MOTM

How feasible is this idea of creating an 'analog delay' using digital RAM?
Is that an oxymoron, or is this a legitimate concept?

The good points of course are that you don't need bucket-brigade device
chips that will go out of production, and extensibility. It sure would be
cool if such a module could be based on, say, PC EDO RAM DIMMs or something,
with lots of extra sockets so that DIMMs could be added later.

Think about it-a few years from now when your Pentium III is a boat anchor,
you can throw the PC out and move the RAM into the delay module to get some
real use out of it!

Hey All -

I know this has been discussed a little in the past, but I'd like to 
bring it up again.

I think it would be very useful to have a simple delay module for my 
MOTM (a resounding *Duh* from the audience).  I know I'm coming late 
to this particular party, but it seems like a good thing for making 
comb filters, phasor and flangy thingys, delaying envelopes, etc.

I'm talkin' really simple here: delay variable from 0 to maybe 5 ms, 
knob for attenuating the mod input, jack for mod input, jack for 
signal input and jack for output.  No feedback needed, no internal 
modualtion.

Wouldn't this be simple to build and highly useful?

Mike

"No feedback needed,...."

I think you would really need feedback!

--
___________________
John Blacet
Blacet Research
http://www.blacet.com

I think it would be amazingly useful, but Paul has stated, on several
occasions, that he does not intend to make such a beast, partially for
the fact that there are others on the market which are plenty good, and
partially due to parts scarcity.

Too bad.  But I'm still hoping he'll reconsider one day...

--PBr

-----Original Message-----
From: mmarsh100 [mailto:mmarsh@stellcom.com]
Sent: Monday, February 25, 2002 4:33 PM
To: motm@yahoogroups.com
Subject: [motm] Delay Module


Hey All -

I know this has been discussed a little in the past, but I'd like to 
bring it up again.

I think it would be very useful to have a simple delay module for my 
MOTM (a resounding *Duh* from the audience).  I know I'm coming late 
to this particular party, but it seems like a good thing for making 
comb filters, phasor and flangy thingys, delaying envelopes, etc.

I'm talkin' really simple here: delay variable from 0 to maybe 5 ms, 
knob for attenuating the mod input, jack for mod input, jack for 
signal input and jack for output.  No feedback needed, no internal 
modualtion.

Wouldn't this be simple to build and highly useful?

Feedback on the suggestion, yes!  On the delay, maybe not.  That's 
what the Time Machine is for!  A 'little' delay would free up a TM to 
do 'bigger' stuff.

Remember, I suggested very small delay times: 0 to 5 ms, or even 2.5 
ms.  Do you still need oodles of bucket-brigade for such short 
delays?  If not, then the chip problem is moot, no?

Mike

--- In motm@y..., John Blacet <blacet@b...> wrote:

> "No feedback needed,...."
> 
> I think you would really need feedback!
> 
> --
> ___________________
> John Blacet
> Blacet Research
> http://www.blacet.com

a) There is not much I can do about the lack of BBD chips

b) That leaves DSP/digital solutions. Which is "allowed" but then get a Boss pedal for $45, rip
guts out, put on blank panel. If I did one, I'd be a $189 kit and everyone would squeal like a
pig.

Paul S.


----- Original Message -----

From: "Brousseau, Paul E (Paul)" <noise@...>
To: <motm@yahoogroups.com>
Sent: Monday, February 25, 2002 6:59 PM
Subject: RE: [motm] Delay Module


> I think it would be amazingly useful, but Paul has stated, on several
> occasions, that he does not intend to make such a beast, partially for
> the fact that there are others on the market which are plenty good, and
> partially due to parts scarcity.
>
> Too bad.  But I'm still hoping he'll reconsider one day...
>
> --PBr
>
> -----Original Message-----
> From: mmarsh100 [mailto:mmarsh@...]
> Sent: Monday, February 25, 2002 4:33 PM
> To: motm@yahoogroups.com
> Subject: [motm] Delay Module
>
>
> Hey All -
>
> I know this has been discussed a little in the past, but I'd like to
> bring it up again.
>
> I think it would be very useful to have a simple delay module for my
> MOTM (a resounding *Duh* from the audience).  I know I'm coming late
> to this particular party, but it seems like a good thing for making
> comb filters, phasor and flangy thingys, delaying envelopes, etc.
>
> I'm talkin' really simple here: delay variable from 0 to maybe 5 ms,
> knob for attenuating the mod input, jack for mod input, jack for
> signal input and jack for output.  No feedback needed, no internal
> modualtion.
>
> Wouldn't this be simple to build and highly useful?
>
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>

Again, coming from a total electronics dumb-bell, with such small 
delay times couldn't you do this with discrete parts?

Mike

--- In motm@y..., "Paul Schreiber" <synth1@a...> wrote:
> a) There is not much I can do about the lack of BBD chips
> 
> b) That leaves DSP/digital solutions. Which is "allowed" but then 
get a Boss pedal for $45, rip
> guts out, put on blank panel. If I did one, I'd be a $189 kit and 
everyone would squeal like a

> pig.
> 
> Paul S.

Paul,

  Just offer a Quantec clone, kit version only.  Use something like 200 
stages and just by parts bulk pre-packed in bags of 200 so you don't have 
to kit them at all.  Then laugh evilly as you think of the people building 
the Quantec clone. ;)

Crow
/**/

I take it from this message that even a small delay in the ms range 
is not feasable?

--- In motm@y..., The Old Crow <oldcrow@o...> wrote:
> 
>   Paul,
> 
>   Just offer a Quantec clone, kit version only.  Use something like 
200 
> stages and just by parts bulk pre-packed in bags of 200 so you 
don't have 
> to kit them at all.  Then laugh evilly as you think of the people 
building 

> the Quantec clone. ;)
> 
> Crow
> /**/

--- In motm@y..., "Paul Schreiber" <synth1@a...> wrote:
> a) There is not much I can do about the lack of BBD chips

I read somewhere that CCD chips might be a viable alternative.
 
> b) That leaves DSP/digital solutions. Which is "allowed" but 
then get a Boss pedal for $45, rip
> guts out, put on blank panel. If I did one, I'd be a $189 kit and 
everyone would squeal like a
> pig.
> 
> Paul S.

But what about voltage control? I use an old Digitech 
TimeMachine 7.8 DDL that has a control voltage input - it's kinda 
crappy sounding, though.

I used to use a Lexicon PCM-42 which also has a CV in - but, 
sold it when the prices got ridiculously high. I think the PCM-41 
has a CV in and those can be had for $300 - $400

On a similar note, I just resurrected my old Moog 12 stage phase 
shifter (CV in & LFO out). Voltage controlled FX are just more fun 
in a patchable context.

Barry

This is kinda too bad because the effect is very nice. I'll post a 
file to the files section that demonstrates one of the effects.

--- In motm@y..., "mmarsh100" <mmarsh@s...> wrote:
> I take it from this message that even a small delay in the ms range 
> is not feasable?
> 
> --- In motm@y..., The Old Crow <oldcrow@o...> wrote:
> > 
> >   Paul,
> > 
> >   Just offer a Quantec clone, kit version only.  Use something 
like 

> 200 
> > stages and just by parts bulk pre-packed in bags of 200 so you 
> don't have 
> > to kit them at all.  Then laugh evilly as you think of the people 
> building 
> > the Quantec clone. ;)
> > 
> > Crow
> > /**/

What's a CCD chip?  I thought those were the light-sensitive things they
used in digital cameras...

--PBr

-----Original Message-----
From: coyoteous [mailto:satori@...]
Sent: Monday, February 25, 2002 5:44 PM
To: motm@yahoogroups.com
Subject: [motm] Re: Delay Module


--- In motm@y..., "Paul Schreiber" <synth1@a...> wrote:
> a) There is not much I can do about the lack of BBD chips

I read somewhere that CCD chips might be a viable alternative.

> Again, coming from a total electronics dumb-bell, with such small
> delay times couldn't you do this with discrete parts?

Not really. You need to have several steps:

1) Conversion and storage means on the incoming audio. A BBD chip is a black of many small
capacitors (like 512 of them). There are 512 analog switches inbetween them. The audio (AC
voltage) is "dumped" on a capacitor.

In digital delays, the audio runs through an ADC (analog-to-digital converter) and the sample is
stored in RAM.

2) Output conversion means from storage to audio. In a BBD, this is "free" because you are
retrieving the original charge (OK with added noise and 'droop'). In a digital delay, that is
done with a DAC (digital-to-analog converter).

3) Clocking means. This "shifts" the sample along in the storage means to get a delay. If you
have 512 BBD stages, and you want 5.12ms of delay, that's...errrr...(looking for slide
rule/MathCAD/abacus) a clock period of 10us or 100Khz. Most BBDs won't operate this fast, which
is why you see "minimum delay" in the 10ms range. In the digital world, we can go
full-tilt-boogie and "cheat" by picking off from ANY RAM location. A BBD has fixed "taps" and so
we are forced to chug through the whole thing.

Paul S.

--- In motm@y..., "Brousseau, Paul E (Paul)" <noise@A...> wrote:
> What's a CCD chip?  I thought those were the light-sensitive 
things they
> used in digital cameras...
> 
> --PBr

Yes, Charge Coupled Device(?) used in cameras - again, I just 
recall someone theorizing it's possible use as an analog delay - 
maybe on SynthDIY or AH.

Barry

--- In motm@y..., "Brousseau, Paul E (Paul)" <noise@A...> wrote:
> What's a CCD chip?  I thought those were the light-sensitive 
things they
> used in digital cameras...
> 
> --PBr

Yes, Charge Coupled Device(?) used in cameras - again, I just 
recall someone theorizing it's possible use as an analog delay - 
maybe on SynthDIY or AH.

Barry

OK, I got it.  Capacitors 'store' charge.  Too bad, the effect is 
nice.

So if I was to buy a Boss delay and MOTMize it, would it be hard to 
match
the power to MOTM?

Mike

--- In motm@y..., "Paul Schreiber" <synth1@a...> wrote:
> 
> 
> > Again, coming from a total electronics dumb-bell, with such small
> > delay times couldn't you do this with discrete parts?
> 
> Not really. You need to have several steps:
> 
> 1) Conversion and storage means on the incoming audio. A BBD chip 
is a black of many small
> capacitors (like 512 of them). There are 512 analog switches 
inbetween them. The audio (AC
> voltage) is "dumped" on a capacitor.
> 
> In digital delays, the audio runs through an ADC (analog-to-digital 
converter) and the sample is
> stored in RAM.
> 
> 2) Output conversion means from storage to audio. In a BBD, this 
is "free" because you are
> retrieving the original charge (OK with added noise and 'droop'). 
In a digital delay, that is
> done with a DAC (digital-to-analog converter).
> 
> 3) Clocking means. This "shifts" the sample along in the storage 
means to get a delay. If you
> have 512 BBD stages, and you want 5.12ms of delay, that's...errrr...
(looking for slide
> rule/MathCAD/abacus) a clock period of 10us or 100Khz. Most BBDs 
won't operate this fast, which
> is why you see "minimum delay" in the 10ms range. In the digital 
world, we can go
> full-tilt-boogie and "cheat" by picking off from ANY RAM location. 
A BBD has fixed "taps" and so

> we are forced to chug through the whole thing.
> 
> Paul S.

Nobody is mentioning the most interesting use for a voltage 
controllable delay with very short delay times - analog physical 
modelling. To do 'tuned resonator' types of blown pipe emulations one 
needs precise voltage control with sub millisecond resolution, along 
with voltage control of feedback. If it has to be digital to achieve 
that much control, so be it - but it will take a purpose-built 
design, not a hacked up pedal.

Check out this article about Ron Berry and his analog acoustic 
modelling techniques, if you are not already familiar with him:
http://www.sospubs.co.uk/sos/1997_articles/jul97/ronberry.html

I've often wished to talk with him in greater depth - I'd dearly love 
to explore this side of synthesis. I wonder what technology he used 
to build his own delays...

Moe

>>>>>>
Feedback on the suggestion, yes! On the delay, maybe not. That's 
what the Time Machine is for! A 'little' delay would free up a TM to 
do 'bigger' stuff.

Remember, I suggested very small delay times: 0 to 5 ms, or even 2.5 
ms. Do you still need oodles of bucket-brigade for such short 
delays? If not, then the chip problem is moot, no?
<<<<<<

Yes!  Very cool, indeed.

Mike

--- In motm@y..., "mate_stubb" <mate_stubb@y...> wrote:
> Nobody is mentioning the most interesting use for a voltage 
> controllable delay with very short delay times - analog physical 
> modelling. To do 'tuned resonator' types of blown pipe emulations 
one 
> needs precise voltage control with sub millisecond resolution, 
along 
> with voltage control of feedback. If it has to be digital to 
achieve 
> that much control, so be it - but it will take a purpose-built 
> design, not a hacked up pedal.
> 
> Check out this article about Ron Berry and his analog acoustic 
> modelling techniques, if you are not already familiar with him:
> http://www.sospubs.co.uk/sos/1997_articles/jul97/ronberry.html
> 
> I've often wished to talk with him in greater depth - I'd dearly 
love 
> to explore this side of synthesis. I wonder what technology he used 
> to build his own delays...
> 
> Moe
> 
> >>>>>>
> Feedback on the suggestion, yes! On the delay, maybe not. That's 
> what the Time Machine is for! A 'little' delay would free up a TM 
to 
> do 'bigger' stuff.
> 
> Remember, I suggested very small delay times: 0 to 5 ms, or even 
2.5 

> ms. Do you still need oodles of bucket-brigade for such short 
> delays? If not, then the chip problem is moot, no?
> <<<<<<

I've been playing around with ideas for delay modules for some time.
I would use a ADC of presumably 16 bits or better on the input with 
a fixed filter at around 20 kHz.  The converted data would go into 
RAM of a fixed length then come back out through a DAC and another 
filter.  Most delay lines vary the point at which the data comes out 
of RAM.  This is why you get zipper noise when varying the delay 
time on most digital delays.  In order to do away with the zipper 
noise you have to vary the clock speed.  You would want to do this 
with voltage control.  I'm not sure if you'd want the frequency 
response of the clock to respond linearly or exponentially or 
logarithmically to the control voltage.  A selection of all three 
would be interesting.  The problem is (and I don't know the answer 
yet) Over what range can you vary the clock rate of the converters 
and RAM?  This is something I'll be playing with in the future if I 
ever get caught up on module building and get started on some of my 
own ideas.  It doesn't sound cheap though.
I agree that you can do a lot of interesting things with short delay 
times.  Although most delays do not go below 1mS you can do some 
great stuff in the 50uS to 1mS range.  One of my favorites is to 
split a mono signal and delay 1 side my a sub ms amount.  Great for 
really strong stereo imaging.
I think you would want the feedback as well.  Ideally, I'd make the 
feedback amount Voltage Controlled with a VCA.
Interesting Stuff.

> Nobody is mentioning the most interesting use for a voltage
> controllable delay with very short delay times - analog physical
> modelling. To do 'tuned resonator' types of blown pipe emulations one
> needs precise voltage control with sub millisecond resolution, along
> with voltage control of feedback. If it has to be digital to achieve
> that much control, so be it - but it will take a purpose-built
> design, not a hacked up pedal.

I thoght about building something like that in the past, but there was
always one think that kept me from really doing it:
The BBD line can be "played in tune" if you have an exponential
clock VCO. But for physical modelling, you need a filter in the
feedback path, don't you ? And every filter has some extra phase shift
(i.e. frequency dependend delay), so it will make the model go out of tune.
In a fully digital implementation you can correct this, but in analogue ?

JH.

The audiophile* version: make it 24 bit and give it a fairly high 
base sampling rate (96 kHz?), so you could sweep down to half 
the base SR and up to double (with 192 kHz capable parts) and 
still use a fixed full audio bandwidth anti-alias filter? Linear 
voltage control and/or exponential for the for the physmo stuff? 
Manual selection of taps would be excellent. You could have 
multiple taps by digital mixing and/or multiple DACs. The same 
basic module could be optimized for longer or shorter delay 
times. Problems: expense,  jitter, RFI - but, could it work?

*From my understanding and experience, this is what it would 
take to get close to high end analog (MOTM quality?) with digital.

Barry

--- In motm@y..., "paulhaneberg" <phaneber@o...> wrote:
> I've been playing around with ideas for delay modules for 
some time.
> I would use a ADC of presumably 16 bits or better on the input 
with 
> a fixed filter at around 20 kHz.  The converted data would go 
into 
> RAM of a fixed length then come back out through a DAC and 
another 
> filter.  Most delay lines vary the point at which the data comes 
out 
> of RAM.  This is why you get zipper noise when varying the 
delay 
> time on most digital delays.  In order to do away with the zipper 
> noise you have to vary the clock speed.  You would want to do 
this 
> with voltage control.  I'm not sure if you'd want the frequency 
> response of the clock to respond linearly or exponentially or 
> logarithmically to the control voltage.  A selection of all three 
> would be interesting.  The problem is (and I don't know the 
answer 
> yet) Over what range can you vary the clock rate of the 
converters 
> and RAM?  This is something I'll be playing with in the future if I 
> ever get caught up on module building and get started on 
some of my 
> own ideas.  It doesn't sound cheap though.
> I agree that you can do a lot of interesting things with short 
delay 
> times.  Although most delays do not go below 1mS you can do 
some 
> great stuff in the 50uS to 1mS range.  One of my favorites is to 
> split a mono signal and delay 1 side my a sub ms amount.  
Great for 
> really strong stereo imaging.
> I think you would want the feedback as well.  Ideally, I'd make 
the 
> feedback amount Voltage Controlled with a VCA.
> Interesting Stuff.

24 bit would certainly have improved sonic quality over 16 bit, but I 
think you could probably get a bigger variation in clock speeds with a 
16 bit.  If the converter is less complex the actual conversion would 
be quicker assuming a succesive-approximation type converter.  I don't 
know what the highest number of bits available in a flash converter 
is, but presumably this could be clocked at higher speeds (up into the 
MHz range)  There are 24 bit converters available that can convert at 
192 kHz, but cost is an issue.  Another possibility is to use a 
one-bit converter at even higher clock speeds.
I would definitely use a filter at a low fixed frequency like 20kHz.
One of the pitfalls of this module ideas is that in order to be truly 
musically useful the voltage to frequency conversion for the clock 
must be very accurate over a wide range.  I'm not sure how possible 
that is in the MHz region.
Multiple taps are easily implemented.  A range of delay times from 
very short to reasonably long could be available.  This is simply a 
function of the amount of RAM.
Even 12 bits would probably be an improvement over an analog delay.
I think a module of this sort would be very useful and popular as long 
as the cost was not too high.  
At some point I will probably try to build this thing and find out.

24 bit 192kHz is a real misnomer. I can't think of any affordable 24bit A/D
that goes anywhere near that speed. "Real" converters that can put out a
real 24bits are only good for a few kHz (the last time I looked)

Audio "type" converters have a 24bit output word, but typically about 5 of
those bits are marketing bits. Look at the SNR and DR of the ADC for the
real story.

I don't think you'll have any luck VCing (by varying the clock) a
delta-sigma converter anyway. They almost all have internal PLLs that
wouldn't like a radically changing clock and many of them have async timing
inside that is very dependant on specific external frequencies. They figure
out what the external freq is and make timing adjustments.

If you want to use variable clocks, you're better off doing it the old
fashioned way (think Deltalab gear)

Tony

> -----Original Message-----
> From: paulhaneberg [mailto:phaneber@...]
> Sent: Tuesday, February 26, 2002 6:42 PM
> To: motm@yahoogroups.com
> Subject: [motm] Re: Delay Module
>
>
> 24 bit would certainly have improved sonic quality over 16 bit, but I
> think you could probably get a bigger variation in clock speeds with a
> 16 bit.  If the converter is less complex the actual conversion would
> be quicker assuming a succesive-approximation type converter.  I don't
> know what the highest number of bits available in a flash converter
> is, but presumably this could be clocked at higher speeds (up into the
> MHz range)  There are 24 bit converters available that can convert at
> 192 kHz, but cost is an issue.  Another possibility is to use a
> one-bit converter at even higher clock speeds.
> I would definitely use a filter at a low fixed frequency like 20kHz.
> One of the pitfalls of this module ideas is that in order to be truly
> musically useful the voltage to frequency conversion for the clock
> must be very accurate over a wide range.  I'm not sure how possible
> that is in the MHz region.
> Multiple taps are easily implemented.  A range of delay times from
> very short to reasonably long could be available.  This is simply a
> function of the amount of RAM.
> Even 12 bits would probably be an improvement over an analog delay.
> I think a module of this sort would be very useful and popular as long
> as the cost was not too high.
> At some point I will probably try to build this thing and find out.
>
>
>
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>

That's why I said "expensive."<g> But, 18/20/24 bit converters - even 
a crappy midiman Flying Cow, are for the most part superior to a 
crappy 16 bit converter. At my day job, I regularly brutalize 16 bit 
recordings to make them into "realy loud" 14-15 bit recordings. About 
10% of the masters I handle are mixed to a resolution greater than 16 
bit - I can almost always make these sound better than their 16 bit 
counterparts. The reason - "the most misunderstood of all bits" - #17. 
This guy lets us turn a 17-18 bit recording into a "really loud" 16 
bit recording. Also, even on on a cheap "24 bit" converter, there is 
some "good" signal at and under the noise floor, just as there is with 
analog tape.

That said, you're right - it's probably not worth the expense/effort. 
Lest I take us even further off topic (MOTM - right?), my real 
interest in all of this is in efficient, cheap and high definition 
exchange of information between the analog and digital worlds, 
particularly control signals. At this point, we're kind of stuck with 
midi or convoluted "ghost electronics." I'd like to have digital 
processing/generation under analog control and analog processing/
generation under digital control. How's about a MOTM or Encore "ghost 
electronics" translation module? It's kind of a natural extension of 
(or off shoot from) the midi<->cv thing and I bet it would sell at 
least as good as a triple reversible attenuator (or maybe not). <g>

Barry


--- In motm@y..., "Tony Karavidas" <tony@e...> wrote:

> 24 bit 192kHz is a real misnomer. I can't think of any affordable 24bit A/D
> that goes anywhere near that speed. "Real" converters that can put out a
> real 24bits are only good for a few kHz (the last time I looked)
> 
> Audio "type" converters have a 24bit output word, but typically about 5 of
> those bits are marketing bits. Look at the SNR and DR of the ADC for the
> real story.
> 
> I don't think you'll have any luck VCing (by varying the clock) a
> delta-sigma converter anyway. They almost all have internal PLLs that
> wouldn't like a radically changing clock and many of them have async timing
> inside that is very dependant on specific external frequencies. They figure
> out what the external freq is and make timing adjustments.
> 
> If you want to use variable clocks, you're better off doing it the old
> fashioned way (think Deltalab gear)
> 
> Tony
> 
> 
> 
> 
> 
> > -----Original Message-----
> > From: paulhaneberg [mailto:phaneber@o...]
> > Sent: Tuesday, February 26, 2002 6:42 PM
> > To: motm@y...
> > Subject: [motm] Re: Delay Module
> >
> >
> > 24 bit would certainly have improved sonic quality over 16 bit, but I
> > think you could probably get a bigger variation in clock speeds with a
> > 16 bit.  If the converter is less complex the actual conversion would
> > be quicker assuming a succesive-approximation type converter.  I don't
> > know what the highest number of bits available in a flash converter
> > is, but presumably this could be clocked at higher speeds (up into the
> > MHz range)  There are 24 bit converters available that can convert at
> > 192 kHz, but cost is an issue.  Another possibility is to use a
> > one-bit converter at even higher clock speeds.
> > I would definitely use a filter at a low fixed frequency like 20kHz.
> > One of the pitfalls of this module ideas is that in order to be truly
> > musically useful the voltage to frequency conversion for the clock
> > must be very accurate over a wide range.  I'm not sure how possible
> > that is in the MHz region.
> > Multiple taps are easily implemented.  A range of delay times from
> > very short to reasonably long could be available.  This is simply a
> > function of the amount of RAM.
> > Even 12 bits would probably be an improvement over an analog delay.
> > I think a module of this sort would be very useful and popular as long
> > as the cost was not too high.
> > At some point I will probably try to build this thing and find out.
> >
> >
> >
> >
> >
> > Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
> >
> >

>The audiophile* version: make it 24 bit and give it a fairly high
>base sampling rate (96 kHz?), so you could sweep down to half
>the base SR and up to double (with 192 kHz capable parts)

That sounds like overkill to me.  Based on research papers that I've read,
and my own ears, 24/96 conversion doesn't sound any better than 24/44.1
conversion.  Higher sampling rates only seem to matter when mixing or
altering sounds within the digital realm itself.  Regardless, I'd take the
Pepsi challenge with my 18/48 PCM-80 against any of these consumer 24/96
delays any day of the week.

>24 bit 192kHz is a real misnomer. I can't think of any affordable 24bit A/D
>that goes anywhere near that speed. "Real" converters that can put out a
>real 24bits are only good for a few kHz (the last time I looked)
>
>Audio "type" converters have a 24bit output word, but typically about 5 of
>those bits are marketing bits. Look at the SNR and DR of the ADC for the
>real story.

This is true.  There are many 16-bit converters that sound better than
24-bit converters.  While there is a noticeable difference between 16-bit
and 20-bit, the difference between 20-bit and 24-bit seems insignificant.

>In order to do away with the zipper noise you have to vary the clock
>speed.  >You would want to do this with voltage control.

I believe this is how the Korg SDD-2000 works, but it doesn't have voltage
control.

>But what about voltage control?

I have five digital delays, and four of them have voltage control.  It's a
very common feature.

However, none of them can pass DC, which is what I would want in an motm
digital delay module.

>b) That leaves DSP/digital solutions. Which is "allowed" but then get a
>>Boss pedal for $45, rip guts out, put on blank panel. If I did one, I'd
>be >a $189 kit and everyone would squeal like a pig.

Could I get a sample of that sound?? :)

> That sounds like overkill to me.  Based on research papers 
that I've read,
> and my own ears, 24/96 conversion doesn't sound any better 
than 24/44.1

I tend to agree, but that wasn't my point. The idea was to get 
20khz bandwidth and the bottom end of the control range - 192 at 
the top, 96 in the middle, 48 at the bottom.

> This is true.  There are many 16-bit converters that sound 
better than
> 24-bit converters.  While there is a noticeable difference 
between 16-bit
> and 20-bit, the difference between 20-bit and 24-bit seems 
insignificant.

Converters tend to come in increments of 8 bits these days. 

> I have five digital delays, and four of them have voltage control.  
It's a
> very common feature.

Can you name one that is still in production?

> However, none of them can pass DC, which is what I would 
want in an motm
> digital delay module.

Now you're talkin', I suppose if it were for cv range only it could be 
pretty cheap to make. Otherwise, that's where the "ghost 
electronics" come in.

Barry