Digital BW, The Print

--- In DigitalBlackandWhiteThePrint@y..., "Austin Franklin" <darkroom@i...> wrote:
> Hi, Oh Prolific Roy ;-)
> 

Gee, Thanks for the moniker.

There seems to be a few instances of disagreements about whether we're
talking about linear versus exponential.  Its usually an issue of what units
and what measuring device.  Before going into the details, let me say that
I'm always referring to a scale as it relates to human perception.  Certainly
its OK to use an electronic device for measuring the number but I always
want to relate it back to how we will perceive it.  For example, dB's are
linear because we hear it linear as well.  Amp wattage is exponential because
we need exponential increases in wattage to produce linear perception.
The examples that have no "human" connection are kind of up for grabs
so I think it'd be best to avoid them (or be explicit).

> 
> dB (as in the result of a dynamic range equation) does not translate back to
> "Height" and "NumSteps", dB is merely a relative scale of power. 
OK.

  You could
> have an absolute range of 1V with the exact same dynamic range as a range
> with 100V, the "heights" as you call them, are entirely different.
?? not sure what's stated here

> Relatively, yes, the max/min are the same...but as I said, there is NO way
> to get back to the original hard (as in actual values, not relative values)
> from dB.  That's simple fact.

Sure, specifying the size a range whether exponential or linear never tells 
the absolute endpoints -- just relative to each other.
ExponRange = Max / Min
or
LinearRange = Max - Min
-- neither range value tells the absolutes.

> 
> > The basic fallacy is the mixing of linear scales and exponential
> > (or geometric)
> > scales
> 
> Er, I didn't mix them anywhere...but you're right, that would be a
> fallacy...
My main point goes back to using the "scale" (expon or linear) as
defined by the human not by the measuring device.

> 
> > You are
> > looking at the DynRange formula and seeing the term (maxpower/minpower)
> > and "feeling"  that its the same thing as the ruler.
> 
> The example of a ruler is fine, providing you understand what it means.  In
> the example I've cited above using 6V/.01V, no matter where you are on the
> scale, you can only measure to .01V.
Seems like there isn't human factor here, saying you can measure 0.01V,
sounds like everything is linear.

> 
> Now, using a ruler for your example of power.  As an aside, every 3db is a
> doubling of power, not 2x the dB value...but that may not matter...it
Yes, the dB unit would always be linear scale in my mind.
> depends on what "unit" your measurement is in.  If you ARE measuring in dB
> (which I would assume you would have to do to give two numbers like 60dB as
> your largest swing, and .5dB as your minimum discernable), and your
> measurements are in fact only accurate to a certain dB level (as your
> example above), then a ruler works fine also.
> 
> Just as an aside note, in audio, just because you have doubled power, you do
> NOT get double the "loudness".  That takes ~10x to double the loudness.
I'm not sure if this is a direct response to something I wrote.  I'm really
only interested in saying power is exponential, and I'll stick to using "loudness"
just as a descriptive term.

> 
> The fallacy would be applying steps of "hard" numbers to a power system, and
> vice versa.  It doesn't work.  You couldn't say you had a max of 60dB and a
> step of .01V.  Both numbers have to stay in the same measurement system
> (unit).
Sure, can't mix units.

> 
> > Its like
> > minpower is the
> > step size and when you divide it into maxpower you get "how many steps".
> 
> If you are measuring in power, then that works fine.
> 
> > This is exactly what you've said in the second paragraph above.
> 
> Actually, I used volts, not power, but it does still apply.
> 
> > This is completely wrong as a concept.  Why?? Because power is an
> > exponential scale from a perceptual point of view.
> 
> It doesn't matter.  All that matters is that you understand what your
> "system" is, and what you are measuring in.  Typically, minimum measurable
> of a system is in the same "units" as the system you are measuring...  If it
> 's power, then it's power.  If it's volts, then it's volts.  Exponential
I know power = volts^2 so either is OK.  
> system or not...  That appears to a possible source of your confusion.
> 
> > I.E.  To get
> > a "linear" feel
> > of loudness in the audio system, the power output has to go up
> > exponentially.
> 
> Correct.

For the last couple of paragraphs, I think you agree that "power" in the
audio sense is an exponential scale.  It looks like you are also maintaining
that  (maxpower/minpower)  has a "how many steps" meaning.
This, to me, is the crucial point in this whole discussion.  Its a purely
mathematical issue that for an exponential scale, the max/min ratio
is just a size of the total range.  There's no meaning of step here.
I thought my different of exponential series in the last post clearly
demonstated that.  


> 
> > Here's a short audio example:
> > We have an audio amplifier -- puts out 100 watts at the max end,
> > and measuring
> > the low end (smallest discernable signal) we get 1 watt.
> 
> A power output of 1 watt at the low end does not necessarily mean you have a
> smallest discernable signal of 1 watt, BTW, but I'm fine with that number,
> as it really doesn't matter.
> 
> > Let's
> > throw that
> > into the DynRange formula:  the inside term is (100/1), you are
> > thinking that
> > this is 100 steps.  Not TRUE!!
> 
> It certainly IS true, IF your measuring device is accurate to only 1 watt,
> as YOU specified above, and you ARE measuring watts...  YOU seem to be
> confusing systems here.  What ever your measuring system is, you are only as
> accurate as the measuring system is...and if your measurement system gives
> measurements in some log form (as a densitometer does, for example, or a
> sound pressure level meter), then your steps are also in log form.
> 
> BUT if you are measuring in "loudness", well, then, you can't very well give
> "watts" as your unit, as watts is NOT the same relative measurement of
> "loudness".  I believe that is a clear mistake.
You can use watts its just that you have to think in exponential series.
This is the point with the various series that I wrote earlier.

> 
> <big snip>
> 
> > See now have:
> > DyR  = maxdb - mindb
> 
> Sure, note both are in dB.  Nothing wrong with that.
My point here was to mathematically change equation and show
it no longer has the divide that (at least I'm surmising) makes the
original form of equation look like "number of steps".


  BUT...if you are
> trying to claim somehow that has something to do with density range, it does
> not!  Density measurements are NOT in dB, period.
Well, yes, density and dB are different.  But I'm glad you noticed the
hint of them having some parallelism.  Mathematically they sure do
look a lot alike, don't they?

Say we had a series of audio powers:
1,  1/2, 1/4, 1/8, 1/16 ...
this would correspond to decreasing dB's:
0, 3, 6, 9, 12, 15 ...

Similarly, we had a series of light intensities:
1,  1/2, 1/4, 1/8, 1/16 ...
this would correspond to densities of:
0, .3, .6, .9, 1.2, 1.5, ...

I'd sure like an awful good reason to think all this parallelism
should be ignored.  If you have a reason, please do tell.

> 
> > All things called "ranges" are just a top end and a bottom end.
> 
> Correct, but then add the word "dynamic" to it, and adds another dimension
I like it still to be a range -- great to add more qualification (isn't that
what adjectives do?).

> to the meaning, or why even bother adding it...especially considering
> dynamic means "characterized by continuous change".  That hardly seems

Well, at least we both have cute quotes now :).   I'm going to have to keep
asking what's "continuously changing".

> applicable to two static numbers!
Aren't your maxpower and minpower measurements basically a "static" numbers.
What are you calling "continuous change"?   I don't see any "continuous
change" when you measure the power amp and a certainly do see any
continuous change when you get to the photo print. 

Are you thinking of the sinewave change of the audio signal?  But you don't
perceive changes in the constant tone.  I guess if you wanted you could
also claim sinewave electromagnetic radiation i.e. light, but you don't
perceive that either.

> 
> > Dynamic range is a range just like any other range,  the word "dynamic"
> > is just a qualifier that says "perceptually interesting part of".
> 
> I don't mean to be snide, but that's just silly.  So many things have ranges
> specified, and they aren't called "dynamic ranges", 
I didn't say all  ranges are dynamic.  I said a "dynamic range" is firstmost
a range and it ought to AT LEAST have the notion of being a range.
The "dynamic" is an ADDITIONAL modifier, qualifying additional limits.

   but they are the range
> within which the device operates.  In fact, many specs include both a range
> of operation, and a dynamic range.

Well we're not looking at the same spec sheet, but it sounds like the range
of operation is the total range where you don't fry the components,
whereas the dynamic range is were the device works "well" i.e. meets the
specs of what it is supposed to do.

> 
> BTW, things can only be perceived if you can "perceive" a difference, which
> fits perfectly with what I've said ;-)
> 
> If I get one of THE "gurus" 
   You think there's any chance of defining "guru"?  :)   If there are gurus 
shouldn't they have a book?  Didn't we fail to agree on any books that either
of us found?

  to say that "dynamic range is NOT the same as
> density range" will you give?

Actually, I love hear somebody else say something about it all.  

   Well it seems that if we're defining DynRange of
something whose key property is density, we ought to end of with some kind
of range whose units are density.  I not saying the DynRange has to be
identical to the total density range -- but it ought to  be some subset.
Similarly you could be talking about the DynRange of a natural
scene --  density wouldn't be to useful since there isn't any.  Here
you ought to use whatever units you can measure.  (My spotmeter does
EV at ISO 100 for instance, so I'd be thinking of DynRange in EV's).

  You're on a crusade to prove something wrong
> that isn't wrong, and you aren't going to be able to do it ;-)
Well somebody is :)

> 
> Regards,
> 
> Austin

Here we go again.

Roy