«

isw <> wrote:
> Wolfgang Weisselberg <> wrote:

>> You're basically trying to describe a circle with a triangle or
>> a rectangle --- once you have 100 corners, it's much smoother.

> It sounds like you know for a fact that the Apple "eyeball" calibration
> method has only a limited ability to manipulate control points, compared
> to a hardware calibrator. Is that really the case? I don't know of any
> reason why it would have to be so.

According to
http://www.youtube.com/watch?v=51gbxDwz0NM
there are only 5 grey/gamma points.

Thus, only 5 control points. Contrast and compare to 100,
500, 836 or more control points.

Yes, 5 control points, at 6, 12.5, 25, 50 and 75% grey may
have been enough for a well behaved CRT[1]. Well, not really.

For a LCD that may not enough[4].

VMRL (.wrl) files on request. I can also run a different 836
(or more, or less) patches check (hitting different spots)
to evaluate the capability of a L*a*b* clut profile from the
original 836 patches, if you're interested.

-Wolfgang

[1] Checked my CRT with a photospectrometer[2]: 0% Black, 100%
White, Red, Green, Blue and the relevant 6%, 12.5%, 25%, 50%
and 75% grey. Created a Gamma+Matrix profile from the data.
Checked the profile against the full 800+ control points[3]
from which these 10 values were pulled. Delta E (CIEDE2000):
RMS and average <1 (probably invisible), maximum >3 however
(clearly visible), and the plot shows that all the dark
colours are well off, due to non-linearity in all 3 colours
in the dark area.

[2] Which delivers absolute values and full spectral data ---
quite unlike your eye.

[3] 85 steps on the gray axis, 65 steps on each colour axis,
the rest well distributed.

[4] Computed as in [1], laptop LCD (TN) screen, white LEDs.
Delta E (CIEDE2000): average >3.6, RMS >4.0, maximum >8.0.
Very bad. Everything's visibly off (even on the gray axis!).
The plot shows most areas are off, only one part of the grey
axis is sorta neutral.

Retry with a Shaper+Matrix profile. Slightly better average
and RMS, maximum DE error now >11. The plot shows the grey
axis being mostly correct in the darkest part and returning
to neutral in 2 or 3 other places.

Eric Stevens <> wrote:
> On Thu, 7 Jun 2012 13:51:45 +0200, Wolfgang Weisselberg
>>isw <> wrote:
>>> Wolfgang Weisselberg <> wrote:

>>>> You're basically trying to describe a circle with a triangle or
>>>> a rectangle --- once you have 100 corners, it's much smoother.

>>> It sounds like you know for a fact that the Apple "eyeball" calibration
>>> method has only a limited ability to manipulate control points, compared
>>> to a hardware calibrator. Is that really the case? I don't know of any
>>> reason why it would have to be so.

>>According to
>>http://www.youtube.com/watch?v=51gbxDwz0NM
>>there are only 5 grey/gamma points.

> There is nothing absolute about a colour management process which
> enables you to "choose your display to be as sharp and vivid as you
> prefer it to be".

With any colour management process your're free to detune your
display and use sharpening and vibrancy settings as you like.

> It might be OK if all you want to do is stare at the
> screen but not if you want to print to a colour-managed printer.

Not with 5 point adjustments.

-Wolfgang

isw <> wrote:
> Wolfgang Weisselberg <> wrote:
>> Eric Stevens <> wrote:
>> > On Thu, 7 Jun 2012 13:51:45 +0200, Wolfgang Weisselberg
>> >>isw <> wrote:
>> >>> Wolfgang Weisselberg <> wrote:

>> >>>> You're basically trying to describe a circle with a triangle or
>> >>>> a rectangle --- once you have 100 corners, it's much smoother.

>> >>> It sounds like you know for a fact that the Apple "eyeball" calibration
>> >>> method has only a limited ability to manipulate control points, compared
>> >>> to a hardware calibrator. Is that really the case? I don't know of any
>> >>> reason why it would have to be so.

>> >>According to
>> >>http://www.youtube.com/watch?v=51gbxDwz0NM
>> >>there are only 5 grey/gamma points.

>> > There is nothing absolute about a colour management process which
>> > enables you to "choose your display to be as sharp and vivid as you
>> > prefer it to be".

>> With any colour management process your're free to detune your
>> display and use sharpening and vibrancy settings as you like.

>> > It might be OK if all you want to do is stare at the
>> > screen but not if you want to print to a colour-managed printer.

>> Not with 5 point adjustments.

> I don't agree --

Fine. Do you have data? Facts?

> it depends on how much interpolation can be done from
> those points,

Oh, you can interpolate as much as you like. The question
is: how close do you come to reality?

> and I suspect that it could result in a pretty decent
> curve.

Sure. A broken clock could result in showing the right time,
too. Happens pretty regularly, twice a day, in fact.

The idea is not to "suspect" a pretty decent result, but to get
one, every time.

> The interpolation should work well, because it is a
> "well-behaved" function.

Do you have data for that? Do you have facts?


> But anyhow, if it's not possible for a human to create a decent match
> *at those points*, why do you think that a machine could do it?

Why not?
Humans cannot fly ... but with machines they can.
Humans cannot see single cells ... but with microscopes they can.
Humans cannot listen to radio waves ... but with radios they can.

Humans cannot measure brightness and saturation by eye with
high accuracy nor absolutely ... but with machines they can.

Still think not?

> As I
> said, if you need an "eleven" and the knob only goes to "ten" ...

Oh, but the problem is you cannot even see the knob properly
without a machine ... much less see what it is set to.

-Wolfgang

isw <> wrote:

> I'm still trying to find someone who actually *knows* in precisely what
> ways the Apple "eyeball" calibration method is flawed.

I gave you the answer a couple posts ago.
Short form: 5 measurement points on the grey axis aren't enough
for many screens even *IF* the eye was a perfect instrument.

Delta E (2000) averages of 4.0 and maximums of 8 are BAD.

FYI:
ΔE means
0.0 … 0.5 practically no difference
0.5 … 1.0 might be visible to the trained eye
1.0 … 2.0 visible difference if you look closely
2.0 … 4.0 clear colour differences
4.0 … 5.0 rarely tolerated colour differences
more than 5.0 that's a completely different colour

So if your screen, on the average, is showing hard to tolerate
colour shifts and for some colours even completely different
colours ... then the calibration method is unusable.

Now add in that your eye is not a perfect instrument. That it
*will* be off by some degree for every measured point by about
1 ∆E or more. (If you want, I can randomize the measured data
for the 5 points by 1-2 ∆E and compute the result based on that.)


> Because from my
> understanding of how it works, it should produce pretty decent results.

Decent as in 'the grey axis will look quite neutral to the
untrained eye, and the colours will be similar', yes.

Decent as in 'If you hold your print next to your monitor it'll
look the same[1][2]', not in a 1000 years.

Decent as in 'The monitor will give you a good idea how the print
will look like', depends a lot on how good an idea you want.

If you've not got a good eye for colour and look at the print
a few minutes after you've seen the monitor and don't have any
comparisons when looking at the print (i.e. you're not trying to
hit a specific colour, say a company logo's colour or a specific
flower and thus are not holding the print next to such an object),
you will be OK.

> Your statements and questions above make it pretty clear that you're not
> going to be the one who provides the answer.

It seems you're not even getting that you have been handed facts.
Or at least that you have problems evaluating what they mean.
OR that you are a fanboi and therefore there are no arguments
that could make you change your mind, since that would challenge
a dogma or something.

If it's the first two, feel free to ask "what does ∆E mean?"
or "How do you arrive at these ∆E numbers?" or similar.

Here:
https://gist.github.com/2929806
is the 836-patches file (Laptop.ti3), the extracted 5-grey-patches
file (5-Laptop.ti3) and the error that happens when you turn
5-Laptop.ti3 into an ICC (matrix+shaper) and test it with
Laptop.ti3 (5-Laptop.wrl) (colour coded by ∆E).

Feel free to play with the values, feel free to experiment with
other ICC generation methods, feel free to report back if you
manage to get a better result (and if so, how).

-Wolfgang

[1] assuming all colours are in the gamut of the print and of
the monitor
[2] you also should use controlled lighting and have the
print corrected for the lighting used (by default D50)
and the monitor set to emulate for the same lighting

isw <> wrote:
> In article <ptvoa9->,
> Wolfgang Weisselberg <> wrote:

>> isw <> wrote:
>>
>> > I'm still trying to find someone who actually *knows* in precisely what
>> > ways the Apple "eyeball" calibration method is flawed.
>>
>> I gave you the answer a couple posts ago.
>> Short form: 5 measurement points on the grey axis aren't enough
>> for many screens even *IF* the eye was a perfect instrument.

> The human eye is very good at making comparisons and determining when
> two things are the same

Within 1-2 ∆E typical. Consumer sensors are more accurate,
much faster and don't tire.

> -- which is the way the Apple method works.

So how do you compare to, say, a GretagMacbeth Colour Checker
(and that's only 24 patches, not hundreds) in D50 normlight (which
costs more than a simple tristimulus sensor) to your monitor?

So how do you compare the brighness to, say, 120 cd/m²?

How do you comare the chromaticity of the 100% white to the
neutral grey or white in D50 (or D65, or whatever you want to
standardize on)?


All you *can* compare is a relative brightness compared to 0% and
100% (i.e. you have no idea if your screen is too bright!) and
the gray neutrality *compared to* the overly blueish white of
the native monitor setting.


> Have you ever used it?

Does one have to use it *and* not compare the result with an
instrument to have a religious experience?

How about *you* use it and then measure, say, 800 odd patches and
tell us how close *you* came using the eyeball method on *your*
monitor?


>> So if your screen, on the average, is showing hard to tolerate
>> colour shifts and for some colours even completely different
>> colours ... then the calibration method is unusable.

> Or (as I suspect might be the case) the screen (and its driving
> hardware) is simply not able to be calibrated properly.

Yep, that's the "sour grapes" strategy.

Do rent a hardware puck for a couple days and tell us. Your
average photo club, your rental service, etc. all can supply
you with one.


> Using the Apple calibration method, I can do a perfectly decent job on
> the Dell 2407;

So, how high is the average ∆E, how bad the worst ∆E and
please show us an ∆E plot ...

Oh, you don't KNOW!

So how do you *know* you can do a perfectly decent job?


> all the test images on the web site you referred to look
> just fine.

Aha.

BTW, which web site did I refer to?
Oh, but I *didn't*.
So all the imaginary test images on an imaginary website look
'just fine', when you don't even have any idea how they're supposed
to look.

I am sure you didn't have them printed on well calibrated systems
(with enough gamut for the whole image) and compared to your
monitor under normlight.


> I cannot achieve similar performance from the MacBook's
> display, whether I use the Apple-provided profile, or one of my own --
> though the latter are (I think) somewhat better.

You're using very insufficient methods, which may work somewhat
on some displays, but which is not good enough for others.

That's like saying a 2 MPix compact camera doesn't resolve
enough for a 1 meter sized image viewed at close distrance,
THEREFORE it's impossible to create such an image at all.


> I have, on several occasions, tried to calibrate other LCD screens as
> well, also mostly with poor results. It is known that LCD screens show
> much more variability in color accuracy than CRTs (usually),

That's true. How could you have both wide gamut screens and
standard gamut screens otherwise?

> and that
> would seem to indicate that not all screens can be properly calibrated.

Ah ... no. Even systems like printers with 8, 10, 12 colours
can be calibrated --- and that even though they're mixing colours
wildly[1] and they're not exactly known to be monotonous.

So, it's easily true that Apple's eyeball method is not enough
for LCDs (and indeed I said as much), but THAT'S ALL.


> Otherwise, why would it even be mentioned in reviews?

Lots of things are mentioned in reviews.
And some monitors are better (!= good, they're much to bright!) as
they come from the factory than others.


>> > Because from my
>> > understanding of how it works, it should produce pretty decent results.

>> Decent as in 'the grey axis will look quite neutral to the
>> untrained eye, and the colours will be similar', yes.

> Well, the Dell 2407 is, and the MacBook is not, capable of producing a
> neutral grey scale; that is precisely the problem.

The MacBook is perfectly capable of producing a neutral grey
scale.
"After calibration the colour is neutral"
http://forums.whirlpool.net.au/archive/957785

It's your technique that's lacking.

> It cannot do that
> even at the five points that you claim are insufficient.

If you cannot set it to neutral grey because the adjustment
slider area is too small, then it is a fault of your technique.
If you cannot set it to neutral because the adjustment steps are
too large, then it is a fault of your technique.
If you cannot set it to neutral because youc cannot see what
neutral would be, that's again a fault of your technique.


> I have a
> feeling that if (as you claim) a hardware calibrator can manage to stick
> values in the LUT that the "eyeball" method cannot, the result would be
> better color accuracy at the expense of some other parameter, possibly
> brightness or contrast ratio.

That's true: if you need to convert your 9600K monitor white
to D50, you pay with some brightness. If you need to correct
your dark colours, you may need to raise the black point a bit.
Both eat contrast ratio.

Just like a not wide open aperture eats light.


> Given that choice I (and perhaps Apple)
> would choose a bright display.

Then STOP TRYING TO CALIBRATE DISPAYS INCLUDING APPLE'S
"EYEBALL" METHOD!

STOP USING ANY PROFILES!

STOP THIS WHOLE THREAD!

You want overly bright, blue-tinted displays, so that every
print looks too dark and too red? FINE. YOUR PROBLEM.


> After all, the Mac is a portable
> computer, not a dedicated photo editor; the screen needs to be usable
> under a wide range of situations.

Yep, that's why overly bright and overly blue is good. NOT!

Ever heard of being able to remove the settings from the LUT
and reverting to "default values"?


>> Decent as in 'If you hold your print next to your monitor it'll
>> look the same[1][2]', not in a 1000 years.

> FWIW, I'm not interested in hard-copy output; it really is pretty
> limited, both in gamut, and in contrast ratio as well.

Really. That must be because you love overly bright, overly
blue screens.


>> Decent as in 'The monitor will give you a good idea how the print
>> will look like', depends a lot on how good an idea you want.

>> If you've not got a good eye for colour and look at the print
>> a few minutes after you've seen the monitor and don't have any
>> comparisons when looking at the print (i.e. you're not trying to
>> hit a specific colour, say a company logo's colour or a specific
>> flower and thus are not holding the print next to such an object),
>> you will be OK.

>> > Your statements and questions above make it pretty clear that you're not
>> > going to be the one who provides the answer.

>> It seems you're not even getting that you have been handed facts.
>> Or at least that you have problems evaluating what they mean.
>> OR that you are a fanboi and therefore there are no arguments
>> that could make you change your mind, since that would challenge
>> a dogma or something.

> If I were a "fanboi", would I be talking about how good the Dell display is, and
> complaining about the one in the Mac?

You're a fanboi of the Apple eyeball method.


> I just want some facts, specifically explaining *how* the Apple method
> is flawed.

How many repetitions do you want?

> All you've done is to say over and over that the mechanized
> method is better. I want to know *why*.

- the eye isn't an absolute instrument. It cannot tell you if
the monitor is too bright or too blue (or too red) without
something to compare against.
- the eye is less sensitive to colour differences than an
instrument. If you use the measurements to extrapolate to other
colours, you need to be more exact ... the larger the range
of colours you extrapolate to, the more exact you need to be.
There are >16 million colours ...
- 5 measurements are about 50 too few --- and even then you
only have a minimal, rough profile.
- 5 measurements on the grey axis don't say anything about the
other colour axes nor about the rest of the colour space.
- 5 measurements (or 10, as it may be) only allow a gamma or a
shaper profile, and they have assumptions on the display that
said display doesn't necessary fulfill.
- Apple's eyeball method is based on relative brightnesses,
compared to 0% and 100% white. Said white is usually completely
wrong, as it leaves the factory. Connect the dots ...
- Apple's eyeball method isn't good enough for many LCDs, as you
have proven (MacBook screen!)

I guess I said all that a couple times now.


> Saying that "you get more, and
> more precise, numbers" doesn't mean anything unless you can state
> precisely *why* that difference matters.

"A color image which straddles the two screens can be a painful
thing to see ..."
-- isw in Message-ID: <isw-4ECE75.14125826052012@[216.168.3.50]>

It's rather obvious, isn't it?

All you need is 2 proper profiles, one for each monitor ... for
which you need more numbers to remove gamma curve assumptions and
the like and for which you want more precise numbers to have less
extrapolation errors. *Especially* for your laptop you need
many more measurements.

If you were living close enough, I'd come over and calibrate and
profile your screens just to prove my point.

So I can just tell you: rent or borrow a hardware unit and try
for yourself.


> It's like the audio folks who
> go on about 24 bit samples at 96 or 192 kilosamples/second, and how much
> "better" that is than 16 bit, 44.1 kilosample sound. Sure, the numbers
> are bigger, but in fact, the difference is simply not audible, so who
> cares?

Do you know what a brick filter is?
Do you know if it's easier to build if the steepness of the filter
can be reduced?
Do you know why a deep pass filter is needed for digital sampling?

Then you know why more samples/s.

Do you know why some people use RAWs with 12 or 14 or more bits
when JPEGs with visually evenly distributed 8 bit are visually
not different at all (and in fact almost all monitors can
only display 8 bit?)

Then you know why 24 bits sample depth is preferred for some
tasks.

Did you know that sometimes it's much easier to throw more
data into a system than to optimize it to the same result
with less data?


>> If it's the first two, feel free to ask "what does ΔE mean?"
>> or "How do you arrive at these ΔE numbers?" or similar.

> I have some knowledge of those, and a few other, things concerning color
> imaging. Prior to retiring, I spent many years in television technology
> including some work with very large screen projection systems, helped
> develop MPEG-2, and designed a few high-performance film printers, among
> other things.

OK. So it's more a case of obstinacy of old age: you think you
already know everything, so there could be no reason someone else
could realistically offer that would change your mind. I still
hope it's not the reason, but as they say, hope's the last thing
to die.

Please report back with the ∆E values of your 'calibrated'
screen.

Or at least tell us with your knowledge which 5 points I
should pick for calibrating my LCD monitor ...

-Wolfgang

[1] for example they're mixing additional CMY (and their light
versions) to the black (and grey and light grey) inks to
reduce speckle (just one or two black or grey dots look more
irregular than a whole handful of dithered LM, LC and Y dots)
and black is even blacker if CMY is mixed into it as well.

Oh, and of course there's no perfectly neutral grey, so there
CMY (or their light versions) needs to be mixed in, too.

isw <> wrote:
> Wolfgang Weisselberg <> wrote:
>> isw <> wrote:
>> > Wolfgang Weisselberg <> wrote:
>> >> isw <> wrote:

>> >> > I'm still trying to find someone who actually *knows* in precisely what
>> >> > ways the Apple "eyeball" calibration method is flawed.

> --snip--

>> > Have you ever used it?

>> Does one have to use it *and* not compare the result with an
>> instrument to have a religious experience?

> Well, it's difficult to understand how you can know how bad it is if you
> don't even know *what* it is or how it works ...

Well, I know *what* a Saturn V is and how it works. Yet I've
never ridden a rocket ... can you spot the error in your logic?

If anything I have done a better test than what Apple's
eyeball offers, with twice as many measurements and more
accurate data ...

> -- snip --

>> >> Decent as in 'If you hold your print next to your monitor it'll
>> >> look the same[1][2]', not in a 1000 years.

>> > FWIW, I'm not interested in hard-copy output; it really is pretty
>> > limited, both in gamut, and in contrast ratio as well.

>> Really. That must be because you love overly bright, overly
>> blue screens.

> Actually, it's because I know how poor color reproduction is, using CMYK
> inks.

Try CcMmYRKLKLLK. Or CcMmYKRG.
Or use a RGB process --- like photosensitive paper, chemically
developed, just like in the days of yore.

And even basic CMYK doesn't automatically mean poor color
reproduction, just a somewhat more limited gamut. (And yet it'll
display colours that are not available for your average screen.)


> -- snip --

>> > It's like the audio folks who
>> > go on about 24 bit samples at 96 or 192 kilosamples/second, and how much
>> > "better" that is than 16 bit, 44.1 kilosample sound. Sure, the numbers
>> > are bigger, but in fact, the difference is simply not audible, so who
>> > cares?

>> Do you know what a brick filter is?
>> Do you know if it's easier to build if the steepness of the filter
>> can be reduced?
>> Do you know why a deep pass filter is needed for digital sampling?

> I was designing sharp-cutoff phase-corrected analog filters for
> broadcast applications over forty years ago.

How sharp, how much supression, at what price?


> I have designed both anti-
> aliasing filters for encoding, and reconstruction filters for decoding,
> digital signals. Not to mention antialiasing of computer-created
> photographic images by the proper overlap of the gaussian brightness
> distribution of the CRT spot. What's your point?

OK, then you should know why oversampling is a good idea.

>> Then you know why more samples/s.

> I know that once you have exceeded the resolution capabilities of the
> final "receiver" (in this case human perception), either in amplitude
> precision or time precision, more bits in either dimension are
> completely pointless.

So your average stereo system does not come with eqalizers?
Never mind the high end systems ...

>> Do you know why some people use RAWs with 12 or 14 or more bits
>> when JPEGs with visually evenly distributed 8 bit are visually
>> not different at all (and in fact almost all monitors can
>> only display 8 bit?)

>> Then you know why 24 bits sample depth is preferred for some
>> tasks.

> And is totally unnecessary for final viewing or listening.

Only when you're using a format that needs absolutely no more
shaping or tweaking (say, colour management or equalizers).
The more you need to shape your signal, the stronger some
parts will be stretched, soon causing visible (or audible)
steps between minimal increments.

Audio needs much less information than images, so it's much easier
to just throw in a couple more bits and to use systems that are
cheaper, but need more signal shaping. Images just need too much
space for that to work easily.


>> Did you know that sometimes it's much easier to throw more
>> data into a system than to optimize it to the same result
>> with less data?

> Easier, and a method commonly used by those who don't understand things
> very well, but usually not better, for lots of reasons.

At least that's your story, and your indignity at people going the
'easy' way and arriving with better results than you do.

-Wolfgang

nospam <> wrote:
> isw <> wrote:

>> I have, on several occasions, tried to calibrate other LCD screens as
>> well, also mostly with poor results. It is known that LCD screens show
>> much more variability in color accuracy than CRTs (usually), and that
>> would seem to indicate that not all screens can be properly calibrated.
>> Otherwise, why would it even be mentioned in reviews? I bought the 2407
>> after learning that it did, indeed, have good color performance.

> actually, lcds are *more* accurate than a crt

By what measure?
They're often more well behaved ... but that's all.

> there is also no drift or warm-up period with an lcd as
> there is with a crt.

CCFL does need a 30 minute warm up period to become stable.
Additionally, CRTs usually have identical brightness all over
the screen, this is not the case for most LCDs ...


> it's the same with audio. you don't measure frequency or loudness by
> ear, you use a frequency counter or sound pressure meter. doctors take
> your temperature with a thermometer, not by touching your forehead like
> parents sometimes do.

Oh, touching the forehead will do as a rough check if the
person is *way* out of the normal temperature range, given
good enough circumstances.

-Wolfgang

isw <> wrote:
> nospam <> wrote:
>> isw <> wrote:

> -- snip --

>> you keep saying how wonderful the apple method is, but with no proof.

> I've never said it was "wonderful". I've been told that it was "no
> good", and I'm just trying to find out *why*. I have described the
> results to some folks who, best I can figure out, have never tried it,
> and know nothing about it, not even the theory behind it. Other than how
> bad it is ...

Well, I know the theory, I know what it matches, and I have
*proved* that even given very accurate measurements (way better
than the eye produces) it's not very good. Between "just
acceptable" and "completely unusable" for colour matching.

All you've been doing is handwaving that another method which
you have never tried, don't understand what it's doing and
know not a shred of theory about cannot be visually better,
won't work on your laptop screen anyways and is wholly
unneccessary.

I've provided data.
You've provided more handwaving.


So is that a case of the pot calling the bridal gown black?


> -- snip --

>> it's the same with audio. you don't measure frequency or loudness by
>> ear, you use a frequency counter or sound pressure meter.

> The point is that in both cases, measuring to arbitrarily great
> precision is possible, but the question (unanswered) is *at what
> precision is human visual resolution equalled or exceeded*?

That has been answered. Roughly around 1 ∆E, depending on the
specific ∆E used and the colour and brightness. Later ∆E
definitions are better in that regard, like the 2000'er definition.

> Because that
> is a limit that it's simply not necessary to exceed.

*sigh*
If you're extrapolating --- and you are, since you're not checking
16 million colours --- any error is multiplied with the distance.

So, yes, if you were checking every colour, you'd need not
to exceed 1 ∆E. Since with the Apple eyeball method you're
extrapolating from 50% gray to 89% blue, you need way more accurate
measurements or way closer measurements. An instrument checking
hundreds of patches provides both.

> I'm perfectly aware that a mechanical calibrator can deliver more
> "accuracy". What nobody who says it is bad seems to know is *in what
> ways* the Apple method is deficient.

See all of my other posts.

See all of my other posts.

See all of my other posts.

See all of my other posts.

See all of my other posts.

See all of my other posts.

See all of my other posts.

See all of my other posts.

See all of my other posts.

See all of my other posts.

See all of my other posts.


I start to feel like a broken record.
Either you're not even grasping the slightest bit of what I am
saying, so you cannot even ask intelligent questions, or you're
trolling. Maybe English isn't your first language?

-Wolfgang

In article <isw-07399C.23355515062012@[216.168.3.50]>, isw
<> wrote:

> > you keep saying how wonderful the apple method is, but with no proof.
>
> I've never said it was "wonderful". I've been told that it was "no
> good", and I'm just trying to find out *why*. I have described the
> results to some folks who, best I can figure out, have never tried it,
> and know nothing about it, not even the theory behind it. Other than how
> bad it is ...

it's not that it's 'no good' it's that it isn't *as* good as a hardware
puck. this should be obvious, without any of the proof that has been
provided.

have you tried a hardware puck? if you have, you'd see the difference
and this discussion would not be taking place. whether you care about
the difference is another story.

In article <isw-AF24F1.21404517062012@[216.168.3.50]>, isw
<> wrote:

> > > > you keep saying how wonderful the apple method is, but with no proof.
> > >
> > > I've never said it was "wonderful". I've been told that it was "no
> > > good", and I'm just trying to find out *why*. I have described the
> > > results to some folks who, best I can figure out, have never tried it,
> > > and know nothing about it, not even the theory behind it. Other than how
> > > bad it is ...
> >
> > it's not that it's 'no good' it's that it isn't *as* good as a hardware
> > puck. this should be obvious, without any of the proof that has been
> > provided.
>
> More precision is in some way "better". The question I still have is,
> *how* is the Apple eyeball method deficient -- what errors will it make,
> that a machine will not?

because your eye is very easily fooled. look at the optical illusions i
posted.

> > have you tried a hardware puck? if you have, you'd see the difference
> > and this discussion would not be taking place. whether you care about
> > the difference is another story.
>
> I want to find out if the investment is worth the money before I buy
> one. To date, nobody has been able to tell me, chapter and verse,
> precisely what is wrong with the Apple method. I'm beginning to suspect
> that nobody knows ...

people have told you, you just don't want to hear it.