| jhaible wrote: | Everyone seems to implement it slightly different.
EMS crossfades with the dry signal, for instance.
I guess your method of using the max(analyzer, filterbank_pot) for each channel will spoil the articulation. |
That might be correct, actually.
(Even if I don't think the weaker bands are very important for the articulation - a bit like when using a parametric eq in a mix, it's the peaks that you hear a lot (and might get problems with when mixing), not the dips.)
Anyway, I'm excited (but not hurried) to hear the result one day!
(Meanwhile, I get myself your Moog 914 Fixed Filter Bank clone in the other thread, hehe) |
|
|
Back to top
|
|
|
j.dilisio
Joined: May 19, 2009
Posts: 200
Location: baltimore
|
| Posted: Sat Jun 05, 2010 9:09 am Post subject:
|
|
|
I can't wait to see/hear the final product.
This is exciting stuff!
_________________
DRONEGOAT |
|
|
Back to top
|
|
|
magnus
Joined: Aug 12, 2007
Posts: 8
Location: Stockholm, Sweden
|
| Posted: Mon Jun 07, 2010 11:29 pm Post subject:
|
|
|
| jhaible wrote: |
Basically, you just use resistor pins instead of shorting pins.
If you want real independent mixing without any crosstalk, you'd need a virtual GND summing node input for each channel. I will *not* include this (not even the VCS3 has it), but you can add it on a piece of veroboard. It's just 40 opamps (10 quads like TL074).
But if you're using resistors of about 4.7k, you should get decent results with just a little crosstalk, without these opamps.
BTW: Are there any interesting demos for cross-patched vocoder channels? I admit that I never saw such a big benefit in it. The formant shifting that is always quoted as a main application can also be done with a pitch shifter in the Analyzer channel, can't it?
Anyway: I'll provide the possibility of cross-patching, of course. On the PCB, it's just a row of connectors where you plug in the matrix, or jumpers instead.
JH. |
For the Synthi-A the cross-talk is a feature. For a vocoder it would cause reduced speech intelligibility. However, it is not often that you do such complex patches on a vocoder. It is more likely that you interchange bands than mix bands from several sources.
To illustrate crosstalk, consider this patch:
| Code: |
I I
1 2
O1 ---x
O2 -x-x |
I1 is supposed to only get output 2, but we want both O1 and O2 to mix into I2 but now O1 leaks into the O2 output due to its relatively high output impedance and then into I1 as cross-talk.
For the Synthi-A, it's a part of it's "it's alive" character. |
|
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
| Posted: Tue Jun 08, 2010 7:14 am Post subject:
|
|
|
| magnus wrote: | | For the Synthi-A the cross-talk is a feature. For a vocoder it would cause reduced speech intelligibility. However, it is not often that you do such complex patches on a vocoder. It is more likely that you interchange bands than mix bands from several sources. |
How do you know that? Maybe I really really want to activate the mid bands of the vocoded minor chord on both the offbeat hihat AND the low toms in the fill-ins... otherwise we get no reggae!
Honestly I'm not sure if I'll ever use the vocoder for robot voices. It would be wise to consider many other applications as useful and likely.
| magnus wrote: | | For the Synthi-A, it's a part of it's "it's alive" character. |
And here "it's a living vocoder!"  |
|
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007
Posts: 2014
Location: Germany
Audio files: 24
|
| Posted: Wed Jun 16, 2010 5:31 am Post subject:
|
|
|
(Tiny) Update
Layout of the Compander section finished.
I'm using as similar compander as in the Arp Quadra Phaser / Storm Tide Flanger. This keep the level of the Voiced Replacement input signal fairly constant, so automatic control of the Noise as Unvoiced Replacement will not be necessary.
Also including an "S" Generator like the EMS 2000 did, but with a completely different implementation.
Rimu Layout program responds better to huge PCB, now that I've finally bought a new PC after 6 years.
JH.
_________________
"I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
|
Back to top
|
|
|
nerdware
Joined: Jul 11, 2009
Posts: 91
Location: UK
|
| Posted: Wed Jun 16, 2010 5:51 am Post subject:
|
|
|
| jhaible wrote: | | Rimu Layout program responds better to huge PCB, now that I've finally bought a new PC after 6 years. |
_________________
http://soundcloud.com/nerdware/
"render unto digital what is due to digital, render unto analogue what is due to analogue" |
|
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007
Posts: 2014
Location: Germany
Audio files: 24
|
| Posted: Wed Jun 16, 2010 5:56 am Post subject:
|
|
|
| nerdware wrote: | | jhaible wrote: | | Rimu Layout program responds better to huge PCB, now that I've finally bought a new PC after 6 years. |
|
... and the new PC is even more quiet than the previous one. The old one was already fanless, and the new one is without hard disk, also. Absolutely no noise.
JH.
_________________
"I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
|
Back to top
|
|
|
nerdware
Joined: Jul 11, 2009
Posts: 91
Location: UK
|
| Posted: Wed Jun 16, 2010 12:16 pm Post subject:
|
|
|
Excellent. I approve of silent computers. 
_________________
http://soundcloud.com/nerdware/
"render unto digital what is due to digital, render unto analogue what is due to analogue" |
|
|
Back to top
|
|
|
toppobrillo
Joined: Dec 10, 2005
Posts: 766
Location: oakland, ca
G2 patch files: 1
|
| Posted: Mon Jun 21, 2010 2:22 am Post subject:
|
|
|
hey Jurgen,
just caught up with this over the past week or so.. looks unbelievable!! sign me up for one |
|
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007
Posts: 2014
Location: Germany
Audio files: 24
|
| Posted: Fri Jun 25, 2010 10:19 am Post subject:
|
|
|
Latest News:
Power Supply will fit on the PCB!
Of course the mains transformer, primary fuse and the power transistors on the heat sink will be external, but everything else - secondary fuses, regulators, capacitors, rectifier - will be on-board!
http://www.jhaible.de/vocoder/working_on_vocoder.jpg
JH.
_________________
"I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
| Posted: Fri Jun 25, 2010 7:09 pm Post subject:
|
|
|
| jhaible wrote: | Latest News:
Power Supply will fit on the PCB! |
Reservoir gods! |
|
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007
Posts: 2014
Location: Germany
Audio files: 24
|
| Posted: Tue Jun 29, 2010 7:06 am Post subject:
|
|
|
First Part of Schematics is up!
http://www.jhaible.de/vocoder/living_vocoder.html
Any hint on errors is very welcome at that stage of the project.
This will be a 40cm x 30cm PCB, and I don't want to do this twice!
JH.
_________________
"I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
| Posted: Tue Jun 29, 2010 3:44 pm Post subject:
|
|
|
| jhaible wrote: | | This will be a 40cm x 30cm PCB, and I don't want to do this twice! |
Could you make a lab/prototype area in the unused upper right corner?
Supplies & Ground pads would be handy to have there as well. |
|
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007
Posts: 2014
Location: Germany
Audio files: 24
|
| Posted: Tue Jun 29, 2010 11:10 pm Post subject:
|
|
|
| /mr wrote: | | jhaible wrote: | | This will be a 40cm x 30cm PCB, and I don't want to do this twice! |
Could you make a lab/prototype area in the unused upper right corner?
Supplies & Ground pads would be handy to have there as well. |
That's exactly what I intended to do.
Not all of it - there are still some minor parts missing - but what space is left, will be like veroboard. I know that everybody has different preferences what to have in a vocoder, apart from the regular / mandatory stuff.
JH.
_________________
"I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007
Posts: 2014
Location: Germany
Audio files: 24
|
| Posted: Mon Jul 05, 2010 10:21 am Post subject:
|
|
|
Update
I just placed my order at the factory for this:
http://www.jhaible.de/vocoder/schematics/living_vocoder_comp_overlay.pdf
JH.
_________________
"I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
|
Back to top
|
|
|
janvanvolt
Joined: Nov 24, 2005
Posts: 285
Location: Mainz, Germany
|
|
|
Back to top
|
|
|
jhaible
Joined: May 25, 2007
Posts: 2014
Location: Germany
Audio files: 24
|
| Posted: Mon Jul 05, 2010 2:30 pm Post subject:
|
|
|
| janvanvolt wrote: | | any final price for the PCB itself ? |
Not sure yet. Probably a bit more expensive than the EUR 150.00 I had in mind for the earlier 14-band vocoder. I'll have to make final calculations when the whole thing is finished (if it's actually working!), and when I can see how many boards will be ordered. But whatever the final figure for the PCB will be, it will be the smallest part of the whole project. Capacitors will be a main price factor, and pots and enclosure, and the optional matrix.
This will *not* be an unexpensive vocoder, but it will be ridiculously unexpensive compared to what you pay for an EMS 5000 or Sennheiser.
Speaking of capacitors: If you don't have much money, but a lot of time, you might buy a couple thousand of unexpensive 10% or 5% capacitors, select the right ones using capacitance meter, and then keep the un-selected ones for other, less critical projects.
JH.
_________________
"I tell you the truth, if anyone says to this mountain, 'Go, throw yourself into the sea,' and does not doubt in his heart but believes that what he says will happen, it will be done for him. Therefore I tell you, whatever you ask for in prayer, believe that you have received it, and it will be yours." (Mk 11,23f) |
|
|
Back to top
|
|
|
janvanvolt
Joined: Nov 24, 2005
Posts: 285
Location: Mainz, Germany
|
| Posted: Tue Jul 06, 2010 12:54 am Post subject:
|
|
|
Okay, so take me in for one PCB (if it works and i'll already start ordering the capacitors (5% will be). What's the ideal value to look for of those two ?
_________________
Homepage - http://www.czmok.de
My dIY - http://diy.czmok.de
Film/Music - http://gfm.me |
|
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
| Posted: Tue Jul 06, 2010 4:41 am Post subject:
|
|
|
| janvanvolt wrote: | | i'll already start ordering the capacitors (5% will be). What's the ideal value to look for of those two ? |
Go upwards in this thread and check jhaible's posts for the answer.
It's not that far. |
|
|
Back to top
|
|
|
|
Forum index » DIY Hardware and Software » Jürgen Haible designs
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
