Picture coordinate readout when worksheet is zoome

[PhilipOakley]

If a picture has been set up to return a pixel coordinate, then when the whole worksheet is zoomed, the coordinate values are zoomed, even when the same location is clicked.

see attached. Requires Lena from the Wavelets pack.

Problem exists in both V11.2a and V14 M030

Philip Oakley

[ptc-1368288]

On 12/17/2009 12:02:35 PM, philipoakley wrote:
>If a picture has been set up
>to return a pixel coordinate,
>then when the whole worksheet
>is zoomed, the coordinate
>values are zoomed, even when
>the same location is clicked.
>
>see attached. Requires Lena
>from the Wavelets pack.
>
>Problem exists in both V11.2a
>and V14 M030
>
>Philip Oakley
_______________________________

YES, "the coordinate
>values are zoomed, even when
>the same location is clicked.

... and you get nothing usable out of a zoomed image, no matter the format whereas "ZOOM" is a destructive image processing function. What was your question really ?

jmG

[PhilipOakley]

On 12/17/2009 2:31:35 PM, jmG wrote:
>On 12/17/2009 12:02:35 PM, philipoakley
>wrote:
>>If a picture has been set up
>>to return a pixel coordinate,
>>then when the whole worksheet
>>is zoomed, the coordinate
>>values are zoomed, even when
>>the same location is clicked.
>>
>>see attached. Requires Lena
>>from the Wavelets pack.
>>
>>Problem exists in both V11.2a
>>and V14 M030
>>
>>Philip Oakley
>_______________________________
>
>YES, "the coordinate
>>values are zoomed, even when
>>the same location is clicked.
>
>... and you get nothing usable out of a
>zoomed image, no matter the format
>whereas "ZOOM" is a destructive image
>processing function. What was your
>question really ?
>
>jmG
>
>
>
Jean,
It is a statement about the fault.
Mona requested I post it here.
The worksheet zoom should NOT affect the coordinate of a given image pixel! It is still at the same place in the matrix.

Philip Oakley

[ptc-1368288]

On 12/18/2009 4:31:28 AM, philipoakley wrote:
>On 12/17/2009 2:31:35 PM, jmG wrote:
>>On 12/17/2009 12:02:35 PM, philipoakley
>>wrote:
>>>If a picture has been set up
>>>to return a pixel coordinate,
>>>then when the whole worksheet
>>>is zoomed, the coordinate
>>>values are zoomed, even when
>>>the same location is clicked.
>>>
>>>see attached. Requires Lena
>>>from the Wavelets pack.
>>>
>>>Problem exists in both V11.2a
>>>and V14 M030
>>>
>>>Philip Oakley
>>_______________________________
>>
>>YES, "the coordinate
>>>values are zoomed, even when
>>>the same location is clicked.
>>
>>... and you get nothing usable out of a
>>zoomed image, no matter the format
>>whereas "ZOOM" is a destructive image
>>processing function. What was your
>>question really ?
>>
>>jmG
>>
>>
>>
>Jean,
>It is a statement about the fault.
>Mona requested I post it here.
>The worksheet zoom should NOT affect the
>coordinate of a given image pixel! It is
>still at the same place in the matrix.
>
>Philip Oakley
________________________

Yes, it must affect whereas past some pixels the all pixels are shifted and most of them are not true anymore. Don't you remember the all work sheets about "pixelation". The answer is there in those work sheets. Try to convince yourself by putting a red dot in Lena center picture, look at that coordinate in the picture tool, then zoom 2 x 2 and look again at the new coordinates. Zoom does not make pixels bigger or smaller, zoom is a function, that adds or remove data values in the image matrix. Therefore the grid system is affected. Zooming a picture is only applicable to technical matters of no concern to the person either directly or indirectly. Zooming a person is an act of forgery.
For what Mona is asking about your interpretation, hard to tell if there is fault or not but you can check easily by putting a NaN and looking back where it locates in the new grid . That will be easier than the red dot in Lena

[PhilipOakley]

On 12/19/2009 3:54:04 PM, jmG wrote:
>On 12/18/2009 4:31:28 AM, philipoakley
>wrote:
>>On 12/17/2009 2:31:35 PM, jmG wrote:
>>>On 12/17/2009 12:02:35 PM, philipoakley
>>>wrote:
>>>>If a picture has been set up
>>>>to return a pixel coordinate,
>>>>then when the whole worksheet
>>>>is zoomed, the coordinate
>>>>values are zoomed, even when
>>>>the same location is clicked.
>>>>
>>>>see attached. Requires Lena
>>>>from the Wavelets pack.
>>>>
>>>>Problem exists in both V11.2a
>>>>and V14 M030
>>>>
>>>>Philip Oakley
>>>_______________________________
>>>
>>>YES, "the coordinate
>>>>values are zoomed, even when
>>>>the same location is clicked.
>>>
>>>... and you get nothing usable out of a
>>>zoomed image, no matter the format
>>>whereas "ZOOM" is a destructive image
>>>processing function. What was your
>>>question really ?
>>>
>>>jmG
>>>
>>>
>>>
>>Jean,
>>It is a statement about the fault.
>>Mona requested I post it here.
>>The worksheet zoom should NOT affect the
>>coordinate of a given image pixel! It is
>>still at the same place in the matrix.
>>
>>Philip Oakley
>________________________
>
>Yes, it must affect whereas past some
>pixels the all pixels are shifted and
>most of them are not true anymore. Don't
>you remember the all work sheets about
>"pixelation". The answer is there in
>those work sheets. Try to convince
>yourself by putting a red dot in Lena
>center picture, look at that coordinate
>in the picture tool, then zoom 2 x 2 and
>look again at the new coordinates. Zoom
>does not make pixels bigger or smaller,
>zoom is a function, that adds or remove
>data values in the image matrix.
>Therefore the grid system is affected.
>Zooming a picture is only applicable to
>technical matters of no concern to the
>person either directly or indirectly.
>Zooming a person is an act of forgery.
>For what Mona is asking about your
>interpretation, hard to tell if there is
>fault or not but you can check easily by
>putting a NaN and looking back where it
>locates in the new grid . That will be
>easier than the red dot in Lena
>
>
>
>
>

Jean,

The reported value is incorrect. Enough said.

Philip Oakley

[ptc-1368288]

Jean,
The reported value is incorrect. Enough said.
Philip Oakley
_______________________

I will look at it next year.
I doubt your conclusion is that simple, enough said.
Jean

[IRstuff]

I don't quite see that behavior in 11.2a. What I get is motion of the cursor, regardless of whether the iamge is selected, or not.

Also, at higher sheet zoom, the cursor's reported position does not correspond to cursor's actual position, so Lena's left eye is reported aas 246,198, while its 100% zoom location is 165,133. However, at 150% zoom, I can't select a cursor position larger than 255,255, which is the original image's size, but that's only about 2/3 of the image, and the remainer is unselectable.

TTFN,
Eden

[ptc-1368288]

On 12/24/2009 8:26:23 AM, sgrode wrote:
>zooming has some many bugs in
>Mathcad that is should have been
>reprogrammed years ago.
>
>Steen Gro�e_
________________________________

As there is no reference code because zoom is a forging function, then you will have hard time to prove the "Mathcad unknown zoom function" could be buggy !!! You are at liberty to exhibit more of your declaration. It seems reasonable to think that the Mathcad zoom is the inverse of the downsampling. Therefore an image, for instance an image of an object that would have been downsampled for reducing the size of a catalogue, that image could be restored [restored !!! ???] in the best reconstruct. But zooming (+) an original image is destructing that image, again: an act of forgery.

The Mathcad Improc downsampling is available for inspection, but not the zoom (+). No source code was found.

jmG

[ptc-1368288]

On 12/17/2009 12:02:35 PM, philipoakley wrote:
>If a picture has been set up
>to return a pixel coordinate,
>then when the whole worksheet
>is zoomed, the coordinate
>values are zoomed, even when
>the same location is clicked.
>
>see attached. Requires Lena
>from the Wavelets pack.
>
>Problem exists in both V11.2a
>and V14 M030
>
>Philip Oakley
______________________________

You have explained nothing about what you attempted. Zoom (�) is pixelation, i.e: a destructive process, it adds/removes pixels from the original matrix. If 127 ... 127 is the center of original Lena, when you zoom the work sheet 75% the Lena matrix has no more 255 ... 255 pixels but 191 ... 191, therefore Lena center is no more @ 127 ... 127.
All like explained before, no defect in Mathcad picture tool 11.2a !

Enough said (as you say).

jmG

[PhilipOakley]

On 1/10/2010 9:28:00 PM, jmG wrote:
>On 12/17/2009 12:02:35 PM, philipoakley
>wrote:
>>If a picture has been set up
>>to return a pixel coordinate,
>>then when the whole worksheet
>>is zoomed, the coordinate
>>values are zoomed, even when
>>the same location is clicked.
>>
>>see attached. Requires Lena
>>from the Wavelets pack.
>>
>>Problem exists in both V11.2a
>>and V14 M030
>>
>>Philip Oakley
>______________________________
>
>You have explained nothing about what
>you attempted. Zoom (�) is pixelation,
>i.e: a destructive process, it
>adds/removes pixels from the original
>matrix. If 127 ... 127 is the center of
>original Lena, when you zoom the work
>sheet 75% the Lena matrix has no more
>255 ... 255 pixels but 191 ... 191,
>therefore Lena center is no more @ 127
>... 127.
>All like explained before, no defect in
>Mathcad picture tool 11.2a !
>
>Enough said (as you say).
>
>jmG
>
>
>

jean,
I think you may have misunderstood the type of zoom being considered. This is the tools on the pallette for visualisation. This provides both pan and zoom so that one can inspect local image regions down to the pixel level, thus the pixel reference values should be unchanged.

In fact as I decribe this it prompts me to look at the effect of 'panning' such a zoomed image, and checking how the pixel displayed at top left is numbered as the image is panned.

From a user perspective, I expect the coordinate to be that of the image bing displayed, not the display coordinate within the window.

Have a go with the file.

Philip Oakley

[PhilipOakley]

On 1/11/2010 4:47:20 AM, philipoakley wrote:
>On 1/10/2010 9:28:00 PM, jmG wrote:
>>On 12/17/2009 12:02:35 PM, philipoakley
>>wrote:
>>>If a picture has been set up
>>>to return a pixel coordinate,
>>>then when the whole worksheet
>>>is zoomed, the coordinate
>>>values are zoomed, even when
>>>the same location is clicked.
>>>
>>>see attached. Requires Lena
>>>from the Wavelets pack.
>>>
>>>Problem exists in both V11.2a
>>>and V14 M030
>>>
>>>Philip Oakley
>>______________________________
>>
>>You have explained nothing about what
>>you attempted. Zoom (�) is pixelation,
>>i.e: a destructive process, it
>>adds/removes pixels from the original
>>matrix. If 127 ... 127 is the center of
>>original Lena, when you zoom the work
>>sheet 75% the Lena matrix has no more
>>255 ... 255 pixels but 191 ... 191,
>>therefore Lena center is no more @ 127
>>... 127.
>>All like explained before, no defect in
>>Mathcad picture tool 11.2a !
>>
>>Enough said (as you say).
>>
>>jmG
>>
>>
>>
>
>jean,
>I think you may have misunderstood the
>type of zoom being considered. This is
>the tools on the pallette for
>visualisation. This provides both pan
>and zoom so that one can inspect local
>image regions down to the pixel level,
>thus the pixel reference values should
>be unchanged.
>
>In fact as I decribe this it prompts me
>to look at the effect of 'panning' such
>a zoomed image, and checking how the
>pixel displayed at top left is numbered
>as the image is panned.
>
>From a user perspective, I expect the
>coordinate to be that of the image bing
>displayed, not the display coordinate
>within the window.
>
>Have a go with the file.
>
>Philip Oakley

OOPS. Got that one wrong (the quoted comment above). It is the whole screen magnification that was the problem. However the the last paragraph still stands.

The zoom in this case is no different to a disability access control for those with poorer eyesight.

Philip Oakley

[ptc-1368288]

When you zoom the work sheet, it zooms the image too and Lena is no more Lena because of the pixelation that adds/removes data values. Consequently, any original X|Y original coordinates are now "relative coordinates" and pixel levels have changed, even if your eye does not see.
What was the problem ?

jmG

[rmix22-disabled]

On 1/11/2010 11:30:01 PM, jmG wrote:
>When you zoom the work sheet,
>it zooms the image too and
>Lena is no more Lena because
>of the pixelation that
>adds/removes data values.
>Consequently, any original X|Y
>original coordinates are now
>"relative coordinates" and
>pixel levels have changed,
>even if your eye does not see.
>What was the problem ?

I think we suppose that even after zooming, the coordinates of, lets say the left eye, should not change. It works correct when you use the zoom from the image palette, zooming the image only and this dispite of the effect of pixelation and "added data values".
The faulty behaviour occurs when You are zooming the whole worksheet (for better visibility e.g. when using a beamer for demonstration).

Regards
RMix

[ptc-1368288]

"The faulty behaviour occurs when You are zooming the whole worksheet ..."
________________________

That's what I have said from the very beginning
i.e: zoom work sheet has destroyed the image.
The destruction occurs between the PC & the screen.

jmG

[rmix22-disabled]

Same effect here. Also with 14M030.

If you zoom (not from the image palette), you are not able to reach the bottom right corner with the crosshair cursor. It stays at the position where the bottom right corner used to be before zooming. MC assumes this is (255;255) in case of lena.

Because the zooming from the image-palette works correct its not simply giving back screen coordinates. I suspect that the routine which translates screen/mouse-coordinates to image coordinates does not consider the worksheet-zoom.

Playing around I just experienced another behaviour which is a little confusing me. When I mirror the image along a vertical axis crosshair cursor and mousecursor are mirrored too. If I mouseclick at the lower left corner the crosshaircursor appears at the upper left!
Didn't work a lot with images in mathcad but I dont't think this is as it should be.

Regards
RMix

[mzeftel]

This has been logged.

Mona