Synth-DIY Yahoo! Groups Archives

New Voice Chips?

2014-05-29 by bperkins211@...

Maybe the topic got buried in the other thread, but there was mention of trying to build a new voice for the HAWK.

I was suggesting perhaps we try to use a simple 8bit PIC with a small wavetable set, one PIC for each voice. So maybe a 6 or 8 chip voice board? Imagine the combo's you could come up with since each voice could be a different wavesample.

I put a line out to the guy who wrote the code for the Electric Druid's VC DCO pic chip. He might give some pointers on how to mod his wavetable PIC code to work with a digital uC instead of all VC.

I did post questions to Mike about some of the features he was after.. I think my post got covered over and lost before he saw it??

~Blaine

Re: [korgpolyex] New Voice Chips?

2014-05-29 by Michael Hawkins

Yes, I am in general agreement regarding using 8 PIC chips, one for each DCO. And hopefully, each can take a voltage control signal for envelope generator VCA control. Because we already have those available in the existing circuitry. Then we also need another voltage control input for frequency control. So that MG and joystick pitch bend can control each PIC oscillator. Finally, we also need detuning (which can be digital control.

Mike

On Wednesday, May 28, 2014 8:28 PM, "bperkins211@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:

Maybe the topic got buried in the other thread, but there was mention of trying to build a new voice for the HAWK.

I was suggesting perhaps we try to use a simple 8bit PIC with a small wavetable set, one PIC for each voice. So maybe a 6 or 8 chip voice board? Imagine the combo's you could come up with since each voice could be a different wavesample.

I put a line out to the guy who wrote the code for the Electric Druid's VC DCO pic chip. He might give some pointers on how to mod his wavetable PIC code to work with a digital uC instead of all VC.

I did post questions to Mike about some of the features he was after.. I think my post got covered over and lost before he saw it??

~Blaine

Re: [korgpolyex] New Voice Chips?

2014-05-29 by Michael Hawkins

And lastly, we need a digital interface that allows selecting note frequency across the same range as the existing TG.

On Thursday, May 29, 2014 6:12 PM, "Michael Hawkins korgpolyex800@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:

Yes, I am in general agreement regarding using 8 PIC chips, one for each DCO. And hopefully, each can take a voltage control signal for envelope generator VCA control. Because we already have those available in the existing circuitry. Then we also need another voltage control input for frequency control. So that MG and joystick pitch bend can control each PIC oscillator. Finally, we also need detuning (which can be digital control.

Mike

On Wednesday, May 28, 2014 8:28 PM, "bperkins211@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:

Maybe the topic got buried in the other thread, but there was mention of trying to build a new voice for the HAWK.

I was suggesting perhaps we try to use a simple 8bit PIC with a small wavetable set, one PIC for each voice. So maybe a 6 or 8 chip voice board? Imagine the combo's you could come up with since each voice could be a different wavesample.

I put a line out to the guy who wrote the code for the Electric Druid's VC DCO pic chip. He might give some pointers on how to mod his wavetable PIC code to work with a digital uC instead of all VC.

I did post questions to Mike about some of the features he was after.. I think my post got covered over and lost before he saw it??

~Blaine

Re: New Voice Chips?

2014-05-30 by bperkins211@...

Take another look at Tom's VC DO
VC'd wavetable monosynth, basically.

http://www.electricdruid.net/index.php?page=projects.vcdo

we can use his code, here
http://www.electricdruid.net/code/VCDO1.ASM
and modify it to drop unneeded VC inputs.
Change them into a bus port and select lines.

PICs are very simple to config. They are RISC Harvard Arch.
we can port over code to a larger chip if we need more ports to config.

You can read his code to get an idea of the formula's he's used and his conversion of the wavetables to a PWM output.. evidently it's a high Hz and needs filtered to sort out the Audio.

Looks like it's not setup for a VCA.. guess we could do that separate using an example circuit from other synths if needed.. or use an off the shelf VCA chip.

Would you be using S&H methods to distribute VC's out from the DAC with a 3 to 8 decoder or some method like that?

Google "Knowles Know It All PIC microcontroller" to find a posted PDF of a very good book on PICs to evaluate. I liked it so much that I bought a book copy on ebay for cheap.. which I highly recommend for you as well.. these are great little devices.

Re: [korgpolyex] Re: New Voice Chips?

2014-05-30 by Michael Hawkins

Don't seem to be able to get to the electricdruid pages. Maybe they're down or maybe I have a problem. I shall check 'em out again tomorrow.

Mike

On Thursday, May 29, 2014 8:34 PM, "bperkins211@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:

Take another look at Tom's VC DO
VC'd wavetable monosynth, basically.

http://www.electricdruid.net/index.php?page=projects.vcdo

we can use his code, here
http://www.electricdruid.net/code/VCDO1.ASM
and modify it to drop unneeded VC inputs.
Change them into a bus port and select lines.

PICs are very simple to config. They are RISC Harvard Arch.
we can port over code to a larger chip if we need more ports to config.

You can read his code to get an idea of the formula's he's used and his conversion of the wavetables to a PWM output.. evidently it's a high Hz and needs filtered to sort out the Audio.

Looks like it's not setup for a VCA.. guess we could do that separate using an example circuit from other synths if needed.. or use an off the shelf VCA chip.

Would you be using S&H methods to distribute VC's out from the DAC with a 3 to 8 decoder or some method like that?

Google "Knowles Know It All PIC microcontroller" to find a posted PDF of a very good book on PICs to evaluate. I liked it so much that I bought a book copy on ebay for cheap.. which I highly recommend for you as well.. these are great little devices.

Re: [korgpolyex] New Voice Chips?

2014-05-30 by John David Duncan

This all sounds great.

I guess there's also D/A converters.
And then combining the new audio signal with the old signal before the input to the filter (without overloading the filter input and clipping).
Am I right about that or am I misunderstanding something?

JD

>
>
> And lastly, we need a digital interface that allows selecting note frequency across the same range as the existing TG.
>
>
> On Thursday, May 29, 2014 6:12 PM, "Michael Hawkins korgpolyex800@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:
>
>
>
>
> Yes, I am in general agreement regarding using 8 PIC chips, one for each DCO. And hopefully, each can take a voltage control signal for envelope generator VCA control. Because we already have those available in the existing circuitry. Then we also need another voltage control input for frequency control. So that MG and joystick pitch bend can control each PIC oscillator. Finally, we also need detuning (which can be digital control.
>
> Mike
>
>
> On Wednesday, May 28, 2014 8:28 PM, "bperkins211@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:
>
>
>
> Maybe the topic got buried in the other thread, but there was mention of trying to build a new voice for the HAWK.
>
> I was suggesting perhaps we try to use a simple 8bit PIC with a small wavetable set, one PIC for each voice. So maybe a 6 or 8 chip voice board? Imagine the combo's you could come up with since each voice could be a different wavesample.
>
> I put a line out to the guy who wrote the code for the Electric Druid's VC DCO pic chip. He might give some pointers on how to mod his wavetable PIC code to work with a digital uC instead of all VC.
>
> I did post questions to Mike about some of the features he was after.. I think my post got covered over and lost before he saw it??
>
> ~Blaine
>
>
>
>
>
>
>

Re: New Voice Chips?

2014-05-30 by bperkins211@...

I just checked and ED's main page is indeed down now.
It was working when I posted the link...

His code page worked for me. But then it could be the cache on my PC opening it up since it is simple text.

Odd.. he mentions Olivier Gillet, from Mutable Instruments, as an influence for his PIC code.

Check it out tomorrow, I'm sure it will be back up then.

Re: [korgpolyex] New Voice Chips?

2014-05-30 by Michael Hawkins

I'll need to see the PIC stuff first before I can make comment. I may try and get a basic functional diagram going so that we can nail down the precise way in which we can plug this into the existing Poly main board and start comparing techniques and component choices.

If the PIC VCO is producing a digital output then we shall need a D2A but they can be constructed with some low tolerance resistor ladders = cheap. I have to also check the levels on the Poly to see if the existing TG is producing 0-5 or -5 - +5 volt signals. Either way, we can easily deal with that too. The challenge is to make this CHEAP.

I have also looked at using a 28C256 chip and using sampled waveforms much like the DW-8000 DWGS but I think the PIC path is a better one with the potential to produce a unique sound, possibly updatable/downloadable code. Possibly a selection of waveforms. If we can use a PIC for each voice that's in the 2-3 buck range then supporting components and PCB should all come somewhere in the 50-100 range total kit cost. I would definitely pay 100 bucks to upgrade the tone generator in the Poly to be true independent clocks along with independent detune, modulation etc plus better actual waveforms themselves. PWM, square, saw, triangle. Apart from the complete lack of MIDI (which was fixed by the HAWK), the waveform generation on the Poly is the worse thing about the sythn. It's one of the reasons it gets a lot of criticism and is often considered a toy.

Mike

On Thursday, May 29, 2014 9:14 PM, "John David Duncan john.david.duncan@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:

This all sounds great.

I guess there's also D/A converters.
And then combining the new audio signal with the old signal before the input to the filter (without overloading the filter input and clipping).
Am I right about that or am I misunderstanding something?

JD

>
>
> And lastly, we need a digital interface that allows selecting note frequency across the same range as the existing TG.
>
>
> On Thursday, May 29, 2014 6:12 PM, "Michael Hawkins korgpolyex800@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:
>
>
>
>
> Yes, I am in general agreement regarding using 8 PIC chips, one for each DCO. And hopefully, each can take a voltage control signal for envelope generator VCA control. Because we already have those available in the existing circuitry. Then we also need another voltage control input for frequency control. So that MG and joystick pitch bend can control each PIC oscillator. Finally, we also need detuning (which can be digital control.
>
> Mike
>
>
> On Wednesday, May 28, 2014 8:28 PM, "bperkins211@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:
>
>
>
> Maybe the topic got buried in the other thread, but there was mention of trying to build a new voice for the HAWK.
>
> I was suggesting perhaps we try to use a simple 8bit PIC with a small wavetable set, one PIC for each voice. So maybe a 6 or 8 chip voice board? Imagine the combo's you could come up with since each voice could be a different wavesample.
>
> I put a line out to the guy who wrote the code for the Electric Druid's VC DCO pic chip. He might give some pointers on how to mod his wavetable PIC code to work with a digital uC instead of all VC.
>
> I did post questions to Mike about some of the features he was after.. I think my post got covered over and lost before he saw it??
>
> ~Blaine

>
>
>
>
>
>
>

------------------------------------

------------------------------------

Feel free to upload into the files section any sysex dumps and tape dumps of patches that you may have but please discuss (with the entire group) the posting of other files *before* posting them. This helps us to keep redundant information from showing up everywhere and also allows us to constantly improve the format and structure of the documentation. We talk about the HAWK-800 quite a bit but we also discuss and help owners of the original Poly-800 models. So don't be shy if you haven't got the HAWK-800.Yahoo Groups Links

<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/korgpolyex/

<*> Your email settings:
Individual Email | Traditional

<*> To change settings online go to:
http://groups.yahoo.com/group/korgpolyex/join
(Yahoo! ID required)

<*> To change settings via email:
korgpolyex-digest@yahoogroups.com
korgpolyex-fullfeatured@yahoogroups.com

<*> To unsubscribe from this group, send an email to:
korgpolyex-unsubscribe@yahoogroups.com

<*> Your use of Yahoo Groups is subject to:
https://info.yahoo.com/legal/us/yahoo/utos/terms/

Re: [korgpolyex] New Voice Chips?

2014-05-30 by bperkins211@...

Looks like the ED page is working again.

If you want to hear a few samples of some of the wavetables, go here..

Waverider Digital VCO | Frequency Central

Waverider Digital VCO | Frequency Central

Waverider Digital VCO – here we go! Now in production, wavetable VCO with tons of features: CV’d interpolating waveforms, CV’d subosc waveforms and octaves, ...

View on www.frequencycentra...

Preview by Yahoo

that's another site that sells modular kits for ED's chips.

The pic he used has only two 8bit ports to configure.. if you make one to be an interface to the HAWK's data bus, that leaves the other port to config. ADC's, timer inputs, etc. from.
Each bit on the ports can be config'd to a fixed list of options.. so only certain pins can be timer ins/outs, ADC's and so on.

If we made one port to listen in on the data bus, we can easily get the data to interact with the PIC's onboard program.
I dont know if you'd want to create a 3bit select out of 3 pins from the other port or how you'd go about to trigger the PIC to capture the data on the bus.. but then you'd only have 5 pins left for the output and ADC's.

He shows a bipolar ADC input interface in his datasheet. so it should be easy to use that form of CV if needed.

If we need a 3rd port to give us another 8bits/pins to work with, I'd recommend the 16F1938
Here is it's datasheet
http://ww1.microchip.com/downloads/en/DeviceDoc/41574B.pdf

Tom's/ED's code should be easily ported over to this chip since it matches the clock, etc. It cost's $3 each.
I did get in touch with Tom about his PIC and the idea to make a poly voice with it.. he said he'd have time to talk after a 3week absence. But I think we can figure out alot of it just by his docs and and other avail. ref's.

~Blaine

Re: [korgpolyex] New Voice Chips?

2014-05-30 by Gordon JC Pearce

On Thu, May 29, 2014 at 07:59:20PM -0700, Michael Hawkins korgpolyex800@... [korgpolyex] wrote:

>
> I'll need to see the PIC stuff first before I can make comment. I may try and get a basic functional diagram going so that we can nail down the precise way in which we can plug this into the existing Poly main board and start comparing techniques and component choices.
>
>
> If the PIC VCO is producing a digital output then we shall need a D2A but they can be constructed with some low tolerance resistor ladders = cheap. I have to also check the levels on the Poly to see if the existing TG is producing 0-5 or -5 - +5 volt signals. Either way, we can easily deal with that too. The challenge is to make this CHEAP.
>

Don't know why you'd bother with PIC, the toolchain is shite and they are electrically "fragile". If you use a low-end ARM chip you may need to do something to convert between 5V and 3.3V logic, but you'll have enough grunt for all eight voices.

Doing two voices and all four outputs for square and saw ought to be doable in a single atmega 328.

Use a PWM output to generate the "analogue" output instead of an R/2R DAC. Use the counter for the PWM to trigger the sample calculation at around 32kHz and allow it to run asynchronously from the control loop. When the PWM rolls over it'll call an ISR that you use to load the comparison register with the calculated sample value.

Pish easy.

--
Gordonjcp MM0YEQ

Re: [korgpolyex] New Voice Chips?

2014-05-30 by bperkins211@...

Don't know why you'd bother with PIC, the toolchain is shite and they are electrically "fragile". If you use a low-end ARM chip you may need to do something to convert between 5V and 3.3V logic, but you'll have enough grunt for all eight voices.

Doing two voices and all four outputs for square and saw ought to be doable in a single atmega 328.

Use a PWM output to generate the "analogue" output instead of an R/2R DAC. Use the counter for the PWM to trigger the sample calculation at around 32kHz and allow it to run asynchronously from the control loop. When the PWM rolls over it'll call an ISR that you use to load the comparison register with the calculated sample value.

Pish easy.

--
Gordonjcp MM0YEQ

I suggested the PIC since it is the only uC I have any kind of experience with and currently studying, for now.. I was trying to offer ANY help on this attempt to revoice the HAWK.

Your suggestion sounds nice. I remember you posted something on this and some code as well awhile back.
Is the PWM outputs from it require any filtering at all?

Do you have any docs to point to on using wavetables on a ARM chip? Can you maybe lend more detailed advice on this?
I take it to get 8 voices total, that we'd need two atmega 328's?

I'd like to know more about any chip used for Arduino projects anyway...
I am looking to start building quadcopters and other UAV's using some of the open source Flight Controllers available like 3dRobotics APM.

~Blaine

Re: [korgpolyex] New Voice Chips?

2014-05-31 by Michael Hawkins

I am also looking at the STM32F107RC. ARM 32-bit-M3 running at 72Mhz, 256K flash, 64K RAM with 2x12 bit D/A converters, 80 I/O ports and tons of other good stuff. All for about 10 bucks.

I also happen to be working on another project that uses the same device. So I have to set up the development environment and get familiar writing code for it anyway.

Mike

On Friday, May 30, 2014 11:45 AM, "bperkins211@... [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:

Don't know why you'd bother with PIC, the toolchain is shite and they are electrically "fragile". If you use a low-end ARM chip you may need to do something to convert between 5V and 3.3V logic, but you'll have enough grunt for all eight voices.

Doing two voices and all four outputs for square and saw ought to be doable in a single atmega 328.

Use a PWM output to generate the "analogue" output instead of an R/2R DAC. Use the counter for the PWM to trigger the sample calculation at around 32kHz and allow it to run asynchronously from the control loop. When the PWM rolls over it'll call an ISR that you use to load the comparison register with the calculated sample value.

Pish easy.

--
Gordonjcp MM0YEQ

I suggested the PIC since it is the only uC I have any kind of experience with and currently studying, for now.. I was trying to offer ANY help on this attempt to revoice the HAWK.

Your suggestion sounds nice. I remember you posted something on this and some code as well awhile back.
Is the PWM outputs from it require any filtering at all?

Do you have any docs to point to on using wavetables on a ARM chip? Can you maybe lend more detailed advice on this?
I take it to get 8 voices total, that we'd need two atmega 328's?

I'd like to know more about any chip used for Arduino projects anyway...
I am looking to start building quadcopters and other UAV's using some of the open source Flight Controllers available like 3dRobotics APM.

~Blaine

Re: [korgpolyex] New Voice Chips?

2014-05-31 by Gordon JC Pearce

On Sat, May 31, 2014 at 06:32:28AM -0700, Michael Hawkins korgpolyex800@... [korgpolyex] wrote:

>
>
> I am also looking at the STM32F107RC. ARM 32-bit-M3 running at 72Mhz, 256K flash, 64K RAM with 2x12 bit D/A converters, 80 I/O ports and tons of other good stuff. All for about 10 bucks.

There's just about enough processor there to run a whole PPG Wave emulation.

Now stick eight of them in the same box.

--
Gordonjcp MM0YEQ

Re: [korgpolyex] New Voice Chips?

2014-05-31 by John David Duncan

Over at http://www.adventurekid.se/ there are around 4,000 open source
single-cycle waveforms. These are 16-bit .WAV files and are about
1300 bytes each. I was thinking it might be very nice to have a few
basic ones built in standard and then another 20 or so that are user
selectable.

JD

On May 31, 2014, at 6:32 AM, Michael Hawkins korgpolyex800@...
[korgpolyex] wrote:

>
>
>
> I am also looking at the STM32F107RC. ARM 32-bit-M3 running at
> 72Mhz, 256K flash, 64K RAM with 2x12 bit D/A converters, 80 I/O
> ports and tons of other good stuff. All for about 10 bucks.
>
> I also happen to be working on another project that uses the same
> device. So I have to set up the development environment and get
> familiar writing code for it anyway.
>
> Mike
>
>
> On Friday, May 30, 2014 11:45 AM, "bperkins211@...
> [korgpolyex]" <korgpolyex@yahoogroups.com> wrote:
>
>
>
> Don't know why you'd bother with PIC, the toolchain is shite and
> they are electrically "fragile". If you use a low-end ARM chip you
> may need to do something to convert between 5V and 3.3V logic, but
> you'll have enough grunt for all eight voices.
>
> Doing two voices and all four outputs for square and saw ought to be
> doable in a single atmega 328.
>
> Use a PWM output to generate the "analogue" output instead of an R/
> 2R DAC. Use the counter for the PWM to trigger the sample
> calculation at around 32kHz and allow it to run asynchronously from
> the control loop. When the PWM rolls over it'll call an ISR that you
> use to load the comparison register with the calculated sample value.
>
> Pish easy.
>
> --
> Gordonjcp MM0YEQ
>
>
>
> I suggested the PIC since it is the only uC I have any kind of
> experience with and currently studying, for now.. I was trying to
> offer ANY help on this attempt to revoice the HAWK.
>
> Your suggestion sounds nice. I remember you posted something on
> this and some code as well awhile back.
> Is the PWM outputs from it require any filtering at all?
>
> Do you have any docs to point to on using wavetables on a ARM chip?
> Can you maybe lend more detailed advice on this?
> I take it to get 8 voices total, that we'd need two atmega 328's?
>
> I'd like to know more about any chip used for Arduino projects
> anyway...
> I am looking to start building quadcopters and other UAV's using
> some of the open source Flight Controllers available like 3dRobotics
> APM.
>
> ~Blaine
>
>
>
>
>
>
>