Reply
 
Thread Tools Display Modes
  #1   Report Post  
Richard Freeman
 
Posts: n/a
Default does this box make sense electronically? (Folcrom)


"xy" wrote in message
om...
http://www.rollmusic.com/systems/folcrom.shtml

I'm not smart enough with electronics to understand if it's a good
ciruit idea or not. It seems cool. But what do I know!

I was hoping some of the circuit-gods around here could offer any
thoughts on it.


Well as described it should work

For example, I'm wondering how something completely passive like that
can maintain proper gain staging,


Well it doesnt. let me quote "requires approximately 30-40dB of make-up
gain" in other words because it is a passive mixer it effectively adds up to
30-40dB of loss meaning it needs to be followed by a pre-amp

and how it could stay quiet and


No active components = bugger all noise

reject rf.


It is shielded


But again, I'm bad at electronics!


On the other hand I just looked at the Price $1000 USD !!!!!!! It would be
expensive at 1/4 that price !

Regards
Richard Freeman


  #2   Report Post  
ThePaulThomas
 
Posts: n/a
Default

Arjan P wrote in message ...
What a joke.

Actually, it's not a joke.
  #3   Report Post  
Paul Stamler
 
Posts: n/a
Default

Electronically it makes sense, with one exception, which I'll get to in a
moment. But unless they're using very high-quality switches for 20 years of
reliability, it's way the hell overpriced. A prosumer version of this could
go for perhaps $250, a higher-quality one (again, that mostly means better
switches) for $450-500.

Like I said, electronically it makes sense; the gain staging should work
okay. If the nominal output level is -35dBu, and the nominal output
impedance is 150 ohms, that'll work just fine with a typical microphone
preamp. If the latter has an input equivalent noise level of -127dBu, then
you'll get a signal-to-noise ratio of 92dB, relative to nominal level, and
however much headroom your converters allow.

The fly in the ointment is the crosstalk spec and the switching. They say
it's -90 dB, but they don't say how that number's derived. My guess is that
it's the figure for how much leaks into the left channel when you're driving
the right channel, or vice versa. All very well and good, if that's all
that's punched in.

But they also give you the facility to swtich an input to both channels at
once, and that will degrade the separation some. I calculate that pushing in
one pair of buttons to switch that channel to the center makes the
separation more like -76 dB, and pushing in more pairs degrades it further.
Maybe the distinction's academic, but it should be noted.

You could build a switchbox like this yourself, and if you remember to only
switch inputs to left or right, not both, getting center stuff by panning,
you could get very nice results, and use up a wad of solder. Of course, then
your results will depend on the quality of your D/A converters.

Peace,
Paul


  #4   Report Post  
Justin Ulysses Morse
 
Posts: n/a
Default

Since I designed the Folcrom and I'm selling it, I don't expect anybody
to take my word for it that it's a good idea. The people who have been
buying it seem to think it's a good idea, as we've gotten a
surprisingly strong and positive response so far. I'll comment on some
of the commentary below.

Paul Stamler wrote:

Electronically it makes sense, with one exception, which I'll get to in a
moment. But unless they're using very high-quality switches for 20 years of
reliability, it's way the hell overpriced. A prosumer version of this could
go for perhaps $250, a higher-quality one (again, that mostly means better
switches) for $450-500.


I've had to defend the price of the Folcrom on several occasions when
DIY types suggest it's nothing more than a box of resistors and that it
should cost some ridiculously paltry amount. I could have thrown in
three dollars' worth of op-amps and a wall wart and it wouldn't be
passive, and nobody would complain about the price. The fact is that
the bulk of the parts cost lies in the chassis, the switches, and the
circuitboards. The difference between a passive circuit and an active
one does not amount to a meaningful percentage of the selling price.
The parts cost only accounts for about a quarter of the selling price,
which is typical for any gear you buy. Aside from parts cost, there's
also dealer mark-up, distribution costs, assembly labor, and facility
overhead to account for the majority of the cost. I'm not quite
getting rich off this thing. The dealers I'm working with and my
partners seem to think the price is a bit too low, and in fact I think
it's right about where it should be. In truth a person who's handy
with a soldering iron and friendly with a metal shop could build one
for about $300 if they soldered everything point-to-point, but that
would take a hell of a long time. The circuitboard design would be a
hurdle for most DIY folks, which makes the selling price of my box seem
pretty reasonable. Is it worth 30 hours of your time to save $400?
The Folcrom is sturdy and well-built, it looks good, and comes with a
warranty.

Like I said, electronically it makes sense; the gain staging should work
okay. If the nominal output level is -35dBu, and the nominal output
impedance is 150 ohms, that'll work just fine with a typical microphone
preamp. If the latter has an input equivalent noise level of -127dBu, then
you'll get a signal-to-noise ratio of 92dB, relative to nominal level, and
however much headroom your converters allow.


Exactly. Many folks don't immediately realize that any normal, active
mixer has a make-up gain amplifier built into it. The noise
performance of the Folcrom coupled with a good low-noise preamp is no
worse than a good quiet mixer with built-in makeup gain. In fact it is
even quieter because we left out all kinds of circuitry that's needed
in a traditional mixer but unnecessary in a DAW-summing application.

The fly in the ointment is the crosstalk spec and the switching. They say
it's -90 dB, but they don't say how that number's derived. My guess is that
it's the figure for how much leaks into the left channel when you're driving
the right channel, or vice versa. All very well and good, if that's all
that's punched in.

But they also give you the facility to swtich an input to both channels at
once, and that will degrade the separation some. I calculate that pushing in
one pair of buttons to switch that channel to the center makes the
separation more like -76 dB, and pushing in more pairs degrades it further.
Maybe the distinction's academic, but it should be noted.


It's actually very difficult to pin down an accurate, meaningful
specification for either crosstalk or noise. The measurements will
always depend on what's feeding the Folcrom and what the Folcrom is
feeding. The crosstalk specification we give (-90dB) is sort of a
best-case scenario for the real world. Paul's right that assigning
numerous channels to mono (both left and right output) will degrade the
crosstalk performance somewhat, but that degradation will be very
slight when the Folcrom is fed by a high quality source. In this
context I'm defining "high quality" as a very low-impedance source. A
theoretically ideal source, having zero impedance, would not degrade
the crosstalk whatsoever. Any high quality DAC ought to have a source
impedance of under 100 ohms, which will have a negligible impact on the
crosstalk even if several channels are bussed mono. I would put the
worst-case scenario, barring any unusually high-impedance sources, at
about -70dB.

Let's also remember that this is a mixer we're talking about. Its
output is a stereo mix that you're going to listen to on a pair of
speakers. Crosstalk between the left and right channels is about the
most unimportant specification I can think of. What's the separation
of an LP being played by a good cartridge through a pair of speakers
placed 10 feet apart? In reality, the potential "degradation" of the
crosstalk Paul describes will be absolutely undetectable to even the
most critical listener.

You could build a switchbox like this yourself, and if you remember to only
switch inputs to left or right, not both, getting center stuff by panning,
you could get very nice results, and use up a wad of solder. Of course, then
your results will depend on the quality of your D/A converters.


You absolutely could build a box like this yourself for less than half
of what you can buy mine for. But it'll be a lot of work. And if you
leave out the switches so that you can't assign an input to more than
one channel, nor mute an unused channel, then you'll shave the cost of
the actual mixer substantially. But then you'll need twice as many DAC
channels to feed your mono signals, and you'll have to build dummy
plugs to short any input that's not in use. Those items will cost
substantially more than the price difference between a stripped-down
DIY version and the store-bought version.

In summary, the Folcrom is actually a pretty good idea. We spent a
long time developing and refining the design, and in practice the
device works very well. Users have been thrilled with the ability to
change the sound of their mix buss by switching make-up gain
amplifiers. Its price reflects its production costs and is on the low
side of industry norms.
  #5   Report Post  
agent86
 
Posts: n/a
Default

What I don't understand about this type of box (both this one & the active
ones made by other companies) is how they can actually accomplish what the
claim to. I've read claims that they "avoid the resolution loss that
occurs when lowering gain digitally" (paraphrase). But if they don't
include pots or faders, you still have to controll the gain digitally. Is
there some "Magic" in running what for all practical purposes is a digital
mix through an analog summing bus? Or is it pretty much the same as mixing
down digital tracks to analog tape?



  #6   Report Post  
xy
 
Posts: n/a
Default

Thanks for the compliment, but I stink at electronics! It baffles me.
For some unexplainable reason, I do have "skill" at soldering. But I
have no idea what I'm wiring up. I need to be told, "do this do
that".


You could build a switchbox like this yourself, and if you remember to only
switch inputs to left or right, not both, getting center stuff by panning,
you could get very nice results, and use up a wad of solder. Of course, then
your results will depend on the quality of your D/A converters.

Peace,
Paul

  #7   Report Post  
xy
 
Posts: n/a
Default

and thanks for all the insights guys. Paul, I appreciate your
thoughts on the crosstalk and noise. I'm really anti-noise and
crosstalk. It's like kryptonite to me. So thanks for the warning!
  #8   Report Post  
Rob Reedijk
 
Posts: n/a
Default

Justin, I agree with almost everything you say. (Although, I believe
a point-to-point version will actually improve it a little further).

I don't think you are charging too much at all. In fact, your basic
concept already exists, but in a more pricey form: The Millennia
Media Mixing Suite. If you were to price out the equivalent in
the 'Suite with passive channels installed (there are actually two
levels of "passivity"), it would end up being a fair bit more than
the Folcrom. However, then you would have the gain section built in
with various other master section features such as headphone outputs,
oscillator etc. The Suite would be about $3 to 4 thousand in the
equivalent form. It is probably a little higher in quality, but
still at $1000, I think you have a very nice price.

Rob R.

Justin Ulysses Morse wrote:
Since I designed the Folcrom and I'm selling it, I don't expect anybody
to take my word for it that it's a good idea. The people who have been
buying it seem to think it's a good idea, as we've gotten a
surprisingly strong and positive response so far. I'll comment on some
of the commentary below.


Paul Stamler wrote:


Electronically it makes sense, with one exception, which I'll get to in a
moment. But unless they're using very high-quality switches for 20 years of
reliability, it's way the hell overpriced. A prosumer version of this could
go for perhaps $250, a higher-quality one (again, that mostly means better
switches) for $450-500.


I've had to defend the price of the Folcrom on several occasions when
DIY types suggest it's nothing more than a box of resistors and that it
should cost some ridiculously paltry amount. I could have thrown in
three dollars' worth of op-amps and a wall wart and it wouldn't be
passive, and nobody would complain about the price. The fact is that
the bulk of the parts cost lies in the chassis, the switches, and the
circuitboards. The difference between a passive circuit and an active
one does not amount to a meaningful percentage of the selling price.
The parts cost only accounts for about a quarter of the selling price,
which is typical for any gear you buy. Aside from parts cost, there's
also dealer mark-up, distribution costs, assembly labor, and facility
overhead to account for the majority of the cost. I'm not quite
getting rich off this thing. The dealers I'm working with and my
partners seem to think the price is a bit too low, and in fact I think
it's right about where it should be. In truth a person who's handy
with a soldering iron and friendly with a metal shop could build one
for about $300 if they soldered everything point-to-point, but that
would take a hell of a long time. The circuitboard design would be a
hurdle for most DIY folks, which makes the selling price of my box seem
pretty reasonable. Is it worth 30 hours of your time to save $400?
The Folcrom is sturdy and well-built, it looks good, and comes with a
warranty.


Like I said, electronically it makes sense; the gain staging should work
okay. If the nominal output level is -35dBu, and the nominal output
impedance is 150 ohms, that'll work just fine with a typical microphone
preamp. If the latter has an input equivalent noise level of -127dBu, then
you'll get a signal-to-noise ratio of 92dB, relative to nominal level, and
however much headroom your converters allow.


Exactly. Many folks don't immediately realize that any normal, active
mixer has a make-up gain amplifier built into it. The noise
performance of the Folcrom coupled with a good low-noise preamp is no
worse than a good quiet mixer with built-in makeup gain. In fact it is
even quieter because we left out all kinds of circuitry that's needed
in a traditional mixer but unnecessary in a DAW-summing application.


The fly in the ointment is the crosstalk spec and the switching. They say
it's -90 dB, but they don't say how that number's derived. My guess is that
it's the figure for how much leaks into the left channel when you're driving
the right channel, or vice versa. All very well and good, if that's all
that's punched in.

But they also give you the facility to swtich an input to both channels at
once, and that will degrade the separation some. I calculate that pushing in
one pair of buttons to switch that channel to the center makes the
separation more like -76 dB, and pushing in more pairs degrades it further.
Maybe the distinction's academic, but it should be noted.


It's actually very difficult to pin down an accurate, meaningful
specification for either crosstalk or noise. The measurements will
always depend on what's feeding the Folcrom and what the Folcrom is
feeding. The crosstalk specification we give (-90dB) is sort of a
best-case scenario for the real world. Paul's right that assigning
numerous channels to mono (both left and right output) will degrade the
crosstalk performance somewhat, but that degradation will be very
slight when the Folcrom is fed by a high quality source. In this
context I'm defining "high quality" as a very low-impedance source. A
theoretically ideal source, having zero impedance, would not degrade
the crosstalk whatsoever. Any high quality DAC ought to have a source
impedance of under 100 ohms, which will have a negligible impact on the
crosstalk even if several channels are bussed mono. I would put the
worst-case scenario, barring any unusually high-impedance sources, at
about -70dB.


Let's also remember that this is a mixer we're talking about. Its
output is a stereo mix that you're going to listen to on a pair of
speakers. Crosstalk between the left and right channels is about the
most unimportant specification I can think of. What's the separation
of an LP being played by a good cartridge through a pair of speakers
placed 10 feet apart? In reality, the potential "degradation" of the
crosstalk Paul describes will be absolutely undetectable to even the
most critical listener.


You could build a switchbox like this yourself, and if you remember to only
switch inputs to left or right, not both, getting center stuff by panning,
you could get very nice results, and use up a wad of solder. Of course, then
your results will depend on the quality of your D/A converters.


You absolutely could build a box like this yourself for less than half
of what you can buy mine for. But it'll be a lot of work. And if you
leave out the switches so that you can't assign an input to more than
one channel, nor mute an unused channel, then you'll shave the cost of
the actual mixer substantially. But then you'll need twice as many DAC
channels to feed your mono signals, and you'll have to build dummy
plugs to short any input that's not in use. Those items will cost
substantially more than the price difference between a stripped-down
DIY version and the store-bought version.


In summary, the Folcrom is actually a pretty good idea. We spent a
long time developing and refining the design, and in practice the
device works very well. Users have been thrilled with the ability to
change the sound of their mix buss by switching make-up gain
amplifiers. Its price reflects its production costs and is on the low
side of industry norms.


  #9   Report Post  
Arny Krueger
 
Posts: n/a
Default

agent86 wrote:

What I don't understand about this type of box (both this one & the
active ones made by other companies) is how they can actually
accomplish what the claim to. I've read claims that they "avoid the
resolution loss that occurs when lowering gain digitally"
(paraphrase).


The basic claim that there is a resolution loss when gain is lowered
digitally ignores all of the DAW software that either does summing in 64 bit
fixed point, or 32 bit floating point. 64 bits gives something like 300 dB
dynamic range, and 32 bit floating point gives about 1,000 dB dynamic range.
Wither is obviously grotesque overkill, but they are the next logical step
up from the 140 or so dB of dynamic range you get with 24 bit fixed point,
minus the losses inherent in mixing.

Paul Stamler's post estimated a best case SNR of about 92 dB, which pales
compared to the 130 dB dynamic range in much of the DAW software that is
supposedly being improved upon.

But if they don't include pots or faders, you still
have to control the gain digitally. Is there some "Magic" in
running what for all practical purposes is a digital mix through an
analog summing bus?


The only magic I see is along the lines of "Look over there, cakes!"

Or is it pretty much the same as mixing down digital tracks to analog

tape?

If the discussion is about dynamic range, don't go anywhere near analog
tape!


  #11   Report Post  
S O'Neill
 
Posts: n/a
Default

Justin Ulysses Morse wrote:

Since I designed the Folcrom and I'm selling it, I don't expect anybody
to take my word for it that it's a good idea. The people who have been
buying it seem to think it's a good idea, as we've gotten a
surprisingly strong and positive response so far. I'll comment on some
of the commentary below.


So what's the maximum input level? 42 dB-what?


  #12   Report Post  
agent86
 
Posts: n/a
Default

Arny Krueger wrote:

agent86 wrote:

What I don't understand about this type of box (both this one & the
active ones made by other companies) is how they can actually
accomplish what the claim to. I've read claims that they "avoid the
resolution loss that occurs when lowering gain digitally"
(paraphrase).


The basic claim that there is a resolution loss when gain is lowered
digitally ignores all of the DAW software that either does summing in 64
bit
fixed point, or 32 bit floating point. 64 bits gives something like 300
dB dynamic range, and 32 bit floating point gives about 1,000 dB dynamic
range. Wither is obviously grotesque overkill, but they are the next
logical step up from the 140 or so dB of dynamic range you get with 24 bit
fixed point, minus the losses inherent in mixing.

Paul Stamler's post estimated a best case SNR of about 92 dB, which pales
compared to the 130 dB dynamic range in much of the DAW software that is
supposedly being improved upon.

But if they don't include pots or faders, you still
have to control the gain digitally. Is there some "Magic" in
running what for all practical purposes is a digital mix through an
analog summing bus?


The only magic I see is along the lines of "Look over there, cakes!"

Or is it pretty much the same as mixing down digital tracks to analog

tape?

If the discussion is about dynamic range, don't go anywhere near analog
tape!



I'm inclined to agree with you're point of view, Arny. But even if you buy
the argument about digital resolution vs decreasing gain, these boxes don't
even address that. All the gain adjustments STILL happen in the DAW. An
analog summing bus without faders or pot is NOT an analog mixer AFAIK.

If the argument were that the mathematical calculations within the digital
summing bus were themselves inherently flawed, I could buy that. (Or I
could at least consider it a point worthy of discussion.) But that would
also raise the question of why one would expect the mathematical
calculations within the DA & AD converters to be any less flawed. Then you
have to decide whether any potential benefit gained by summing in analog
exceded the potential loss associated with two extra generations of
conversion.

It's a damn wonder anybody finds time to actually record anything.

  #13   Report Post  
Justin Ulysses Morse
 
Posts: n/a
Default

S O'Neill wrote:

So what's the maximum input level? 42 dB-what?


It's pretty arbitrary, really. There's nothing in the Folcrom that can
be overloaded, but if you put enough signal in then eventually the 1/4W
resistors will overheat. I calculated that to be an input signal of
100.75 volts RMS on any single input channel. I believe that's
42.28dBv.

However, if all 16 inputs were driven with equal and fully-correlated
(identical) signals and all fed to the same (left or right) output
buss, then the maximum level for each channel would be 76.49VRMS before
the summing (output) resistors would overheat. I think that's
+39.89dBv. Obviously none of things are ever going to happen in the
real world, so there's not really a need for a MOL specification other
than an academic one.


Rob Reedijk wrote:

It is probably a little higher in quality, but still at $1000, I think you have a very nice price.


Thanks for the support. And actually, the street price of the Folcrom
is $795.

ulysses
  #14   Report Post  
Justin Ulysses Morse
 
Posts: n/a
Default

agent86 wrote:

If the argument were that the mathematical calculations within the digital
summing bus were themselves inherently flawed, I could buy that. (Or I
could at least consider it a point worthy of discussion.)


This is the argument for analog summing. We know that digital summing
should theoretically be perfect but so far all of the claims that
real-world software is in fact perfect seem to grossly underestimate
some software companies' ability to screw it up. That's my current
view of the situation, anyway. Unfortunately this is one of the most
difficult areas to perform conclusive, objective tests because there
are just so many variables.

But that would also raise the question of why one would expect the
mathematical calculations within the DA & AD converters to be any
less flawed.


The DACs themselves simply aren't asked to perform this particular kind
of calculation.

Then you have to decide whether any potential benefit
gained by summing in analog exceded the potential loss associated
with two extra generations of conversion.


And that's where the real meat lies. People who have tried analog
summing have overwhelmingly concluded that they like the effect it has
on the sound of their mixes and their approach to mixing. That's what
really matters to most engineers I know.

It's a damn wonder anybody finds time to actually record anything.


When's the last time I posted about a recording project of my own?

ulysses
  #15   Report Post  
Arny Krueger
 
Posts: n/a
Default

agent86 wrote:
Arny Krueger wrote:

agent86 wrote:

What I don't understand about this type of box (both this one & the
active ones made by other companies) is how they can actually
accomplish what the claim to. I've read claims that they "avoid the
resolution loss that occurs when lowering gain digitally"
(paraphrase).


The basic claim that there is a resolution loss when gain is lowered
digitally ignores all of the DAW software that either does summing
in 64 bit
fixed point, or 32 bit floating point. 64 bits gives something like
300 dB dynamic range, and 32 bit floating point gives about 1,000 dB
dynamic range. Wither is obviously grotesque overkill, but they are
the next logical step up from the 140 or so dB of dynamic range you
get with 24 bit fixed point, minus the losses inherent in mixing.

Paul Stamler's post estimated a best case SNR of about 92 dB, which
pales compared to the 130 dB dynamic range in much of the DAW
software that is supposedly being improved upon.

But if they don't include pots or faders, you still
have to control the gain digitally. Is there some "Magic" in
running what for all practical purposes is a digital mix through an
analog summing bus?


The only magic I see is along the lines of "Look over there, cakes!"

Or is it pretty much the same as mixing down digital tracks to
analog tape?


If the discussion is about dynamic range, don't go anywhere near
analog tape!



I'm inclined to agree with you're point of view, Arny. But even if
you buy the argument about digital resolution vs decreasing gain,
these boxes don't even address that. All the gain adjustments STILL
happen in the DAW. An analog summing bus without faders or pot is
NOT an analog mixer AFAIK.


Agreed.

If the argument were that the mathematical calculations within the
digital summing bus were themselves inherently flawed, I could buy
that. (Or I could at least consider it a point worthy of discussion.)
But that would also raise the question of why one would expect the
mathematical calculations within the DA & AD converters to be any
less flawed.


Different products, different vendors, different development cycles,
hardware versus software.

But, none of these factors would apply to every implmentation, just some of
them.

Then you have to decide whether any potential benefit
gained by summing in analog exceded the potential loss associated
with two extra generations of conversion.


A digital mixer is a picture of simplicity compared to a converter. For
example a digital mixer need not have any brick wall filters or
frequency-dependent processing.

It's a damn wonder anybody finds time to actually record anything.


It's easy, just stick to business.

If obsession strikes, just do a PCABX DBT.




  #16   Report Post  
agent86
 
Posts: n/a
Default

Arny Krueger wrote:


If obsession strikes, just do a PCABX DBT.


Somehow, I KNEW you were going to say that ;-)

  #17   Report Post  
Ben Bradley
 
Posts: n/a
Default

On Thu, 06 May 2004 13:40:48 -0400, agent86
wrote:

I'm inclined to agree with you're point of view, Arny. But even if you buy
the argument about digital resolution vs decreasing gain, these boxes don't
even address that. All the gain adjustments STILL happen in the DAW. An
analog summing bus without faders or pot is NOT an analog mixer AFAIK.


Ironically, changing the gain in 'digital' is a multiplication
function, and that increases the word length and requires dither to go
back to the original word length, and all that.
Summation (the calculation replaced by this box of resistors) is
literally adding numbers in a processor, and as long as you keep up
with overflows (which may require double precision integers, but
that's not a lot of overhead), the result is always 'exact'.

I suppose the reason these things are made is because people can
hear the difference, and many people like what comes out of the
resistor box better than what comes out of a digitally summed bus. I
hesistate to even post in this thread, because I don't think I can
hear the difference myself, and that seems to be the big criterion
here on RAP.

Perhaps the difference is in two D/A converters (giving the exact
same distortion to everything) vs. 8 or 16 D/A converters (each one
does something slightly different to the signal its carrying), but
that's just a hypothesis.

-----
http://mindspring.com/~benbradley
  #18   Report Post  
agent86
 
Posts: n/a
Default

Ben Bradley wrote:

On Thu, 06 May 2004 13:40:48 -0400, agent86
wrote:

I'm inclined to agree with you're point of view, Arny. But even if you
buy the argument about digital resolution vs decreasing gain, these boxes
don't
even address that. All the gain adjustments STILL happen in the DAW. An
analog summing bus without faders or pot is NOT an analog mixer AFAIK.


Ironically, changing the gain in 'digital' is a multiplication
function, and that increases the word length and requires dither to go
back to the original word length, and all that.
Summation (the calculation replaced by this box of resistors) is
literally adding numbers in a processor, and as long as you keep up
with overflows (which may require double precision integers, but
that's not a lot of overhead), the result is always 'exact'.

I suppose the reason these things are made is because people can
hear the difference, and many people like what comes out of the
resistor box better than what comes out of a digitally summed bus. I
hesistate to even post in this thread, because I don't think I can
hear the difference myself, and that seems to be the big criterion
here on RAP.



You can hear a difference in most anything if you pay attention. That's
why we get all the discussions about audio-phool cables & such. But
hearing a difference doesn't mean you can objectively say one is BETTER
than another.

I usually mix in analog, just because an analog mixer FEELS more natural to
me than a mouse (or even a digital mixer. I already know without thinking
about it, what all the knobs & faders & pushbuttons do, & I don't have to
spend a lot of time thinking about it. Granted, most of the stuff I'm
doing is acoustin & I seldom run more than 8 tracks & not a lot of effects.
On the rare occasion that I have enough going on that I think automation
would be handy, I do it in the computer. If I ever bothered to do an
analog & a digital mix of the exact same program material & do one of
Arny's ABX tests on it I suspect I would hear a difference. But I doubt it
would be enough of a difference for me to worry about. Life's just too
short to obsess about that kind of ****. It's going to end up on a 44.1/16
CD anyway. And it's probably going to get played back on some piece of
crap Sony all-in-one through Bose speakers.

I'm no Scott Dorsey, but I'm pretty sure my mixing room & monitor setup is
WAY more accurate than ANYthing my listeners are playing my stuff back on.
If it sound good in my mix room & it also sounds good in my living room,
then kife's too short to worry about whether it might sound a tiny little
bit better if I used a different summing bus. The only thing left to do is
see how it sounds in the car, & that's a good enough excuse to grab the ol'
fly rod & head for the nearest trout stream.

  #19   Report Post  
Arny Krueger
 
Posts: n/a
Default

agent86 wrote:
Arny Krueger wrote:


If obsession strikes, just do a PCABX DBT.


Somehow, I KNEW you were going to say that ;-)


Well, that's why we came up with ABX in the first place. People would read
stuff in TAS or someplace and get really obsessed with some new purported
*improvement*. Some old grizzled veteran would say "That's a piece of crap",
and then we'd do a DBT to see who was right. Take a look at this:

http://www.pcavtech.com/abx/abx_coh.htm

How many such controversies do you see potentially settled here?


  #20   Report Post  
Arny Krueger
 
Posts: n/a
Default

agent86 wrote:

If I ever bothered to do an analog & a digital mix
of the exact same program material & do one of Arny's ABX tests on it
I suspect I would hear a difference.


If it was anything but a static mix, the level-matching requirement (0.1 dB
or so) would be practically impossible to manage.

However, let's say that you heard a difference, what would that mean? The
analog mixer could have frequency response or noise problems, while the
digital mixer can be for all practical purposes, perfectly clean and flat.
But no way would ever I hope to move faders around on the DAW in real time
like I do with the analog console. No, I draw gain and pan envelopes on the
DAW.

I often mix the same music live on an analog mixer, and mix it again hours
or days later on a DAW. I think that most people would say that the DAW mix
is the better-sounding one. Of course, the results sound different. They are
different artistic works.

I don't worry about the reasons why they sound different, because its all
about getting each job done right. And, they are two vastly different jobs.




  #22   Report Post  
agent86
 
Posts: n/a
Default

Arny Krueger wrote:

agent86 wrote:

If I ever bothered to do an analog & a digital mix
of the exact same program material & do one of Arny's ABX tests on it
I suspect I would hear a difference.


If it was anything but a static mix, the level-matching requirement (0.1
dB or so) would be practically impossible to manage.

However, let's say that you heard a difference, what would that mean?


My point exactly!

  #23   Report Post  
Ben Bradley
 
Posts: n/a
Default

On Fri, 7 May 2004 05:06:11 -0400, "Arny Krueger"
wrote:

Well, that's why we came up with ABX in the first place. People would read
stuff in TAS or someplace and get really obsessed with some new purported
*improvement*. Some old grizzled veteran would say "That's a piece of crap",
and then we'd do a DBT to see who was right. Take a look at this:

http://www.pcavtech.com/abx/abx_coh.htm

How many such controversies do you see potentially settled here?


Now I can see why ABX is itself SO controversial - if controversies
were settled, many vocal people would have nothing to say!

-----
http://mindspring.com/~benbradley
  #24   Report Post  
Arny Krueger
 
Posts: n/a
Default

Ben Bradley wrote:
On Fri, 7 May 2004 05:06:11 -0400, "Arny Krueger"
wrote:

Well, that's why we came up with ABX in the first place. People
would read stuff in TAS or someplace and get really obsessed with
some new purported *improvement*. Some old grizzled veteran would
say "That's a piece of crap", and then we'd do a DBT to see who was
right. Take a look at this:

http://www.pcavtech.com/abx/abx_coh.htm

How many such controversies do you see potentially settled here?


Now I can see why ABX is itself SO controversial - if controversies
were settled, many vocal people would have nothing to say!


There is evidence that for every controversy settled, more than one new one
springs up.

;-)


  #25   Report Post  
Justin Ulysses Morse
 
Posts: n/a
Default

Ben Bradley wrote:

Ironically, changing the gain in 'digital' is a multiplication
function, and that increases the word length and requires dither to go
back to the original word length, and all that.
Summation (the calculation replaced by this box of resistors) is
literally adding numbers in a processor, and as long as you keep up
with overflows (which may require double precision integers, but
that's not a lot of overhead), the result is always 'exact'.


I have a sneaking suspicion that, for precisely the reasons you state
above, software designers don't devote as much computational power to
the summation as they do to the gain changes. Summing is supposed to
be easy and perfect, and a lot of people assume that it is (remember
when people used to assume that CDs were perfect?) which is why they
haven't taken the time to verify it or improve it.

I suppose the reason these things are made is because people can
hear the difference, and many people like what comes out of the
resistor box better than what comes out of a digitally summed bus. I
hesistate to even post in this thread, because I don't think I can
hear the difference myself, and that seems to be the big criterion
here on RAP.


I make no claims about my own hearing and listening abilities. I
frankly can't hear any difference when comparing most of the things
that people talk about hearing. Comparing cheap old converters to
cheap newer converters, I hear no difference except when one is audibly
broken (which happens a lot). Cables? Nope, I can't tell 'em apart.
I even souped up my SX202 using Monte's recipe, and I can't tell it
apart from my stock units. In general I agree with Arny about DBTs
and our tendency to hear things that aren't there. But the comparisons
of internal and external mixes I've listened to were not subtle. They
were more substantial than a lot of other things we obsess over every
day. Whether you'd prefer the analog mix to the internal bounce is
another issue, but you would most definitely hear it. Even the one
person who's returned a Folcrom talked about the differences he heard.

Perhaps the difference is in two D/A converters (giving the exact
same distortion to everything) vs. 8 or 16 D/A converters (each one
does something slightly different to the signal its carrying), but
that's just a hypothesis.


That's certainly a possibility. Even if that were the explanation, I
can't think of another way to get around it besides using an external
summing device. I think there's more to it than that, but I'm also
speculating.

ulysses


  #26   Report Post  
Monte P McGuire
 
Posts: n/a
Default

In article ,
Justin Ulysses Morse wrote:
agent86 wrote:

If the argument were that the mathematical calculations within the digital
summing bus were themselves inherently flawed, I could buy that. (Or I
could at least consider it a point worthy of discussion.)


This is the argument for analog summing. We know that digital summing
should theoretically be perfect but so far all of the claims that
real-world software is in fact perfect seem to grossly underestimate
some software companies' ability to screw it up. That's my current
view of the situation, anyway. Unfortunately this is one of the most
difficult areas to perform conclusive, objective tests because there
are just so many variables.


My take on it is that most folks have crappy DACs that don't sound
good with complex program material. Basically, they don't sound good
playing back entire mixes. If you split your program out so that each
channel only handles a single instrument, bad cross modulation stuff
happens less.

The problem remains that your mix is now analog and must be converted
again, and people don't seem to want to monitor with a DAC on the
other side of that conversion. So, it's not a fair comparison at all.

IMHO, you can avoid all of these problems by buying a good DAC.

But that would also raise the question of why one would expect the
mathematical calculations within the DA & AD converters to be any
less flawed.


The DACs themselves simply aren't asked to perform this particular kind
of calculation.


Yeah... they get to do much more complicated ones... upsampling a wide
word at a low sample rate to a narrow word at 64x to 256x the original
sample rate. Lotsa multiply-accumulates, and we haven't accounted for
the calculations inside of the A/D's decimator, needed to get back
into the digital realm.


Regards,

Monte McGuire

Reply
Thread Tools
Display Modes

Posting Rules

Smilies are On
[IMG] code is On
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
rec.audio.car FAQ (Part 3/5) Ian D. Bjorhovde Car Audio 0 March 6th 04 07:54 AM
How do you make a Jeep sound good (or at least decent :) Simon Juncal Car Audio 3 January 29th 04 06:31 PM
MAKE CASH FAST TO PAY FOR YOR SYSTEM! Shaynemcgill Car Audio 0 October 1st 03 08:10 PM


All times are GMT +1. The time now is 04:05 PM.

Powered by: vBulletin
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 AudioBanter.com.
The comments are property of their posters.
 

About Us

"It's about Audio and hi-fi"