Bilinear, bicubic and pixel resize
Jean-Luc ERNST
I enlarge 800x an image (crop of part of a screen copy).
I make 3 enlargements (resize) via the three possibilities in PSP7 :
bicubic, bilinear and pixel resize.
I see no differences between bilinear and pixel resize (arithmetic tool
gives a black image when set on differences between the two images).
Bicubic is selected by the Smart size option in PSP when image is enlarged
but gives, IMHO, a blurred image.
Could someone explain why no differences between bilinear and pixel resize ?
Is it the same system with two different names ? Why those 2 options
because they give exactly the same result ?
Thank you.
Jean-Luc
Ronald Vick
wrote:
From the help files:
"· Bicubic resample, which uses a process called interpolation to
minimize the raggedness normally associated with expanding an image.
As applied here, interpolation smoothes out rough spots by estimating
how the "missing" pixels should appear, and then filling them with the
appropriate color. It produces better results than the Pixel resize
method with photo-realistic images and with images that are irregular
or complex. Use Bicubic resample when enlarging an image.
· Bilinear resample, which reduces the size of an image by
applying a similar method as Bicubic resample. Use it when reducing
photo-realistic images and images that are irregular or complex.
· Pixel Resize, where Paint Shop Pro duplicates or removes
pixels as necessary to achieve the selected width and height of an
image. It produces better results than the resampling methods when
used with hard-edged images.
Vixen of Verbiage"
Note that Bilinear is the normal choice for the Reduction of an image,
not enlargement. I suspect that using it for an enlargement causes
it to work much like pixel resize.
Selecting the 'Smart' automatically selects the correct mode for the
current operation so the user need not worry about selection of the
modes.
Jaggiemeister Ron - PSP terrorist
Questions? Answers at http://campratty.com/questions.html
Kris Zaklika
>
> Hello !
> I enlarge 800x an image (crop of part of a screen copy).
No need to use such a large factor - it just means you have
to examine more pixels and can't look at each individual
pixel. For example, you will get the same result if you
increase your image size two-fold (200%). Now repeat your
experiment resizing the image by a non-integer factor, e.g.
1.23 (123%). Then tell me what you see.
It may be easier if you use an easily comprehensible image.
Make a 10 x 10 pixel image. Put white in the left half and
black in the right. Set grey (128,128,128) as the color and,
with a one-pixel size brush, draw a diagonal line from top
left to bottom right.
Using this source image try each of the resampling methods
for increasing your image size 123%. Compare them at a
zoom of 20:1. Before you dismiss bicubic and decide bilinear
is better, compare these two result images side by side at
2:1 zoom. In which is the diagonal line easier to see? Why?
(Hint: look carefully at contrast using 20:1 zoom again.)
> I make 3 enlargements (resize) via the three possibilities in PSP7 :
> bicubic, bilinear and pixel resize.
> I see no differences between bilinear and pixel resize (arithmetic tool
> gives a black image when set on differences between the two images).
> Bicubic is selected by the Smart size option in PSP when image is enlarged
> but gives, IMHO, a blurred image.
Experiments that use factors for resizing that are exactly
divisible by two don't tell the whole story.
> Could someone explain why no differences between bilinear and pixel resize ?
> Is it the same system with two different names ? Why those 2 options
> because they give exactly the same result ?
After following my suggestions you can probably draw your
own conclusions about these questions :)
> Thank you.
>
> Jean-Luc
--
----------------------------------------------------------------------
Kris Zaklika Jasc Software, Inc. The
Product Ideas: id...@jasc.com Power
Customer Service: customer...@jasc.com To
Technical Support: tec...@jasc.com Create
----------------------------------------------------------------------
Jean-Luc ERNST
I'm forced to eat my hat ! (don't know is this french expression exists in
english).
Thank you for switching on the lighthouse !
Jean-Luc
"Kris Zaklika" <kzak...@jasc.com> a écrit dans le message de news:
3D550E41...@jasc.com...
Kris Zaklika
>
> Hey Kris, you are a real devil ! Or a sorcerer ! ...
> I'm forced to eat my hat ! (don't know is this french expression exists in
> english).
The equivalent English expression is "to eat one's hat" :)
It's very clever of you to have attempted to investigate
the resampling systematically. You're the first one in
my experience who has.
> Thank you for switching on the lighthouse !
The more you know the more effective you can be with PSP.
I'm happy to help.
> Jean-Luc
[snip]
Sally Beacham
"Jean-Luc ERNST" <j...@wanadoo.be> wrote in message
news:aj4541$1dfob$1...@sinclair.be.wanadoo.com...
> Hey Kris, you are a real devil ! Or a sorcerer ! ...
> I'm forced to eat my hat ! (don't know is this french expression exists in
> english).
> Thank you for switching on the lighthouse !
>
> Jean-Luc
>
>
Jean-Luc....
You crack me up. ;-))))
--
Sally Beacham / www.dizteq.com
www.lvsonline.com / PSP, Filter Frenzy, Xara X
FilterMunky / www.psppower.com
reply to sbeachamATdizteq.com
Kennedy McEwen
<kzak...@jasc.com> writes
>No need to use such a large factor
Well actually, yes there is, because it is just such a scale factor
which reveals the bug!
>
>Experiments that use factors for resizing that are exactly
>divisible by two don't tell the whole story.
>
Experiments which use factors less than two clearly don't tell the whole
story either!
Integer multiples, whether odd or even, are only special cases of
general resizing. Bilinear resizing should not discriminate special
cases.
Three isn't a scale that is exactly divisible by two. Try the PSP7
resizing with a scale of 300% and you might see the difference between
what bilinear resizing should do and the rather poor implementation this
edition of PSP actually provides. :-(
If you had even bothered to look at what Jean-Luc suggested you might
realise why he asks the following:
>> Is it the same system with two different names ? Why those 2 options
>> because they give exactly the same result ?
>
The real answer is that PSP doesn't implement bilinear resizing any
more. Instead, it implements something which is not even a close
approximation of bilinear resizing. Given your politician like response
to divert attention away from the odd effect of the function, I suspect
that you are only too aware of its obvious deficiencies!
>After following my suggestions you can probably draw your
>own conclusions about these questions :)
>
My own conclusion is that if this is what you intended it to do, Kris,
then you really should call it something unique to PSP, because it
certainly *ISN'T* bilinear resizing, as used throughout the industry
long before either PSP (or even your expensive competitor) came on the
scene!
Bilinear resizing is a mathematical function that is well defined and
certainly doesn't give the results that PSP7 produces for resizing
scales greater than x2. If my memory serves me correctly, bilinear
resizing has been in common usage for more than 40 years. I know I
certainly used it on a PDP11 image processing suite over 20 years ago
and I don't think I have seen such that operation ever give the effects
of PSP7!
Bilinear resizing should fit new pixels according to where they lie on a
flat linear surface drawn between the original four original samples
which encompass the new sample. That is WHY it is called bilinear. Two
original points in each axis - a LINEAR fit between them in both axes.
Anything other than that is not bilinear!
PSP7 APPEARS to use pixel resizing to get up to the integer scaling
below the required factor and then bilinear resizing to implement the
remaining fractional part. This is why, when there is no fractional
part, it is indistinguishable from the pixel resizing result - exactly
as Jean-Luc noted. That is a guess based on what I see from some
samples I have tried although for some percentage scalings and pixel
sized images it doesn't even seem to be x-y symmetric!
I am sure you can tell us exactly what is going on, but IT ISN'T
BILINEAR RESIZING! It's a PSP specific algorithm (albeit rather less
useful than bilinear resizing, as Jean-Luc noted), so give it a unique
PSP name, not one that is an industry standard for a completely
different function! That is just asking for trouble and only likely to
cause confusion for anyone using your software seriously.
Strangely, the last version of PSP that I seriously used (Version 4.14)
actually implemented bilinear scaling pretty much correctly, so I am
tempted to assume that this is a programming bug rather than a new
scaling function that Jasc have simply misnamed in the hope that nobody
will notice. ;-)
Do you implement bicubic correctly or does that use another programming
shortcut that has gone wrong as well?
--
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers (replace 'nospam' with 'kennedym' when replying)
Jean-Luc ERNST
I don't promise to learn all because I suppose this is advanced maths but
I'm interested - for my general knowledges - to read something about it.
I could use Google but if you have something relevant, it would spare my
time.
Thank you,
Jean-Luc
"Kris Zaklika" <kzak...@jasc.com> a écrit dans le message de news:
3D55A772...@jasc.com...
Jean-Luc ERNST
Jean-Luc
"Sally Beacham" <p...@tension.you> a écrit dans le message de news:
aj4dr...@enews4.newsguy.com...
Jean-Luc ERNST
Thank you...
Jean-Luc
"Kris Zaklika" <kzak...@jasc.com> a écrit dans le message de news:
3D55A772...@jasc.com...
--
...[snip]...
>. You're the first one in
> my experience who has.
>
Uni
Gee, Kennedy, don't criticize Kris's baby, Paint Shop Pro. It's
professional software, as you can tell :)
Uni
Kennedy McEwen
>
>Gee, Kennedy, don't criticize Kris's baby, Paint Shop Pro. It's
>professional software, as you can tell :)
>
what he wanted it to do, but the nomenclature is wrong, misguided and
misleading. It may be that the software is wrong too but only Kris, or
his colleagues, can comment on that aspect.
Incidentally, just to prove that it isn't only PSP that is wrong, I
should correct one of my comments in the previous post. After I
returned to work today I had a chance to check the version of PSP4.1.4
on my old machine there and guess what - it doesn't do bilinear
filtering correctly either (although, in fairness, it doesn't claim that
it does either). Instead it makes the same coarse filtering that PSP7
does - so it could be deliberate after all!
Just shows you the tricks memory can play - I haven't used the function
in years and now I remember why!
Kris Zaklika
>
> Some URLs about bilinear and bicubic ?
> I don't promise to learn all because I suppose this is advanced maths but
> I'm interested - for my general knowledges - to read something about it.
>
> I could use Google but if you have something relevant, it would spare my
> time.
I don't have anything lying around except technical papers
about different algorithms for image resampling. I doubt
these are much use. You could search for "Mitchell Netravali"
to get an old but basic description.
The general idea is this. Let's for ease of discussion suppose
you only want to change the width of the image so the problem
becomes one-dimensional. The difficulty is that for some
arbitrary value of the resize factor you have to calculate
a pixel value at some position between the original pixels
of the image. For instance, if you halve the width of your
image you need new pixels that are half way between your
old ones. You could compute these by fitting a linear
function to your pixels and solving the equation to find
the color components of a pixel lying somewhere between
the original pixels. A linear equation needs a slope and
an intercept to describe it so you can use as few as two
image pixels to determine the equation that gives you the
new pixel values. What you actually use depends on the
factor by which you are changing the size of the image.
If you cut it to one-quarter of its original size, for
instance, the number of original pixels contributing to a
new image pixel would be four. A cubic equation can also be
fitted to image data and this requires four coefficients
to be determined (those at x cubed, x squared, x and the
intercept). The advantage of the cubic equation is that it
has a point of inflection and can better follow complicated
variations in the image colors from pixel to pixel. The
negative aspect is that you need to use more pixels in
the calculation and then it is slower. So that's linear and
cubic. Bilinear and bicubic is just doing these things
in two dimensions (width and height) simultaneously.
There are, in fact, a whole host of other mathematical
functions you could use to fit to your image data. The
choice is usually determined by a desire to preserve the
highest spatial frequencies in the image, the avoidance of
artifacts due to aliasing and the need for speed.
Kennedy McEwen
<j...@wanadoo.be> writes
>Some URLs about bilinear and bicubic ?
>I don't promise to learn all because I suppose this is advanced maths but
>I'm interested - for my general knowledges - to read something about it.
>
This one is quite handy, because you can cut and paste the source image
into PSP and then implement the various 450% resizing algorithms it
claims to offer, compare the result and then see how short changed PSP
is leaving you:
http://www.dpreview.com/learn/Glossary/Digital_Imaging/Interpolation_01.h
tm
These are more mathematical:
http://www.cs.wisc.edu/graphics/Courses/cs-638-1999/image_scaling.htm
http://ise0.stanford.edu/class/ee368a_proj00/project12/2.html
http://tem692.gsfc.nasa.gov/public/reference/Bilinear%20Interpolation
http://www.mathworks.com/access/helpdesk/help/techdoc/math_anal/poly_i15.
shtml
http://www.pcigeomatics.com/cgi-bin/pcihlp/GCPWORKS%7CTheory%7CResampling
http://www.cs.ubc.ca/spider/little/505/lectures/subsection3_9_2.html
http://www.wisdom.weizmann.ac.il/~maksimf/ex5/Resample.html
http://www.omatrix.com/manual/interp2.htm
http://crssa.rutgers.edu/courses/remsens/remsensing5/tsld019.htm
Here is one on bicubic scaling and how to do that properly:
http://members.bellatlantic.net/~vze2vrva/design.html
Finally, a rather interesting paper which concludes that fractal
resizing is actually quantifiably INFERIOR to both bicubic and bilinear
(surprised? so was I!) methods when they are implemented correctly - ie.
not using PSP!
http://www.gisdevelopment.net/aars/acrs/1997/ps2/ps4029pf.htm
Kennedy McEwen
<kzak...@jasc.com> writes
NO! The number of original pixels contributing to ANY new pixel in one
dimensional linear resampling IS ALWAYS TWO, irrespective of how big you
scale it!!!
Kris Zaklika
[snip]
> NO! The number of original pixels contributing to ANY new pixel in one
> dimensional linear resampling IS ALWAYS TWO, irrespective of how big you
> scale it!!!
I have a 100 x 1 pixel image in which each pixel has a
unique color. I resize this to a 1 x 1 pixel image. Which
two pixels of the original image, in your opinion, should
be chosen to form the result image?
> --
> Kennedy
> Yes, Socrates himself is particularly missed;
> A lovely little thinker, but a bugger when he's pissed.
> Python Philosophers (replace 'nospam' with 'kennedym' when replying)
--
Uni
>
> In article <3D57E5FD...@usa.net>, Uni <plg...@usa.net> writes
> >
> >Gee, Kennedy, don't criticize Kris's baby, Paint Shop Pro. It's
> >professional software, as you can tell :)
> >
> Well I wasn't necessarily criticising the software - it may do exactly
> what he wanted it to do, but the nomenclature is wrong, misguided and
> misleading.
What else is new? Remember, if you feel it's wrong, then YOU ARE WRONG,
in Kris's mind. I haven't heard him yet admit to any fault. Why should
he? That wouldn't "sell". Remember, mum's the word :)
> It may be that the software is wrong too but only Kris, or
> his colleagues, can comment on that aspect.
>
> Incidentally, just to prove that it isn't only PSP that is wrong, I
> should correct one of my comments in the previous post. After I
> returned to work today I had a chance to check the version of PSP4.1.4
> on my old machine there and guess what - it doesn't do bilinear
> filtering correctly either (although, in fairness, it doesn't claim that
> it does either). Instead it makes the same coarse filtering that PSP7
> does - so it could be deliberate after all!
>
> Just shows you the tricks memory can play - I haven't used the function
> in years and now I remember why!
No need to shout, I'm NOT disagreeing with you :)
Just do what most folks here do, make the best of it.
Regards,
Uni
UNI - que
> in Kris's mind. I haven't heard him yet admit to any fault. Why should
> he? That wouldn't "sell". Remember, mum's the word :)
Good point, Uni friend!!!
I'm sure Elements does bilinear, trilinear and quadrilinear resampling in a
wink, doesn't it. (Please say it does, Uni friend, I need it badly to resize
some cutie pictures I got)
Uni - que
Kennedy McEwen
<kzak...@jasc.com> writes
>Kennedy McEwen wrote:
>[snip]
>
>> NO! The number of original pixels contributing to ANY new pixel in one
>> dimensional linear resampling IS ALWAYS TWO, irrespective of how big you
>> scale it!!!
>
>I have a 100 x 1 pixel image in which each pixel has a
>unique color. I resize this to a 1 x 1 pixel image. Which
>two pixels of the original image, in your opinion, should
>be chosen to form the result image?
>
has been the topic for the rest of the thread. Certainly, when
DOWNSAMPLING you take all the data from the original samples which fall
inside the new pixel, linearly weighted.
But now you are here, and getting back to the original question posed in
Jean-Luc's first mesage on the thread: why doesn't bilinear scaling work
correctly in PSP when UPSCALING?
Which is it - bug or design? If the latter, why call it what it ain't?
@compuserve.com Mark Johnson
>Hello !
>I enlarge 800x an image (crop of part of a screen copy).
>I make 3 enlargements (resize) via the three possibilities in PSP7 :
>bicubic, bilinear and pixel resize.
>I see no differences between bilinear and pixel resize (arithmetic tool
>gives a black image when set on differences between the two images).
>Bicubic is selected by the Smart size option in PSP when image is enlarged
>but gives, IMHO, a blurred image.
Enlarging twice, 200%, should be fine. Enlarging much more than that
would only look okay when reduced back down in print. Otherwise,
obviously, it will look blurry, on screen.
Bilinear is used for reductions. And as you say, smart size
automatically picks bicubic for enlargements. I would assume bilinear
is ignored for pixel resize >100% if bilinear is specifically
selected. I noticed, myself, last year, the difference when wondering
if I could enlarge a shot and not introduce too much aliasing. And up
to about 200%, you can enlarge rather smoothly, with bicubic. Take a
1mp shot, and turns it 4mp . . almost. Same for 4mp. If you need an
25meg shot for print, you could go 200% on the 4mp, and get your
resolution, just without _quite_ the detail of a genuine 16mp. If more
is needed, you'd need more resolution (or a wide frame).
Jean-Luc ERNST
a real work.
I'm using PSP mostly for photo retouching and I'm always attentive to the
possible degradation when working with successive tools...
I'm using a scanner and when enlargment is needed, I start to enlarge with
the scanner first.
I don't use big resamples (like 800%) with PSP, just little adjustments.
But, I was curious about the bilinear and bicubic resampling and which is
better when enlarging photos in PSP. I'm testing myself, even is PSP says
this or this is better. Testing gives me more comprehension about the
software.
If a take a glass magnifier, I can see the grain in a photo (argentic) and
the dust spots, etc.
If a go to 800% in PSP, I suppose I'll see problems if they occur... That
was my meaning.
My mistake was to apply 800% and not 753 % or 872 %.
And as Kris said, on a little image with a reduced enlargment, I can already
see how the two systems are working (not need of maths, only my eyes) and to
understand which is the better.
Friendly,
Jean-Luc
"Mark Johnson" <102334.12 @ compuserve.com> a écrit dans le message de news:
7mggluk3been68hb3...@4ax.com...
Jean-Luc ERNST
I bi-understand...
Jean-Luc
"Kris Zaklika" <kzak...@jasc.com> a écrit dans le message de news:
3D57F816...@jasc.com...
> Jean-Luc ERNST wrote:
> >
> > Some URLs about bilinear and bicubic ?
> > I don't promise to learn all because I suppose this is advanced maths
but
> > I'm interested - for my general knowledges - to read something about it.
> >
> > I could use Google but if you have something relevant, it would spare my
> > time.
--
...[snip]...
>
> The general idea is this. > > > [snip]
Kennedy McEwen
<j...@wanadoo.be> writes
>My mistake was to apply 800% and not 753 % or 872 %.
No, your mistake was obviously to believe to the waffle that Kris told
you. Go and try 753% or 872% with PSP - IT JUST DOESN'T WORK, because
it ISN'T doing bilinear resizing!
If you like, I will send you a very small image and a version using
PROPER bilinear resizing of any scale greater than 200% that you like -
integer or not, multiple of two or not. You can then compare the result
to what PSP actually provides.
Better still, download the example from the following web page that I
already gave you and try it yourself:
http://www.dpreview.com/learn/Glossary/Digital_Imaging/Interpolation_01.h
tm
PSP is UNABLE to upscale correctly using bilinear resizing.
It is either a bug or a lie but, apparently, Kris isn't prepared to say
which.
Uni
>
> In article <ajc4du$1c1mf$2...@sinclair.be.wanadoo.com>, Jean-Luc ERNST
> <j...@wanadoo.be> writes
>
> >My mistake was to apply 800% and not 753 % or 872 %.
>
> No, your mistake was obviously to believe to the waffle that Kris told
> you.
Be forgiving of Kris, Kennedy. His mind was probably consumed with more
important things.....
http://community.webshots.com/photo/42670345/46945073DtOxTI
Uni
@compuserve.com Mark Johnson
>But, I was curious about the bilinear and bicubic resampling and which is
>better when enlarging photos in PSP. I'm testing myself, even is PSP says
http://www.nut-n-but.net/CCPCUG/TechBrief%20Articles/TB0204.html
For example.
Kennedy McEwen
<1023...@compuserve.com> writes
Oh dear, another author taken in by Jasc lies. That document says:
"For example, Adobe’s Photoshop program allows the user to choose from
one of three techniques: nearest neighbor, bilinear, and bicubic.
PaintShopPro also offers three techniques: pixel resize, bilinear, and
bicubic"...
But we know, and Jean-Luc has seen it with his own eyes, that
PAINTSHOPPRO DOESN'T DELIVER BILINEAR - IT ONLY OFFERS IT!!
Kris Zaklika
[snip]
> Oh dear, another author taken in by Jasc lies.
I really don't have the time (or even the inclination) to
engage you about the topic of your monomania. I do, however,
resent being called a liar personally and I resent the
implication that the company I work for tells lies. Both
assertions are untrue.
I suggest you go and watch a few Monty Python tapes until
you calm down and learn to be civil.
Sally Beacham
"Kris Zaklika" <kzak...@jasc.com> wrote in message
news:3D5ADAB1...@jasc.com...
> Kennedy McEwen wrote:
> [snip]
>
> > Oh dear, another author taken in by Jasc lies.
>
> I really don't have the time (or even the inclination) to
> engage you about the topic of your monomania. I do, however,
> resent being called a liar personally and I resent the
> implication that the company I work for tells lies. Both
> assertions are untrue.
>
> I suggest you go and watch a few Monty Python tapes until
> you calm down and learn to be civil.
>
Thank goodness you told him. His hysteria was scaring the cats.
Kennedy McEwen
<kzak...@jasc.com> writes
>Kennedy McEwen wrote:
>[snip]
>
>> Oh dear, another author taken in by Jasc lies.
>
>I really don't have the time (or even the inclination) to
>engage you about the topic of your monomania. I do, however,
>resent being called a liar personally and I resent the
>implication that the company I work for tells lies. Both
>assertions are untrue.
>
>I suggest you go and watch a few Monty Python tapes until
>you calm down and learn to be civil.
>
simple bug or a misnomer - the thread has been running for more than 4
days yet your responses to it have avoided the issue entirely. Instead,
you have used that time to mislead anyone who asked what the function
you call "bilinear" actually did and why it differed from real bilinear
resizing. Now you claim you don't have time to admit to which of the
two simple errors I suggested it could be! That makes it quite plain
that your, and Jasc's, intention of calling this function "bilinear" is
to mislead. I consider deliberate deception to be worse than lying, but
YMMV!
Kennedy McEwen
<p...@tension.you> writes
>
>Thank goodness you told him. His hysteria was scaring the cats.
>
glasses you use to look at PSP's "bilinear" resizing operations through!
You can fool some of the people all of the time, perhaps all of the
people some of the time, but sooner or later every cover up is exposed
for what it is.
I have asked Kris a simple question several times, I have posted
explanations of what is wrong with his company's program together with
web references of how the job should be done - that isn't hysteria, it
is an enquiry! You might see the difference if you get rid of those
rose tinted spectacles!
Sally Beacham
"Kennedy McEwen" <r...@nospam.demon.co.uk> wrote in message
news:x8+dy6Hq...@kennedym.demon.co.uk...
> In article <ajels...@enews1.newsguy.com>, Sally Beacham
> <p...@tension.you> writes
> >
> >Thank goodness you told him. His hysteria was scaring the cats.
> >
> Perhaps you should try hysteria - it shake off those rose coloured
> glasses you use to look at PSP's "bilinear" resizing operations through!
>
> You can fool some of the people all of the time, perhaps all of the
> people some of the time, but sooner or later every cover up is exposed
> for what it is.
>
> I have asked Kris a simple question several times, I have posted
> explanations of what is wrong with his company's program together with
> web references of how the job should be done - that isn't hysteria, it
> is an enquiry! You might see the difference if you get rid of those
> rose tinted spectacles!
Methinks you doth philosophize and proselytize too much. PSP does for me
what I need it to do. I'm happy. I have no intention of looking to get
unhappy.
If you'd like a kinder/gentler reception, perhaps you ought to check out the
water before you try to dynamite the fish. It's not the wisest action in
the world to accuse people you don't know, and companies you aren't familiar
with, of lying and obfuscating, when in fact you want them to do something
that to this point, I've only seen YOU make an issue of, after lo these many
years of the software's existence.
Unless of course, you're just the sock puppet I think you are.
Uni
>
> In article <ajels...@enews1.newsguy.com>, Sally Beacham
> <p...@tension.you> writes
> >
> >Thank goodness you told him. His hysteria was scaring the cats.
> >
> Perhaps you should try hysteria - it shake off those rose coloured
> glasses you use to look at PSP's "bilinear" resizing operations through!
>
> You can fool some of the people all of the time, perhaps all of the
> people some of the time, but sooner or later every cover up is exposed
> for what it is.
>
> I have asked Kris a simple question several times, I have posted
> explanations of what is wrong with his company's program together with
> web references of how the job should be done - that isn't hysteria, it
> is an enquiry! You might see the difference if you get rid of those
> rose tinted spectacles!
Kennedy, don't waste your time arguing with the little fish. Go for the
Gold Fish, named Kris.
I see he's doing his typical stalling for time, which isn't unusual.
Good luck.
Cheers,
Uni
Uni
>
> Kennedy McEwen wrote:
> >
> > In article <ajels...@enews1.newsguy.com>, Sally Beacham
> > <p...@tension.you> writes
> > >
> > >Thank goodness you told him. His hysteria was scaring the cats.
> > >
> > Perhaps you should try hysteria - it shake off those rose coloured
> > glasses you use to look at PSP's "bilinear" resizing operations through!
> >
> > You can fool some of the people all of the time, perhaps all of the
> > people some of the time, but sooner or later every cover up is exposed
> > for what it is.
> >
> > I have asked Kris a simple question several times, I have posted
> > explanations of what is wrong with his company's program together with
> > web references of how the job should be done - that isn't hysteria, it
> > is an enquiry! You might see the difference if you get rid of those
> > rose tinted spectacles!
>
> Kennedy, don't waste your time arguing with the little fish. Go for the
> Gold Fish, named Kris.
> I see he's doing his typical stalling for time, which isn't unusual.
If you want to see how nicely an image can be scaled at approx. 1000%,
visit the
news:alt.binaries.paint-shop-pro newsgroup and check out my wallpaper,
which I scaled up for another contributor. What I used (a plug-in)
certainly blew away anything PSP could produce.
Uni
Steve-A
Kennedy,
Is someone twisting your arm to use it? If not, quit your bitching.
You sound like a 12 year old.
--
Steve-A
remove panties to reply
Steve-A
news:ajf65...@enews1.newsguy.com...
Sally,
He is a sock puppet, just like uni and mikey. The only problem is, he
appears on the surface to be about 12 years old. Reminds me of a child
that isn't getting his way, and throws a temper tantrum.
Nightingail
news:ajg1on$1b19ih$1...@ID-73017.news.dfncis.de...
> Sally,
> He is a sock puppet, just like uni and mikey. The only problem is,
he
> appears on the surface to be about 12 years old. Reminds me of a child
> that isn't getting his way, and throws a temper tantrum.
Sheesh, I don't always have a positive opinion about everything the
other regulars do around here (for ex. I personally think that playing
games with Uni, ranting about him, and/or polluting this newsgroup with
a repetition of what he posts on another ng all serve no purposes other
than raising the noise level in here, making the poster feel better, and
giving Uni yet more attention). Of course they're entitled to post what
they wish, and I am entitled to either read it or not, as I choose.
That's my opinion - does that make me a sock-puppet or a troll too?
Gail
--
Nightingail's Gallery
http://www.nightingail.com/
Digital Artwork
Free Web Art and Graphics
Steve-A
news:bZO69.10920$WJ3.1...@news1.news.adelphia.net...
> Sheesh, I don't always have a positive opinion about everything the
> other regulars do around here (for ex. I personally think that playing
> games with Uni, ranting about him, and/or polluting this newsgroup
with
> a repetition of what he posts on another ng all serve no purposes
other
> than raising the noise level in here, making the poster feel better,
and
> giving Uni yet more attention). Of course they're entitled to post
what
> they wish, and I am entitled to either read it or not, as I choose.
> That's my opinion - does that make me a sock-puppet or a troll too?
>
> Gail
Unlikely. You offer good info/help.
Uni
But Steve-A NEVER does!!!!!!!!!!!!!!! :)
Uni
Mike
Steve-A wrote:
>
> "Nightingail" <night...@hotmail.com> wrote in message
> news:bZO69.10920$WJ3.1...@news1.news.adelphia.net...
>
> does that make me a sock-puppet or a troll too?
>
> Unlikely. You offer good info/help.
I find it amusing that Steve is not sure though. He writes "unlikely"
as evidence of his paranoia.
":^) ®
--
Mike
• Logo Design •
Put some fun in your next logo!
Site at: http://www.artistmike.com
Uni
I don't see him as that, Steve. However, I do view Kris Zaklika as being
a coward by not answering this individual.
Uni
Kennedy McEwen
>
>If you want to see how nicely an image can be scaled at approx. 1000%,
>visit the
>news:alt.binaries.paint-shop-pro newsgroup and check out my wallpaper,
>which I scaled up for another contributor. What I used (a plug-in)
>certainly blew away anything PSP could produce.
>
bilinear interpolation and I am more than familiar with the many
alternatives, having developed quite a few myself.
Kennedy McEwen
<p...@tension.you> writes
>
>Methinks you doth philosophize and proselytize too much. PSP does for me
>what I need it to do. I'm happy. I have no intention of looking to get
>unhappy.
>
has no influence on my life whatsover, and it is rather arrogant of you
to assume that it should - so please indicate where I asked you to get
unhappy. If recognising that PSP is flawed in at least this aspect
makes you unhappy to the extent that you would rather ignore the
possibility then that is your problem to deal with, not mine. A lot of
people in the world are quite happy to live in squalor - it doesn't mean
I should as well!
>If you'd like a kinder/gentler reception, perhaps you ought to check out the
>water before you try to dynamite the fish. It's not the wisest action in
>the world to accuse people you don't know, and companies you aren't familiar
>with, of lying and obfuscating, when in fact you want them to do something
>that to this point, I've only seen YOU make an issue of, after lo these many
>years of the software's existence.
>
I don't give a damn what the reception is, I didn't come here for a tea
party but to have a specific question about a FAULT in the program
answered. I used PSP for years, since Version 2.01 according to my
records (1991?), but possibly even before that and have worked in
digital image processing for over 25 years. I am MORE than familiar
with Jasc having provided them with considerable income over the years,
although in a professional context and not in a personal one. I looked
in here because I recently downloaded Version7 of PSP, having rejected
it several years ago due to other flaws, under some pressure to do so by
exactly the sort of people who are wittering here that it is good enough
for the job. You don't have a clue about my application so you haven't
a clue whether this flaw can or should be ignored. You should have
enough common sense to realise that I wouldn't be wasting my time in
this group if I thought that it could!
You might be happy being deceived, people who don't care whether what
they are doing is wrong usually are, but others prefer to be actually
examine the results before blindly using any tool. Usually that is the
result of experience, having been stung before and, yes, in my case that
was by PSP!
Kennedy McEwen
<lake_vu...@hotmail.com> writes
>
>Kennedy,
> Is someone twisting your arm to use it?
Well actually, several people are metaphorically doing just that! At
least one of them being an IT budget holder, the rest being the clueless
masses that blindly believe the product's claims and then question our
project's technical development budgets because they can't see the
benefit of certain processing steps and "can demonstrate its minimal
effect by using PSP"!
So, yes, I DO have a damned good motivation for getting to the bottom of
this and it would certainly be in Jasc's interest to state whether they
intend to fix the bug, change the misnomer or just continue the
deception. All three serve my purposes - only one benefits Jasc, but
their rep on the group would rather mislead and deceive.
>If not, quit your bitching.
>
>You sound like a 12 year old.
>
Whereas you sound like some fool who has no idea of what the
consequences of false claims are! Compared to you, most 12 year olds
are a LOT smarter!
Steve-A
news:3D5BDCB4...@usa.net...
> I don't see him as that, Steve. However, I do view Kris Zaklika as
being
> a coward by not answering this individual.
>
> Uni
John,
Ordinarily, I do not like responding to you, but in this case, I will
make an exception. If I were you, I would certainly hope that Kris Z
isn't a spiteful type of person with a lawyer that resembles F. Lee
Bailey.
I will also relate it to something similiar, like answering any of your
questions, what's the point? Regardless of what Kris says, you, Kennedy
and Mike will still be contrary to his position just because you can,
and you apparently have nothing better to do than try and get people to
have some semblance of a conversation with you by goading them into it
(Stay the F__K out of it Mike). It is quite obvious that none of the
three of you have any interaction with people outside of this or a few
other newsgroups, otherwise you would not be in here posting some of
your brain dead drivel, stirring up flame wars, etc.
Also, if Kris wants to answer Kennedy's question, he will. But, as he
has pointed out many times, he is not getting paid to come in here and
play tech support, mother, baby sitter and moderator. The fact that you
think he has to answer someone's question just because he works for JASC
makes you out to be a bigger fool than you already appear to be.
Consider yourself lucky that he even comes in here at all.
Now to return you to our regular program.
Steve-A
news:3D5BD550...@artistmike.com...
>
>
> Steve-A wrote:
> >
> > "Nightingail" <night...@hotmail.com> wrote in message
> > news:bZO69.10920$WJ3.1...@news1.news.adelphia.net...
> >
> > does that make me a sock-puppet or a troll too?
>
> >
> > Unlikely. You offer good info/help.
>
>
> I find it amusing that Steve is not sure though. He writes "unlikely"
> as evidence of his paranoia.
>
Give the Psychology diploma back, you failed the course.
Your writing is evidence of your lack of intelligence.
Steve-A
> I don't give a damn what the reception is, I didn't come here for a
tea
> party but to have a specific question about a FAULT in the program
> answered. I used PSP for years, since Version 2.01 according to my
> records (1991?), but possibly even before that and have worked in
> digital image processing for over 25 years. I am MORE than familiar
> with Jasc having provided them with considerable income over the
years,
> although in a professional context and not in a personal one.
You have purchased THAT many copies?!?
I looked
> in here because I recently downloaded Version7 of PSP, having rejected
> it several years ago due to other flaws, under some pressure to do so
by
> exactly the sort of people who are wittering here that it is good
enough
> for the job.
Oh, I see, someone came to your house and held a gun to your head and
told you to use PSP.
You don't have a clue about my application so you haven't
> a clue whether this flaw can or should be ignored.
No, you're right we don't have a clue as to what YOUR application is,
but then again why should we care?
> You should have
> enough common sense to realise that I wouldn't be wasting my time in
> this group if I thought that it could!
Okay, so you have apparently come to the conclusion that it won't do
what YOU want it to do, so why, in your words, are you wasting your time
in here?
>
> You might be happy being deceived, people who don't care whether what
> they are doing is wrong usually are, but others prefer to be actually
> examine the results before blindly using any tool.
As others have stated, it does for them what they want it to do, so
what's the problem?
> Usually that is the
> result of experience, having been stung before and, yes, in my case
that
> was by PSP!
> --
> Kennedy
Running around beating your chest like some gorilla will not get you
invited to the beach party.
Steve-A
> In article <ajg1j2$1av0m9$1...@ID-73017.news.dfncis.de>, Steve-A
> <lake_vu...@hotmail.com> writes
> >
> >Kennedy,
> > Is someone twisting your arm to use it?
>
> Well actually, several people are metaphorically doing just that!
Bull.
>
> So, yes, I DO have a damned good motivation for getting to the bottom
of
> this and it would certainly be in Jasc's interest to state whether
they
> intend to fix the bug, change the misnomer or just continue the
> deception.
Sounds like you must have some financial stake in JASC.
> >If not, quit your bitching.
> >
> >You sound like a 12 year old.
> >
> Whereas you sound like some fool who has no idea of what the
> consequences of false claims are! Compared to you, most 12 year olds
> are a LOT smarter!
Hmm, I don't come in here acting like a spoiled child that stomps my
feet on the floor if I don't get what I want if that's what you mean.
> --
> Kennedy
Kennedy McEwen
<lake_vu...@hotmail.com> writes
>
>Hmm, I don't come in here acting like a spoiled child that stomps my
>feet on the floor if I don't get what I want if that's what you mean.
>
asked him a simple question and given him more than adequate time to
reply. Instead he would rather counter with misleading information and
deception before sulking off. You, on the other hand, have added
nothing to the discussion but aggression. Can we expect to see any
positive contribution from you some time in the current millennium?
Kennedy McEwen
<lake_vu...@hotmail.com> writes
>Regardless of what Kris says, you, Kennedy
>and Mike will still be contrary to his position just because you can,
And your evidence of this is what?
I have already apologised to Kris because, owing to him addressing
completely the opposite of the topic under discussion, I misread one of
his previous messages.
For all you know, Kris may indicate that this is a bug which will be
addressed in the next issue. Why, having suggested that this is the
most likely situation myself and irrespective of the impression his
previous responses have generated, would I continue to take a position
contrary to his? You really don't understand to concept of postulate,
debate and conclude, do you?
BY NOT answering the question Kris has served my purpose even more
admirably than he ever could have by entering into a technical exchange.
Kennedy McEwen
<lake_vu...@hotmail.com> writes
>
>You have purchased THAT many copies?!?
>
rather than the software that does the job properly!
>
>Oh, I see, someone came to your house and held a gun to your head and
>told you to use PSP.
>
>
>Okay, so you have apparently come to the conclusion that it won't do
>what YOU want it to do, so why, in your words, are you wasting your time
>in here?
>
deception ended, either in an upgrade or in future versions!
>
> As others have stated, it does for them what they want it to do, so
>what's the problem?
>
hold up PSP faults as evidence of their position!
Steve-A
> In article <ajgqvc$1bbfhr$1...@ID-73017.news.dfncis.de>, Steve-A
> <lake_vu...@hotmail.com> writes
> >
> >Hmm, I don't come in here acting like a spoiled child that stomps my
> >feet on the floor if I don't get what I want if that's what you mean.
> >
> Seems to me like Kris and you are the ones stomping their feet. I
have
> asked him a simple question and given him more than adequate time to
> reply. Instead he would rather counter with misleading information
and
> deception before sulking off. You, on the other hand, have added
> nothing to the discussion but aggression. Can we expect to see any
> positive contribution from you some time in the current millennium?
> --
> Kennedy
You apparently missed one of my previous posts where I stated that Kris
is NOT obligated to come in here and answer everyone/anyone's questions.
Me? I don't stomp my feet unless I like the beat of the music.
It would depend on what your definition of "positive contribution" is.
If your previous posts are an indication of positive contribution, then
no, you won't see anything like that from me. But, if you you mean
helping people with questions pertaining to PSP, then the answer would
be a resounding yes.
And, I believe that Ron Lacey gave you the perfect avenue to vent your
frustrations.
With that, EOT for me.
Jackie Laderoute
<r...@nospam.demon.co.uk> wrote:
>In article <ajgqvc$1bbfhr$1...@ID-73017.news.dfncis.de>, Steve-A
><lake_vu...@hotmail.com> writes
>>
>>Hmm, I don't come in here acting like a spoiled child that stomps my
>>feet on the floor if I don't get what I want if that's what you mean.
>>
>Seems to me like Kris and you are the ones stomping their feet. I have
>asked him a simple question and given him more than adequate time to
>reply. Instead he would rather counter with misleading information and
>deception before sulking off. You, on the other hand, have added
>nothing to the discussion but aggression. Can we expect to see any
>positive contribution from you some time in the current millennium?
You are missing the point here - it is not Kris' job to answer your
questions in this forum (nor anyone else's). Neither you nor I know
what Kris' area of program development is, nor whether he had/has/will
have any input into the resizing algorithms.
If you want technical aspects of the program clarified for you, then
contact Jasc directly. No employee of the company is likely willing or
legally able to discuss these matters with you in a public forum.
I suggest you take Ron's excellent advice and go directly to Jasc with
your issues - you are much more likely to get the answers you require
that way. If you persist in waiting for a public answer in a Usenet
group, you are exceedingly likely to be disappointed.
Jackie
--
< o \"/ Don't play cat and mouse with me! (
---@ ) Master PSP today - http://www.lvsonline.com ()-()
< o /"\ Jackie Laderoute jflad...@shaw.ca (o o)
*****************************************************************/\o/\
Uni
>
> In article <3D5B318B...@usa.net>, Uni <plg...@usa.net> writes
> >
> >If you want to see how nicely an image can be scaled at approx. 1000%,
> >visit the
> >news:alt.binaries.paint-shop-pro newsgroup and check out my wallpaper,
> >which I scaled up for another contributor. What I used (a plug-in)
> >certainly blew away anything PSP could produce.
> >
> There are reasons why I and my colleagues want, and often need, true
> bilinear interpolation and I am more than familiar with the many
> alternatives, having developed quite a few myself.
I understand that and I know where you're coming from.
If you haven't received an explanation from Kris by now, you probably
never will.
Kris AVOIDS replying to ANY question which MAY hurt the sales of PSP.
Would YOU admit to making a mistake with YOUR software, if your job
depended on it?
But take it from me, it is of no value to argue with the PSP worshippers
here. They'll defend PSP, regardless of how bad it is, in any way, shape
or form.
Cheers,
Uni
Jean-Luc ERNST
I'm not a mathematician...
I'm working with my eyes for judging what gives a better result.
I' don't know what is the "legal" or "correct" or "unique" formula for
bilinear or bicubic resizing.
But I can see what PhotoShop makes and I'm not sure that PhotoShop is the
reference !
Honestly, I prefer the PSP system in bicubic rendering (and pixel resize
too).
I have observed that PhotoShop suffers from the "camel syndrom".
It means that the formula used by PhotoShop doesn't give a good result with
the 10x10 pixels image resize test (Kris's test image).
I have resized the 10x10 pixels image by 123 % in PSP 7.04 and in PS 5.0
I have put the result here :
http://cf.photos.yahoo.com/jeanlusque
If you want load them, the 4 images are not big (from 23 K to approx. 90K) :
"original", "pixel resize", "binilear" (sic) and "bicubic".
You can see in the PS5 "pixel resize" test that 2 humps appear on the
diagonal line. They don't appear in PSP7.
In PS5 "bicubic resize" those 2 humps are visible too (not in the PSP7
resize). For me, it is a defect in PhotoShop and PSP7 is better for this
rendering.
PSP and PS are not using exactly the same formulae. Perhaps because patent
reason : Corel PhotoPaint9 in "pixel resize" creates 2 humps too, one near
the center of the line (after 4 pixels) and one at the end bottom right.
It seems that each software has a different formula.
Note : during my tests, I observed that PaintShop Pro was able to zoom at
3200 % on screen but PS5 was limited to 1600% zoom.
Friendly,
Jean-Luc
"Kennedy McEwen" <r...@nospam.demon.co.uk> a écrit dans le message de news:
KXKj4HG5...@kennedym.demon.co.uk...
> In article <4qeklug913713gvei...@4ax.com>, Mark Johnson
> <1023...@compuserve.com> writes
> >"Jean-Luc ERNST" <j...@wanadoo.be> wrote:
> >
> >>But, I was curious about the bilinear and bicubic resampling and which
is
> >>better when enlarging photos in PSP. I'm testing myself, even is PSP
says
> >
> >http://www.nut-n-but.net/CCPCUG/TechBrief%20Articles/TB0204.html
> >
> >For example.
>
> Oh dear, another author taken in by Jasc lies. That document says:
> "For example, Adobe's Photoshop program allows the user to choose from
> one of three techniques: nearest neighbor, bilinear, and bicubic.
> PaintShopPro also offers three techniques: pixel resize, bilinear, and
> bicubic"...
>
> But we know, and Jean-Luc has seen it with his own eyes, that
> PAINTSHOPPRO DOESN'T DELIVER BILINEAR - IT ONLY OFFERS IT!!
Ember
wrote:
>Hello !
>I'm not a mathematician...
>I'm working with my eyes for judging what gives a better result.
>I' don't know what is the "legal" or "correct" or "unique" formula for
>bilinear or bicubic resizing.
>But I can see what PhotoShop makes and I'm not sure that PhotoShop is the
>reference !
>Honestly, I prefer the PSP system in bicubic rendering (and pixel resize
>too).
>I have observed that PhotoShop suffers from the "camel syndrom".
>It means that the formula used by PhotoShop doesn't give a good result with
>the 10x10 pixels image resize test (Kris's test image).
>I have resized the 10x10 pixels image by 123 % in PSP 7.04 and in PS 5.0
>I have put the result here :
>http://cf.photos.yahoo.com/jeanlusque
>If you want load them, the 4 images are not big (from 23 K to approx. 90K) :
>"original", "pixel resize", "binilear" (sic) and "bicubic".
I looked at the comparison images. Are those miniscule differences
what Mr.McEwen has been setting his hair on fire over?
Ember
Donald F. MacLean
> This one is quite handy, because you can cut and paste the source image
> into PSP and then implement the various 450% resizing algorithms it
> claims to offer, compare the result and then see how short changed PSP
> is leaving you:
> http://www.dpreview.com/learn/Glossary/Digital_Imaging/Interpolation_01.h
> tm
snip
I would like to enter the controversy over the quality of the resizing
algorithms in PSP 7.4. When I read Mr McEwens's comment above, I went to
the site suggested to see what it was all about, did as he said,
captured the images, and attempted to replicate the resizing to 450%.
First of all, to try to keep a level playing field, I used the magic
wand to select the borders round the images on the site, inverted the
selection, and cropped to selection to start clean. The results are
posted at abcg - in PSP format (7.x)
Image 1 is the original small image.
Image 2 is the example bi-linear resize at 450%. Image 4 is the PSP
bilinear resize of the little image - again at 450%
Image 3 is the example bicubic rezize, with the PSP equivalent as image 5.
To my eye, there are some additional jaggies in the PSP bilinear upsize,
however it's not too different AND the bilinear method is NOT
recommended for upsizing - only downsizing.
There is little difference that I can see between the two bicubic examples.
I tried to use the arithmetic function to compare the outputs, but the
examples given on the site do not appear to be directly based on the
little image. So an overlay is not coincident.
To cut to the chase, I don't agree that PSP 'sells us short'
Note - I am NOT an employee of JASC :)
[I apologize for the typos in the abcg posting - I think I've caught
them all here ;)]
DFM
Kennedy McEwen
Laderoute <jflad...@shaw.ca> writes
>
>You are missing the point here - it is not Kris' job to answer your
>questions in this forum (nor anyone else's). Neither you nor I know
>what Kris' area of program development is, nor whether he had/has/will
>have any input into the resizing algorithms.
>
he has had any imput to the resizing algorithms, he did that all by
himself by chirping in with totally misleading information in an attempt
to divert attention AWAY from the question. It would have been easier
and simpler for all involved, including his employer, if instead of
attempting deception he had just answered it straight. If he isn't
empowered to do so then he should have left it well alone.
Kennedy McEwen
<REMe...@operamail.com> writes
>
>I looked at the comparison images. Are those miniscule differences
>what Mr.McEwen has been setting his hair on fire over?
No, they are not - try reading the thread before posting stupid
comments.
Kennedy McEwen
<j...@wanadoo.be> writes
>Hello !
>I'm not a mathematician...
>I'm working with my eyes for judging what gives a better result.
>I' don't know what is the "legal" or "correct" or "unique" formula for
>bilinear or bicubic resizing.
>But I can see what PhotoShop makes and I'm not sure that PhotoShop is the
>reference !
>Honestly, I prefer the PSP system in bicubic rendering (and pixel resize
>too).
>I have observed that PhotoShop suffers from the "camel syndrom".
>It means that the formula used by PhotoShop doesn't give a good result with
>the 10x10 pixels image resize test (Kris's test image).
>I have resized the 10x10 pixels image by 123 % in PSP 7.04 and in PS 5.0
>I have put the result here :
>http://cf.photos.yahoo.com/jeanlusque
>If you want load them, the 4 images are not big (from 23 K to approx. 90K) :
>"original", "pixel resize", "binilear" (sic) and "bicubic".
>You can see in the PS5 "pixel resize" test that 2 humps appear on the
>diagonal line. They don't appear in PSP7.
>In PS5 "bicubic resize" those 2 humps are visible too (not in the PSP7
>resize). For me, it is a defect in PhotoShop and PSP7 is better for this
>rendering.
>
from in the resizing process - the top left corner or the centre. You
will find similar effects in either program (for the functions which
actually work!) if you use a different start or end size. Starting with
10x10 pixels and resizing up by 123% results in 12x12 pixels - for pixel
resize those extra two pixels have to be input somewhere.
You will note that the PS result is symmetric about the centre whilst
the PSP result isn't - as a consequence of different rounding in the
vertical and horizontal axes. The same offset between horizontal and
vertical is present, although softer, in the bicubic results. It is
just where the programmer sets his data origin - not a difference in
algorithm per se, and certainly not an algorithm failure as in PSP
bilinear scaling of more than 200%.
Try comparing bilinear scales of 320%, not 120%, in both PSP and PS if
you want to see the difference I have been questioning. On properly
interpolates BILINEARLY, the other just uses nearest neighbour for 200%
and then applies bilinear for the remaining factor of 110%. You can
even prove this for yourself by implementing a 200% pixel resize
followed by 110% bilinear resize and comparing with a single 320%
bilinear - they are identical! The same applies to any other scale,
whether integer or not.
Sally Beacham
> In article <ajf65...@enews1.newsguy.com>, Sally Beacham
> <p...@tension.you> writes
> >
> >Methinks you doth philosophize and proselytize too much. PSP does for me
> >what I need it to do. I'm happy. I have no intention of looking to get
> >unhappy.
> >
> I don't care one way or the other - Sally Beacham's personal happiness
> has no influence on my life whatsover, and it is rather arrogant of you
> to assume that it should - so please indicate where I asked you to get
> unhappy.
Please indicate where I indicate that you asked me to get unhappy, sock
puppet.
If recognising that PSP is flawed in at least this aspect
> makes you unhappy to the extent that you would rather ignore the
> possibility then that is your problem to deal with, not mine. A lot of
> people in the world are quite happy to live in squalor - it doesn't mean
> I should as well!
I'm not the slightest bit unhappy with PSP (well, except for that little
background layer thing, but then again, I love my daughter, except for that
little belly button ring thing, too.) So I don't have a problem, now do I?
Even if sock puppets would like to give me one (right after they finish
haranguing Jasc at large.)
>
> >If you'd like a kinder/gentler reception, perhaps you ought to check out
the
> >water before you try to dynamite the fish. It's not the wisest action in
> >the world to accuse people you don't know, and companies you aren't
familiar
> >with, of lying and obfuscating, when in fact you want them to do
something
> >that to this point, I've only seen YOU make an issue of, after lo these
many
> >years of the software's existence.
> >
>
> I don't give a damn what the reception is, I didn't come here for a tea
> party but to have a specific question about a FAULT in the program
> answered. I used PSP for years, since Version 2.01 according to my
> records (1991?), but possibly even before that and have worked in
> digital image processing for over 25 years. I am MORE than familiar
> with Jasc having provided them with considerable income over the years,
> although in a professional context and not in a personal one. I looked
> in here because I recently downloaded Version7 of PSP, having rejected
> it several years ago
How very odd... you define less than 2 years as "several"? My, I would
assume someone with your insistence on preciseness would have had the exact
date of release of PSP 7 at your fingertips.
>due to other flaws, under some pressure to do so by
> exactly the sort of people who are wittering here that it is good enough
> for the job. You don't have a clue about my application so you haven't
> a clue whether this flaw can or should be ignored. You should have
> enough common sense to realise that I wouldn't be wasting my time in
> this group if I thought that it could!
I have plenty of common sense, and oftentimes I use it to get people to do
what I want them to do, by leading them to water, not hitting between the
eyes with a sledge hammer. But hey, if you NEED a sledge hammer, I can
swing one.
You are free to waste your time any way you desire, sock puppet.
Sally Beacham
> In article <ajgqp9$1aqd63$1...@ID-73017.news.dfncis.de>, Steve-A
> <lake_vu...@hotmail.com> writes
> >
> >You have purchased THAT many copies?!?
> >
> In the past, YES, and some cretins in IT want to purchase some more
> rather than the software that does the job properly!
Perhaps the cretins in IT have instructions from the morons in Accounting
re: the budget in these precarious economic times.
p.s. I bet you are even MORE popular at work than you are here. Remember,
if you can't schmooze 'em with your personality, you better be able to
bolster the bottom line.
Sally Beacham
> In article <ajgqvc$1bbfhr$1...@ID-73017.news.dfncis.de>, Steve-A
> <lake_vu...@hotmail.com> writes
> >
> >Hmm, I don't come in here acting like a spoiled child that stomps my
> >feet on the floor if I don't get what I want if that's what you mean.
> >
> Seems to me like Kris and you are the ones stomping their feet. I have
> asked him a simple question and given him more than adequate time to
> reply.
Listen, Einstein - this is Usenet. This isn't the Jasc company hotline.
Kris is here as a "private citizen" and as such as the same rights as
everyone else - he doesn't have to do anything/ say anything/be anything/or
damn well exist AT ALL in here if he doesn't want to/doesn't have the
time/fell off a turnip truck.
What makes you think you have the right to waltz in here and DEMAND he do
anything? Why don't you just trot next door to comp.graphics.apps.photoshop
and see how far your little hissy fits will get you in THERE?
Considering he works like a dog, and then comes here all hours of the day
and night, you should consider yourself LUCKY he even took the time to say
hello, when he probably could have used some well-earned sleep!
Puppet.
Bob Dietz
>
> Try comparing bilinear scales of 320%, not 120%, in both PSP and PS if
> you want to see the difference I have been questioning. On properly
> interpolates BILINEARLY, the other just uses nearest neighbour for 200%
> and then applies bilinear for the remaining factor of 110%. You can
> even prove this for yourself by implementing a 200% pixel resize
> followed by 110% bilinear resize and comparing with a single 320%
> bilinear - they are identical! The same applies to any other scale,
> whether integer or not.
200% * 110% = 220%
At least it did when I when to school.
Did you mean 320% bilinear compared with 200% pixel resize followed by
160% bilinear resize?
I do see what you are complaining about, but may I ask why you want to
use a bilinear resize rather than bicubic resize when up-scaling an image?
Just wondering.
Bob
Uni
>
> "Kennedy McEwen" <r...@nospam.demon.co.uk> wrote in message
> news:LAXTtINNe$W9E...@kennedym.demon.co.uk...
> > In article <ajgqvc$1bbfhr$1...@ID-73017.news.dfncis.de>, Steve-A
> > <lake_vu...@hotmail.com> writes
> > >
> > >Hmm, I don't come in here acting like a spoiled child that stomps my
> > >feet on the floor if I don't get what I want if that's what you mean.
> > >
> > Seems to me like Kris and you are the ones stomping their feet. I have
> > asked him a simple question and given him more than adequate time to
> > reply.
>
> Listen, Einstein - this is Usenet. This isn't the Jasc company hotline.
> Kris is here as a "private citizen" and as such as the same rights as
> everyone else - he doesn't have to do anything/ say anything/be anything/or
> damn well exist AT ALL in here if he doesn't want to/doesn't have the
> time/fell off a turnip truck.
>
> What makes you think you have the right to waltz in here and DEMAND he do
> anything? Why don't you just trot next door to comp.graphics.apps.photoshop
> and see how far your little hissy fits will get you in THERE?
>
> Considering he works like a dog,
Oh, yes, we can see new software coming out of Jasc by the
truckloads!!!!!!!
Z' = Zzzzzzzzz -L-
Uni
Kris Zaklika
>
> In article <7mdlluklia48ftutb...@4ax.com>, Jackie
> Laderoute <jflad...@shaw.ca> writes
> >
> >You are missing the point here - it is not Kris' job to answer your
> >questions in this forum (nor anyone else's). Neither you nor I know
> >what Kris' area of program development is, nor whether he had/has/will
> >have any input into the resizing algorithms.
> >
> I have not suggested that it is Kris's job to answer questions, or that
> he has had any imput to the resizing algorithms, he did that all by
> himself by chirping in with totally misleading information in an attempt
> to divert attention AWAY from the question. It would have been easier
> and simpler for all involved, including his employer, if instead of
> attempting deception he had just answered it straight. If he isn't
> empowered to do so then he should have left it well alone.
The question was "Could someone explain why no differences
between bilinear and pixel resize ? Is it the same system
with two different names ? Why those 2 options because
they give exactly the same result ?"
I explained that they do not give the same result and how
to see that. In consequence of this simple answer you
accused me (and Jasc) of deception, lies, misleading and
so on. You are a rude and deranged individual with whom I
will not communicate.
> --
> Kennedy
> Yes, Socrates himself is particularly missed;
> A lovely little thinker, but a bugger when he's pissed.
> Python Philosophers (replace 'nospam' with 'kennedym' when replying)
--
----------------------------------------------------------------------
Kris Zaklika Jasc Software, Inc. The
Product Ideas: id...@jasc.com Power
Customer Service: customer...@jasc.com To
Technical Support: tec...@jasc.com Create
----------------------------------------------------------------------
Uni
>
> Kennedy McEwen wrote:
> >
> > In article <7mdlluklia48ftutb...@4ax.com>, Jackie
> > Laderoute <jflad...@shaw.ca> writes
> > >
> > >You are missing the point here - it is not Kris' job to answer your
> > >questions in this forum (nor anyone else's). Neither you nor I know
> > >what Kris' area of program development is, nor whether he had/has/will
> > >have any input into the resizing algorithms.
> > >
> > I have not suggested that it is Kris's job to answer questions, or that
> > he has had any imput to the resizing algorithms, he did that all by
> > himself by chirping in with totally misleading information in an attempt
> > to divert attention AWAY from the question. It would have been easier
> > and simpler for all involved, including his employer, if instead of
> > attempting deception he had just answered it straight. If he isn't
> > empowered to do so then he should have left it well alone.
>
> The question was "Could someone explain why no differences
> between bilinear and pixel resize ? Is it the same system
> with two different names ? Why those 2 options because
> they give exactly the same result ?"
>
> I explained that they do not give the same result and how
> to see that. In consequence of this simple answer you
> accused me (and Jasc) of deception, lies, misleading and
> so on. You are a rude and deranged individual
Hey, Tone, maybe this is why that UK magazine article was published
about PSP. It's not all Jasc customers that get to be called "deranged"
by one of their employees!!!
I wonder if Jasc's technical support is just as nice?
You go, Z'!!!
Uni
Uni
Hey, Z', just curious. Since we couldn't see you, did you give this
customer the single finger salute as well?
Uni
Jackie Laderoute
>It's not all Jasc customers that get to be called "deranged"
>by one of their employees!!!
No, you can buy an Adobe product and be accused of "hallucinations" by
Chris Cox.
150820022239408486%cc...@mindspring.com
And that's not even the foaming at the mouth he does with Timo. ;-)
And for that, you get to pay six times the price.
Uni
You know, Z', now I have it figured why all the regulars admire you.
They too like to beat people up in this newsgroup! It must have felt
good to tell Kennedy off in a public newsgroup, rather than by e-mail,
eh?
Uni
Ember
<r...@nospam.demon.co.uk> wrote:
>In article <mvfolukvbm64lssi2...@4ax.com>, Ember
><REMe...@operamail.com> writes
>>
>>I looked at the comparison images. Are those miniscule differences
>>what Mr.McEwen has been setting his hair on fire over?
>
>No, they are not - try reading the thread before posting stupid
>comments.
I wasn't talking to you. And don't ever tell me how or what to post.
Uni
See, Kris, how well you taught Sally Beachedwhale?
You have her well trained, just like an attack dog!
Uni
Uni
>
> On Thu, 15 Aug 2002 19:48:45 +0100, Kennedy McEwen
> <r...@nospam.demon.co.uk> wrote:
>
> >In article <ajgqvc$1bbfhr$1...@ID-73017.news.dfncis.de>, Steve-A
> ><lake_vu...@hotmail.com> writes
> >>
> >>Hmm, I don't come in here acting like a spoiled child that stomps my
> >>feet on the floor if I don't get what I want if that's what you mean.
> >>
> >Seems to me like Kris and you are the ones stomping their feet. I have
> >asked him a simple question and given him more than adequate time to
> >reply. Instead he would rather counter with misleading information and
> >deception before sulking off. You, on the other hand, have added
> >nothing to the discussion but aggression. Can we expect to see any
> >positive contribution from you some time in the current millennium?
>
> You are missing the point here - it is not Kris' job to answer your
> questions in this forum (nor anyone else's). Neither you nor I know
> what Kris' area of program development is, nor whether he had/has/will
> have any input into the resizing algorithms.
>
> If you want technical aspects of the program clarified for you, then
> contact Jasc directly. No employee of the company is likely willing or
> legally able to discuss these matters with you in a public forum.
>
> I suggest you take Ron's excellent advice and go directly to Jasc with
> your issues
I guess Kris couldn't spare 20 seconds of his valuable time to suggest
this himself.
But as I know, Kris really wanted to tell this poor soul off here! It's
more enjoyable that way.
Uni
Jackie Laderoute
Still obsessed with me, Uni?
And here I thought you'd shot your bolt yesterday.
But do keep your fantasies straight - if you want to post your
confabulations about Kris, do it in a follow-up to him. And here I
thought I was your number one target.
Of course, given your track record of lying, misrepresentation and
fabrication, you can't expect anyone to take you seriously. And you
make it so obvious, too.
Silly boy.
Uni
Fitting in well, I see, Ember! Lucky your gun isn't loaded!
Uni
Uni
Jean-Luc ERNST
I'm not a mathematician...
I'm working with my eyes for judging what gives a better result.
I' don't know what is the "legal" or "correct" or "unique" formula for
bilinear or bicubic resizing.
But I can see what PhotoShop makes and I'm not sure that PhotoShop is the
reference !
Honestly, I prefer the PSP system in bicubic rendering (and pixel resize
too).
I have observed that PhotoShop suffers from the "camel syndrom".
It means that the formula used by PhotoShop doesn't give a good result with
the 10x10 pixels image resize test (Kris's test image).
I have resized the 10x10 pixels image by 123 % in PSP 7.04 and in PS 5.0
I have put the result here :
http://cf.photos.yahoo.com/jeanlusque
If you want load them, the 4 images are not big (from 23 K to approx. 90K) :
"original", "pixel resize", "binilear" (sic) and "bicubic".
You can see in the PS5 "pixel resize" test that 2 humps appear on the
diagonal line. They don't appear in PSP7.
In PS5 "bicubic resize" those 2 humps are visible too (not in the PSP7
resize). For me, it is a defect in PhotoShop and PSP7 is better for this
rendering.
PSP and PS are not using exactly the same formulae. Perhaps because patent
reason : Corel PhotoPaint9 in "pixel resize" creates 2 humps too, one near
the center of the line (after 4 pixels) and one at the end bottom right.
It seems that each software has a different formula.
Note : during my tests, I observed that PaintShop Pro was able to zoom at
3200 % on screen but PS5 was limited to 1600% zoom.
Friendly,
Jean-Luc
"Kennedy McEwen" <r...@nospam.demon.co.uk> a écrit dans le message de news:
zjet1qDm...@kennedym.demon.co.uk...
> In article <3D550E41...@jasc.com>, Kris Zaklika
> <kzak...@jasc.com> writes
--
...[snip]...
> The real answer is that PSP doesn't implement bilinear resizing any
> more. Instead, it implements something which is not even a close
> approximation of bilinear resizing. Given your politician like response
> to divert attention away from the odd effect of the function, I suspect
> that you are only too aware of its obvious deficiencies!
>
> >After following my suggestions you can probably draw your
> >own conclusions about these questions :)
> >
> My own conclusion is that if this is what you intended it to do, Kris,
> then you really should call it something unique to PSP, because it
> certainly *ISN'T* bilinear resizing, as used throughout the industry
> long before either PSP (or even your expensive competitor) came on the
> scene!
>
> Bilinear resizing is a mathematical function that is well defined and
> certainly doesn't give the results that PSP7 produces for resizing
> scales greater than x2. If my memory serves me correctly, bilinear
> resizing has been in common usage for more than 40 years. I know I
> certainly used it on a PDP11 image processing suite over 20 years ago
> and I don't think I have seen such that operation ever give the effects
> of PSP7!
>
> Bilinear resizing should fit new pixels according to where they lie on a
> flat linear surface drawn between the original four original samples
> which encompass the new sample. That is WHY it is called bilinear. Two
> original points in each axis - a LINEAR fit between them in both axes.
> Anything other than that is not bilinear!
>
> PSP7 APPEARS to use pixel resizing to get up to the integer scaling
> below the required factor and then bilinear resizing to implement the
> remaining fractional part. This is why, when there is no fractional
> part, it is indistinguishable from the pixel resizing result - exactly
> as Jean-Luc noted. That is a guess based on what I see from some
> samples I have tried although for some percentage scalings and pixel
> sized images it doesn't even seem to be x-y symmetric!
>
> I am sure you can tell us exactly what is going on, but IT ISN'T
> BILINEAR RESIZING! It's a PSP specific algorithm (albeit rather less
> useful than bilinear resizing, as Jean-Luc noted), so give it a unique
> PSP name, not one that is an industry standard for a completely
> different function! That is just asking for trouble and only likely to
> cause confusion for anyone using your software seriously.
>
> Strangely, the last version of PSP that I seriously used (Version 4.14)
> actually implemented bilinear scaling pretty much correctly, so I am
> tempted to assume that this is a programming bug rather than a new
> scaling function that Jasc have simply misnamed in the hope that nobody
> will notice. ;-)
>
> Do you implement bicubic correctly or does that use another programming
> shortcut that has gone wrong as well?
Donald F. MacLean
gaussian blur of 1.4 to my image 4 - PSP bilinear - resulted in
[subjectively] an almost identical appeaance to the test image [image2]
Perhaps JASC <might/could> incorporate such a blur in their basic
algorithmic approach to bilinear enhancement, but my reading of this and
other forums seems to indicate that a judicious use of blurring is
adviseable when resizing is done? Also, as has again been pointed out
many times by Kris and others, resizing upwards necessarily entails
<inventing> data where none exist.
So I STILL don't understand your assertion that JASC is selling its
customers short, and lying to them, or deceiving them, in any way.
[I accessed the other sites you mention - unfortunately my maths are no
longer up to snuff enough to understand them - let alone comment!!! :)]
DFM
Donald F. MacLean
solid black image. However, as noted before, the original 100% and
450% images do not appear to <sync> perfectly, so there remains a
lighter <ghosting> around the iris and the pupil, etc. Otherwise - good
enough for government work! :)
DFM
Kennedy McEwen
>To my eye, there are some additional jaggies in the PSP bilinear
>upsize, however it's not too different AND the bilinear method is NOT
>recommended for upsizing - only downsizing.
>
applied on the fractional part of the upsize. Bilinear is, by
definition, as capable of upsizing as bicubic, however PSP recommends
using bicubic because it produces smoother results. That is not - and
never has been in question - the issue is why they choose not to
implement bilinear correctly, or why they choose to call what they do
implement "bilinear" when it is not.
>There is little difference that I can see between the two bicubic examples.
>
Bicubic isn't an issue.
>I tried to use the arithmetic function to compare the outputs, but the
>examples given on the site do not appear to be directly based on the
>little image. So an overlay is not coincident.
>
>To cut to the chase, I don't agree that PSP 'sells us short'
>
It does if the algorithm that you are looking for is bilinear = you get
a result which is mainly pixel resize, as you just discovered. How
*important* that is to you depends on what you are trying to do with the
result. In my case it is.
Kennedy McEwen
<p...@tension.you> writes
>
>"Kennedy McEwen" <r...@nospam.demon.co.uk> wrote in message
>news:sxpKB$L2U$W9E...@kennedym.demon.co.uk...
>> In article <ajgqp9$1aqd63$1...@ID-73017.news.dfncis.de>, Steve-A
>> <lake_vu...@hotmail.com> writes
>> >
>> >You have purchased THAT many copies?!?
>> >
>> In the past, YES, and some cretins in IT want to purchase some more
>> rather than the software that does the job properly!
>
>Perhaps the cretins in IT have instructions from the morons in Accounting
>re: the budget in these precarious economic times.
>
claims to then the cost in rework or wrong development expenditure that
arise from its use more than offset any saving in its procurement. That
is why doing what you say you do CORRECTLY is important.
Kennedy McEwen
<p...@tension.you> writes
>
>
>You are free to waste your time any way you desire, sock puppet.
>
know to be faulty. If anyone is a "sock puppet" here Sally, it is the
people who do exactly the "yes its faulty but its good enough for me so
why do you care if neither I nor Jasc do" attitude - JUST LIKE YOU HAVE!
When the emperor is told he has no clothes it isn't the person telling
him that is cow towing.
Kennedy McEwen
<kzak...@jasc.com> writes
>
>The question was "Could someone explain why no differences
>between bilinear and pixel resize ? Is it the same system
>with two different names ? Why those 2 options because
>they give exactly the same result ?"
>
>I explained that they do not give the same result and how
>to see that. In consequence of this simple answer you
>accused me (and Jasc) of deception, lies, misleading and
>so on. You are a rude and deranged individual with whom I
>will not communicate.
>
NOT address it. Instead you replied with "you should rescale to this
size and you will see that it works", which merely diverted attention
from the fact that it didn't work at all at the original scale or
anywhere near it. My question was whether this was a bug (ie. something
that will be fixed now you are aware of it) or the wrong use of an
industry standard term. You have answered that at all - if you have,
please refer to the post in which you did, because it has neither
reached my server nor the Usenet archives.
Fred Hiltz
[snip]
> never has been in question - the issue is why they choose not to
> implement bilinear correctly, or why they choose to call what they do
> implement "bilinear" when it is not.
Kennedy, we get your point. Further repetition is not likely to convince any
more of us that it is an important one, nor will further ad hominem attacks.
Your writing reveals an intelligent person. I'm looking forward to your
contributions--even constructive criticism--on other topics. Welcome to the
group.
--
Fred Hiltz, fhiltz at yahoo dot com
Sally Beacham
> In article <ajhih...@enews4.newsguy.com>, Sally Beacham
> <p...@tension.you> writes
> >
> >
> >You are free to waste your time any way you desire, sock puppet.
> >
> I certainly am - but then I don't waste it trying to defend something I
> know to be faulty. If anyone is a "sock puppet" here Sally, it is the
> people who do exactly the "yes its faulty but its good enough for me so
> why do you care if neither I nor Jasc do" attitude - JUST LIKE YOU HAVE!
> When the emperor is told he has no clothes it isn't the person telling
> him that is cow towing.
It's not my issue, it's your issue. If you don't think it's good enough FOR
YOU, you have been told countless times the appropriate channels to use to
express it. You persist in ranting hysterically here, where absolutely
NOTHING can be done about it. You pose as a business person with an
important stake in the matter, yet you don't approach it from a
business-like standpoint.
Kris has informed you you aren't worthy of his further response, and I'm
taking a cue from him. You can rave to your heart's content, but it's
falling on deaf ears.
Kennedy McEwen
<REMe...@operamail.com> writes
IMAGINED. Any responses, other than from me, to the question you asked
can only be mere conjecture on the part of the responder. You appear to
have a problem with receiving a response from the ONLY PERSON capable of
providing the facts rather than conjecture.
Read the thread and you will find out what the topic really is about
rather than what you or anyone else IMAGINES it is about. Even better,
print it out and save it - it will be many years before you are mature
enough to tell the difference!
Kennedy McEwen
<mac...@capitalnet.com> writes
>Further to my earlier post here, and on abcg, I found that applying a
>gaussian blur of 1.4 to my image 4 - PSP bilinear - resulted in
>[subjectively] an almost identical appeaance to the test image [image2]
>
blur, after the PSP 'bilinear' resize then you will approach the correct
result from a true bilinear resize and this is also true if you start
with pixel resize (which is what the major component of PSP algorithm is
in large scalings. However, a gaussian blur will not give you the
correct result, only something that looks roughly right. The reason it
looks almost right is that it filters out the higher harmonics that
pixel resize introduces - but it does this at the expense of also
filtering some of the fundamental image content as well if the gaussian
spread is large enough to cut out all of the erroneous content.
You can string several steps together to get true bilinear rescaling in
PSP, but only over a limited range and only for integer scalings. First
start with pixel resizing and then use a "User Defined" effect with the
terms in the matrix set to 1 for the scale of resize that you have
implemented. For example, for a 2x resize, set the central 2x2 terms to
1 and all others to zero, for a 3x resize set the central 3x3 terms, and
so on. Then set the division factor to the total of the terms in the
matrix and the bias to 0. I used to use this technique on early
versions of PSP when only pixel resize was available, but it does have
its limitations, such as being limited to the size of the user defined
filter and to integer scalings, which were further compounded by bugs in
the User Defined filter which have since been corrected.
> Perhaps JASC <might/could> incorporate such a blur in their basic
>algorithmic approach to bilinear enhancement, but my reading of this
>and other forums seems to indicate that a judicious use of blurring is
>adviseable when resizing is done? Also, as has again been pointed out
>many times by Kris and others, resizing upwards necessarily entails
><inventing> data where none exist.
>
No - adding steps to an already faulty algorithm is no solution. Would
you put up a new building on foundations that have already cracked? The
solution is to correct the algorithm first, so that it implements what
it claims to do. If anyone then wants to add additional filtering to
the result then they have the freedom to do that, but forcing it into
the basic algorithm isn't a fix for the existing errors, only a bodge.
>So I STILL don't understand your assertion that JASC is selling its
>customers short, and lying to them, or deceiving them, in any way.
>
OK, here are the steps:
Jasc claim that PSP implements bilinear resizing directly.
You have seen for yourself that this is not the case - you have to bodge
it to get something that is even approximately correct, but it isn't
bilinear resizing.
Claiming you do something that you don't is a lie. (Kris and Sally may
not like this fact, but fact it is - whether the lie is intentional or
not is irrelevant, it remains a lie.)
Directing customers to specific examples where the errors in the
algorithm just cancel out or reduce to zero in an attempt to claim that
the algorithm is not defective is deception.
>[I accessed the other sites you mention - unfortunately my maths are no
>longer up to snuff enough to understand them - let alone comment!!! :)]
>
If you are struggling with the maths, try plotting the value of a few
pixels on graph paper with the pixels spaced apart by the number of
units that you are scaling the image by. Then draw straight lines
between the pixels. The values of the new pixels, interpolated
linearly, are the height of the line in each of the columns of cells.
That gives you the scaling in one axis for one row of pixels, and the
data should correspond EXACTLY with the bilinear scaled pixel amplitudes
between original samples in PSP. I don't suggest you go much further
than this because it will involve a LOT of graph drawing, but you can if
you feel you need to in order to understand what is going on. If you
repeat this for all of the rows then you have scaled the entire image in
one axis. Then repeat the process in the other direction using the
scaled information and you get bilinear, or bi-directional linear,
scaling.
Showing bicubic scaling in this graphical manner is also possible, but
much more complicated to do without involving maths because it requires
curves of specific types to implement the interpolation between the
original samples.
The whole point of these interpolation schemes is to retain as much
information of the original image when scale it up without scaling the
pixel artefacts that the image is artificially composed of. Pixel
resizing scales the image and all of the pixel artefacts together.
Bilinear scales none of the pixel artefacts, but loses some of the image
as well. Bicubic retains more of the image, but also more of the pixel
artefacts and so on. In principle, a function known as sinc
interpolation would scale the image only, but this is very complex to
implement and impossible to achieve complete accuracy on a finite sized
image. Some scaling algorithms use various approximations to sinc
interpolation though, such as "Hanning Windowed Sinc" for example, which
comes very close to the ideal - although probably not of sufficient
benefit over bicubic to be of interest to the casual image processing
user. However each type of scaling has specific benefits and drawbacks,
and falsely claiming that you implement one when you actually do
something completely different only serves to cause confusion as to what
these benefits and drawbacks are, which can be problematical when people
who aren't prepared to go through the mathematics (like most users here,
probably) start to make assumptions from what they see. This is
compounded when others refer only to the erroneous algorithm as one of
the capabilities of a particular package, as you have probably seen on a
few of the websites I referenced.
Kennedy McEwen
<p...@tension.you> writes
>
>
>It's not my issue, it's your issue.
Actually, it is Jasc's issue, and since it is an issue that didn't
bother you before, why should my discussion of it mean you have a
problem with it now? Your idyllic world of ignorance can remain
unchanged - that is under your control. As I said before, I don't care
how you feel about PSP. Even if it was a crock of ****, which it
certainly isn't IMO, then some people would still be happy with it. The
fact that you are happy with the product does not mean that I or anyone
else has to be, or that we should not discuss questions and details of
its defects on this unmoderated newsgroup.
Since when was Sally Beacham appointed Usenet moderator-in-chief?
You really crack me up Sally. You object to someone highlighting faults
or bugs in the application and then you throw around childish names in a
vain attempt to appear adult. Names which, incidentally, are more
befitting of someone under a juvenile infatuation with PSP!
> If you don't think it's good enough FOR
>YOU, you have been told countless times the appropriate channels to use to
>express it.
Are you totally unable to read? I have said often enough that I stopped
using PSP because of its bugs which, incidentally remained through many
upgrades despite being reported directly to Jasc at the time.
Unfortunately, many people like you would rather stop all discussion of
defects because it doesn't affect you, so questions like the one which
initiated this thread go unanswered or worse, deceptively answered by
Jasc staff, as was precisely the case with Kris's response.
If you are happy with the product then you have nothing to contribute to
a thread which discusses ONE of its defects - but then you haven't
really contributed anything to the discussion anyway, just muddied the
waters with personal venom. Quite a common trait of those who
contribute nothing to society in general, unfortunately.
Kennedy McEwen
<rbdiet...@yahoo.com> writes
>Kennedy McEwen wrote:
>> Try comparing bilinear scales of 320%, not 120%, in both PSP and PS
>>if you want to see the difference I have been questioning. On
>>properly interpolates BILINEARLY, the other just uses nearest
>>neighbour for 200% and then applies bilinear for the remaining factor
>>of 110%. You can even prove this for yourself by implementing a 200%
>>pixel resize followed by 110% bilinear resize and comparing with a
>>single 320% bilinear - they are identical! The same applies to any
>>other scale, whether integer or not.
>
>200% * 110% = 220%
>At least it did when I when to school.
>
>Did you mean 320% bilinear compared with 200% pixel resize followed by
>160% bilinear resize?
>
deduced. ;-)
>I do see what you are complaining about, but may I ask why you want to
>use a bilinear resize rather than bicubic resize when up-scaling an image?
>Just wondering.
>
To quote George Mallory, "Because it's there!" ;-)
Several reasons actually, from both professional and amatuer
perspectives.
Professionally, I design real time imaging sensors, some of which offer
internal image processing functions such as electronic zoom etc. This
is normally implemented using simple pixel resize, but the image quality
is far better when some interpolation scheme is used, just like you get
on your domestic video camera. Bubic would be nice, but impossible to
implement within the speed and power constraints of the system. Bilinear
fits within these constraints but implementing it in the design means
additional effort and budget. It doesn't help when some little oik in
accounts hauls out his personal copy of PSP and "proves" that it gives
no benefit over the pixel resize that is already in the product! It is a
big bad world out there and no doubt our competitors will soon have
products which do implement such a feature, many higher power slower
pixel rate systems already do, which would leave us at a disadvantage.
Fortunately this very thread has raised questions about Jasc's
credibility with that particular individual, who now realises that he
should not use toys for professional work. Hence my comment about
Kris's attitude already having served my purpose admirably. ;-)
My other reason is more for personal use of image manipulation programs.
When scaling an image the bilinear algorithm completely rejects pixel
structure which is scaled entirely with pixel resize and at reduced
amplitude with bicubic. Of course, bilinear achieves this at the
expense of some image content, however this is often irrelevant, for
example when the image resolution being somewhat less than the full
sampling density. So for certain images, print scalings and viewing
distances bilinear can actually produce a better result than either
bicubic or pixel resize - if it is implemented correctly. Implementing
it wrong means that this middle ground can never be achieved. :-(
I have been pretty amazed at the reaction of certain individuals on this
forum though as they attempt to defend the indefensible. The function
is, without question, flawed in PSP. That the flaw is insignificant to
many is immaterial, to others it is important and more so when certain
references direct users to PSP BECAUSE it offers to implement this
particular function. Who knows how many have just concluded that they
are better off with the bundled version of PS5 or PS Elements as a
consequence of the way PSP implements this or other faulty functions. I
know that I deserted the program some years ago because an error in the
"User Defined" filter remained uncorrected over several version
upgrades. I would have expected devotees of PSP (like Ms Beeacham) to
be more than keen to have all known deficiencies corrected - but no,
they seem to relish the fact that PSP is rejected by many in preference
to its competitors because of the flaws.
Bob Dietz
> In article <3D5C5EEE...@yahoo.com>, Bob Dietz
> <rbdiet...@yahoo.com> writes
>
>> use a bilinear resize rather than bicubic resize when up-scaling an
>> image?
>> Just wondering.
>>
> To quote George Mallory, "Because it's there!" ;-)
While I can empathize with the sentiment, experience has taught me that
business entities regard such points as moot. An image of Don Quixote
springs to mind.
> Several reasons actually, from both professional and amatuer perspectives.
>
> Professionally, I design real time imaging sensors, some of which offer
> internal image processing functions such as electronic zoom etc. This
> is normally implemented using simple pixel resize, but the image quality
> is far better when some interpolation scheme is used, just like you get
> on your domestic video camera. Bubic would be nice, but impossible to
> implement within the speed and power constraints of the system. Bilinear
> fits within these constraints but implementing it in the design means
> additional effort and budget. It doesn't help when some little oik in
> accounts hauls out his personal copy of PSP and "proves" that it gives
> no benefit over the pixel resize that is already in the product! It is a
> big bad world out there and no doubt our competitors will soon have
> products which do implement such a feature, many higher power slower
> pixel rate systems already do, which would leave us at a disadvantage.
Accountants will always require you to account for expenditures and
proposed expenditures. Generally, the challenge is educate them
sufficiently that they are able to see the need for the expenditures.
I can certainly see where PSP's nomenclature didn't expedite the process
for you in this instant.
> Fortunately this very thread has raised questions about Jasc's
> credibility with that particular individual, who now realises that he
> should not use toys for professional work.
I wonder if that "little oik in accounts" has continued to read this
thread. If so, I have little doubt regarding his appreciation of being
call a "little oik."
> Hence my comment about
> Kris's attitude already having served my purpose admirably. ;-)
The first 'attitude' I noticed in this thread was when you said,
"Given your politician like response to divert attention away from
the odd effect of the function, I suspect that you are only too aware
of it's obvious deficiencies!"
Here you assigned a purpose and intent to Kris' post that I did not
and do not see.
> My other reason is more for personal use of image manipulation programs.
> When scaling an image the bilinear algorithm completely rejects pixel
> structure which is scaled entirely with pixel resize and at reduced
> amplitude with bicubic.
I read most of the links you provided. I don't recall any of them
discussing 'pixel structure.' What do you mean by 'completely rejects
pixel structure?'
> Of course, bilinear achieves this at the
> expense of some image content, however this is often irrelevant, for
> example when the image resolution being somewhat less than the full
> sampling density.
Again, you've lost me with the phrase 'image resolution being somewhat less
than the full sampling density.'
> So for certain images, print scalings and viewing
> distances bilinear can actually produce a better result than either
> bicubic or pixel resize - if it is implemented correctly. Implementing
> it wrong means that this middle ground can never be achieved. :-(
A specific example would be very useful here.
> I have been pretty amazed at the reaction of certain individuals on this
> forum though as they attempt to defend the indefensible.
Is this your first time on usenet? <g>
On a serious note, I think you misconstrue what most of these people have
said. You seem to regard this PSP flaw as a personal affront, rather
than a mere bug to be corrected. You seem to think that we must all
regard this as fatal flaw and conclude that JASC and it's employees
are corrupt and deceptive.
> The function is, without question, flawed in PSP.
It is without question, not the same function outlined in the articles
you sited. Nor is it the same as the "bilinear resample" implemented
in PhotoShop 5.0 LE. (and I presume PhotoShop 5,5.5,6,7 and Elements).
For photographs, PhotoShop's 'bilinear resample' appears to be superior
to PSP's 'bilinear resample.'
> That the flaw is insignificant to
> many is immaterial,
That the flaw is insignificant to many is immaterial to you.
I'd think that JASC accounting oik's would find it highly material.
> to others it is important and more so when certain
> references direct users to PSP BECAUSE it offers to implement this
> particular function.
> Who knows how many have just concluded that they
> are better off with the bundled version of PS5 or PS Elements as a
> consequence of the way PSP implements this or other faulty functions. I
> know that I deserted the program some years ago because an error in the
> "User Defined" filter remained uncorrected over several version
> upgrades. I would have expected devotees of PSP (like Ms Beeacham) to
> be more than keen to have all known deficiencies corrected - but no,
> they seem to relish the fact that PSP is rejected by many in preference
> to its competitors because of the flaws.
All software has flaws.
Bob
Jean-Luc ERNST
Excuse me, but they are not identical.
In PSP7, I had a 40x40 pixels image.
I made a 200% pixel resize.
And I followed with both a 110 % and a 120% bilinear resize (you say 200+110
and for me 200+120 = 320 but I tried both).
They are not the same as the direct 320% bilinear resize.
I suppose the explanation is elsewhere...
Friendly,
Jean-Luc
"Kennedy McEwen" <r...@nospam.demon.co.uk> a écrit dans le message de news:
IZoDbrG3...@kennedym.demon.co.uk...
> In article <ajh9pm$11eff$1...@sinclair.be.wanadoo.com>, Jean-Luc
--
...[snip]...
> Try comparing bilinear scales of 320%, not 120%, in both PSP and PS if
> you want to see the difference I have been questioning. On properly
> interpolates BILINEARLY, the other just uses nearest neighbour for 200%
> and then applies bilinear for the remaining factor of 110%. You can
> even prove this for yourself by implementing a 200% pixel resize
> followed by 110% bilinear resize and comparing with a single 320%
> bilinear - they are identical! The same applies to any other scale,
> whether integer or not.
Jean-Luc ERNST
I'm not so "technician" as you and don't have such specific needs as you.
What I'm trying to test is the "visual". I mean, how the image looks after
a resizing. Which system gives the best visually result.
As I don't know how the result must be according to such or such algorithm,
I can only compare the result given by two programs : Paint Shop Pro 7 and
Photo Shop 5 that I have. I don't know if those softwares are using correct
or not correct algorithms. What I can only judge is the appearance of the
image after resizing.
I have made a test on an image suggested by Kris : 10x10 pixels and 123%
pixel resize and bicubic resize :
Conclusion : visually, Paint Shop Pro 7 gives a better result than Photo
Shop 5. Better because they are less artefacts on the image generated by
PSP7. I'm judging principally the grey line. In PSP7, the line doesn't
have "blobs".
I have made a second test :
http://cf.photos.yahoo.com/jeanlusque
The new image is labelled "bicubic resize" (approx 82K).
The original image was a 40x40 pixels image, white on the left and black on
the right. A 1 pixel width grey line (128,128,128) goes from upper left
corner to lower right corner.
I applied a 123% bicubic resize both in PaintShopPro 7 and PhotoShop 5.
As you can see, the result is better in PaintShopPro. You can see the
defect in PhotoShop5 when you observe the lower part of the grey line on the
black background. The grey along the middle of the line is changed : you
see successive pixels that go from dark grey to light grey and reverse.
With the eyedropper, you may see that the lighter grey is 128,128,128 and
the darker is 91,91,91 (a variation in value of 37). If you measure the
grey pixels in the PSP7 image, you see that they go from 121,121,121 to
99,99,99 (a variation of 22 only). Visually, the PSP7 grey line is more
homogeneous.
Conclusion, PSP7 is visually better in bicubic resizing than PhotoShop5 on
those images.
Friendly,
Jean-Luc
"Kennedy McEwen" <r...@nospam.demon.co.uk> a écrit dans le message de news:
KfEja0aF...@kennedym.demon.co.uk...
--
...[snip]...
Kennedy McEwen
<j...@wanadoo.be> writes
>Hello !
>Excuse me, but they are not identical.
>In PSP7, I had a 40x40 pixels image.
>I made a 200% pixel resize.
>And I followed with both a 110 % and a 120% bilinear resize (you say 200+110
>and for me 200+120 = 320 but I tried both).
>They are not the same as the direct 320% bilinear resize.
>I suppose the explanation is elsewhere...
>
of those numbers. You will get identical results from 200% resize
followed by 160% bilinear. Also the scales multiply as they are
applied, rather than add, so the comparison you are looking for is
200%x160% = 320%.
Kennedy McEwen
<rbdiet...@yahoo.com> writes
>
>I read most of the links you provided. I don't recall any of them
>discussing 'pixel structure.' What do you mean by 'completely rejects
>pixel structure?'
>
as a delta function and the volume under the delta function is the
signal. This might seem a difficult concept to grasp, but think of
making a spot smaller and smaller, then the brightness of the spot needs
to increase inversely to its area for the spot to remain as visible as
it previously did - the volume is just the size multiplied by the
brightness. The square pixel shape is a synthetic representation of the
pixel which avoids the singularities of true delta functions, and is
chosen because it is simple to generate and an array of them produces a
full 100% coverage without gaps or overlaps. Assigning this shape also
assigns it a structure. You could equally have different pixel
structures that would be just as valid - for example the pixels produced
by some systems are hexagonal whilst others are diamond shaped (squares
rotated by 45deg). Most computer display systems are based on square
pixel formats, but this is not universal - digital studio TV formats,
such as CCIR601, use rectangular pixels.
The type of algorithm used for resizing changes the ratio of image
information to pixel structure in the output. If you get it wrong then
you will see pixel structure that should not be present.
>
>Is this your first time on usenet? <g>
>
I am sure that Steve has already trawled the archives as he seems to be
doing with Uni. ;-)
>I think you misconstrue what most of these people have
>said. You seem to regard this PSP flaw as a personal affront, rather
>than a mere bug to be corrected. You seem to think that we must all
>regard this as fatal flaw and conclude that JASC and it's employees
>are corrupt and deceptive.
>
Not at all, but I was appalled when Kris's response to Jean-Luc's
original question was to tell him to try a different scale factor. That
is a technique I have seen implemented so often by dodgy car salesmen
and spinning politicians. It sets alarm bells ringing every time.
>
>It is without question, not the same function outlined in the articles
>you sited. Nor is it the same as the "bilinear resample" implemented
>in PhotoShop 5.0 LE. (and I presume PhotoShop 5,5.5,6,7 and Elements).
>For photographs, PhotoShop's 'bilinear resample' appears to be superior
>to PSP's 'bilinear resample.'
>
There is only one bilinear function and it is mathematically defined.
The fact that PSP's bilinear resample isn't as good as others or even
the mathematical stepwise implementation of integer bilinear scaling
that I explained earlier, is evidence enough of its flaws. Similar
evidence exists of flaws in the bicubic scaling in the case of
downsampling, but I haven't even discussed that here because there seems
little hope of people even recognising the much more obvious deficiency
of the bilinear case.
>
>All software has flaws.
>
Sure is, but even Bill Gates admits that Windows has bugs!
Bob Dietz
> Bob Dietz <rbdiet...@yahoo.com> writes:
>>
>> I read most of the links you provided. I don't recall any of them
>> discussing 'pixel structure.' What do you mean by 'completely rejects
>> pixel structure?'
>>
> In sampling theory the pixel itself is an infinitely fine point, known
> as a delta function and the volume under the delta function is the
> signal. This might seem a difficult concept to grasp, but think of
> making a spot smaller and smaller, then the brightness of the spot needs
> to increase inversely to its area for the spot to remain as visible as
> it previously did - the volume is just the size multiplied by the
> brightness. The square pixel shape is a synthetic representation of the
> pixel which avoids the singularities of true delta functions, and is
> chosen because it is simple to generate and an array of them produces a
> full 100% coverage without gaps or overlaps. Assigning this shape also
> assigns it a structure. You could equally have different pixel
> structures that would be just as valid - for example the pixels produced
> by some systems are hexagonal whilst others are diamond shaped (squares
> rotated by 45deg). Most computer display systems are based on square
> pixel formats, but this is not universal - digital studio TV formats,
> such as CCIR601, use rectangular pixels.
>
> The type of algorithm used for resizing changes the ratio of image
> information to pixel structure in the output. If you get it wrong then
> you will see pixel structure that should not be present.
Thanks for the clarification.
>> Is this your first time on usenet? <g>
>>
> I am sure that Steve has already trawled the archives as he seems to be
> doing with Uni. ;-)
Uni is a "special" case.
>> I think you misconstrue what most of these people have
>> said. You seem to regard this PSP flaw as a personal affront, rather
>> than a mere bug to be corrected. You seem to think that we must all
>> regard this as fatal flaw and conclude that JASC and it's employees
>> are corrupt and deceptive.
>>
> Not at all, but I was appalled when Kris's response to Jean-Luc's
> original question was to tell him to try a different scale factor. That
> is a technique I have seen implemented so often by dodgy car salesmen
> and spinning politicians. It sets alarm bells ringing every time.
Here, the best I can do is agree to disagree with you.
>> It is without question, not the same function outlined in the articles
>> you sited. Nor is it the same as the "bilinear resample" implemented
>> in PhotoShop 5.0 LE. (and I presume PhotoShop 5,5.5,6,7 and Elements).
>> For photographs, PhotoShop's 'bilinear resample' appears to be superior
>> to PSP's 'bilinear resample.'
>>
> There is only one bilinear function and it is mathematically defined.
> The fact that PSP's bilinear resample isn't as good as others or even
> the mathematical stepwise implementation of integer bilinear scaling
> that I explained earlier, is evidence enough of its flaws. Similar
> evidence exists of flaws in the bicubic scaling in the case of
> downsampling, but I haven't even discussed that here because there seems
> little hope of people even recognising the much more obvious deficiency
> of the bilinear case.
>
>>
>> All software has flaws.
>>
> Sure is, but even Bill Gates admits that Windows has bugs!
Last I knew, Microsoft was calling them "issues?" <g>
Bob
Sally Beacham
>
> Since when was Sally Beacham appointed Usenet moderator-in-chief?
Last week. I take it you didn't get the memo. Probably the office staff
has as much fondness for you as the oik in Accounting.
>
> You really crack me up Sally. You object to someone highlighting faults
> or bugs in the application and then you throw around childish names in a
> vain attempt to appear adult. Names which, incidentally, are more
> befitting of someone under a juvenile infatuation with PSP!
Sock puppet? Look it up.
>
> > If you don't think it's good enough FOR
> >YOU, you have been told countless times the appropriate channels to use
to
> >express it.
>
> Are you totally unable to read?
Why, actually, no. I have an IQ of 152, I read 1200 wpm and I managed a
measly 780 on my English SAT's many years ago. Unfortunately, my brilliant
son topped me at 800... but that's good genes. I run a very successful
business, I teach, I volunteer, I manage a household (haphazardly, it's
true) three children, two cats, two husbands (one ex- thankfully,) four
cars, a wide circle of friends, fulfilling community involvement and oh yes,
I'm pursuing another degree. So yes, I think I can read. It's also why
they made me Usenet moderator-in-chief. You just didn't get the memo.
Tsk tsk, must be nicer to those secretaries, you know.
>
> If you are happy with the product then you have nothing to contribute to
> a thread which discusses ONE of its defects - but then you haven't
> really contributed anything to the discussion anyway, just muddied the
> waters with personal venom. Quite a common trait of those who
> contribute nothing to society in general, unfortunately.
You were rude and abusive to Kris. You deserved it. You continue to
deserve it. So sayeth the Usenet moderator-in-chief.
JoAnn
"Sally Beacham" <p...@tension.you> wrote in message
news:ajn11...@enews2.newsguy.com...
So sayeth the Usenet moderator-in-chief.
Congratulations Sally!! :-)))
--
Jo
=============================
Got PSP Questions? Get Answers!
http://campratty.com/questions.html
=============================
Kris Zaklika
>
> Kennedy McEwen wrote:
[snip]
> > Not at all, but I was appalled when Kris's response to Jean-Luc's
> > original question was to tell him to try a different scale factor. That
> > is a technique I have seen implemented so often by dodgy car salesmen
> > and spinning politicians. It sets alarm bells ringing every time.
>
> Here, the best I can do is agree to disagree with you.
Good plan. I've sold precisely 6 used cars in my life, all
mine, and nobody complained. I have never held any office
(not even class president) and nobody has accused me of
political skills. So far I've been labeled as a "dodgy car
salesman", "a spinning politician" and been accused of
ignorance, of not bothering to read posts, of diverting
attention, of asking for trouble, of causing confusion, of
waffling, of contributing to "Jasc lies", of a cover up, of
deception, of being an emperor with no clothes, of deceptive
answers, of an intention to mislead, of countering "with
misleading information and deception before sulking off",
of "chirping in with totally misleading information in an
attempt to divert attention away from the question", of
Jasc claiming "that PSP implements bilinear resizing
directly", and a few other largely repetitive and redundant
things. All this from a person with whom I have exchanged
no words beyond comments about his manners and who is
talking to himself on the topic of bilinear resampling.
The only apology has been "I apologise Kris, I thought you
were discussing UPSCALING", where his rush to find fault with
me resulted in his not actually reading the post. According to
the voluminous postings of this individual he is in a position
to ignore supposed defects in PSP and write exactly the code
he wants. Nonetheless he chooses to kvetch instead. In these
same voluminous postings he took the opportunity to insult
a few other posters too. It is little surprise he has trouble
persuading his coworkers to agree with him.
This little extravaganza gives you all a fascinating insight
into what it is like to participate in a newsgroup as a Jasc
employee. Fortunately, the vast majority of posts and e-mails
I receive are not of this type but this one is representative
of some of the minority I am obliged to put up with in
exchange for continuing to volunteer my time in this
newsgroup.
Just because I volunteer my time here I don't see the need
to take any crap from a monomaniac poster. If I were to
behave like this poster I would assert that he is a person who
lies about his name, which is obviously "McEwan" and not the
misleading, deceptive and criminally misrepresentative "McEwen"
that he falsely intimates in a transparent attempt to divert
attention from his behavior. Of course, I don't behave like
that. Nor do I communicate with people like that. By his
behavior he has made himself an unperson.
[snip]
--
----------------------------------------------------------------------
Kris Zaklika Jasc Software, Inc. The
Product Ideas: id...@jasc.com Power
Customer Service: customer...@jasc.com To
Technical Support: tec...@jasc.com Create
----------------------------------------------------------------------
Mike
Sally Beacham wrote:
>
> You persist in ranting hysterically here, where absolutely
> NOTHING can be done about it.
His choice, and YOU can do nothing about it.
":^) ®
--
Mike
• Logo Design •
Put some fun in your next logo!
Site at: http://www.artistmike.com
Mike
Sally Beacham wrote:
>
> "Kennedy McEwen" <r...@nospam.demon.co.uk> wrote in message
> news:vOX6uodI...@kennedym.demon.co.uk...
> >
> > Since when was Sally Beacham appointed Usenet moderator-in-chief?
>
> Last week.
Not true. So the rest of your reply can be disregarded also.
Uni
You poor baby.
Uni
Sally Beacham
"JoAnn" <jowhi...@verizonmail.com> wrote in message
news:SuE79.42361$dr.61...@e3500-atl1.usenetserver.com...
>
> "Sally Beacham" <p...@tension.you> wrote in message
> news:ajn11...@enews2.newsguy.com...
>
> So sayeth the Usenet moderator-in-chief.
>
>
> Congratulations Sally!! :-)))
Thank you thank you, I really have to thank you all, I couldn't have done it
without your love and loyal support. I promise, I'll remember you someday
when I am Empress of the Internet.
There'll be free broadband for all of you. Oh hell, you were with me when I
was just a pup, let's make it a T-3 line.
*giggle*