0.53 Win: Display errors still present

[JohnFredC]

The display anomalies discussed in this post are still present in 0.53 on my XP system.

The lack of resolution of this issue over the last couple of years is frustrating. All image types show the lines. Different zoom levels show different patterns of lines or none at all.

For instance, in an informal test on a small 636x816x24 jpg displayed in an MP 0.53 image tab, the rectilinear line patterns appear at the following zoom levels:

150%
240%
250%
480%
560%

The patterns of lines differ in each case. All other zoom levels are clean.

For a larger, 3456x4608x24 jpg the lines appear at the following zoom levels:

30%
60%
70%
90%
110%
120%
130%
140%
150%
170%
180%

Again, the line patterns differ and the zoom levels not listed are clean (I quit at 180).

Very strange.

Is it a problem with the display library code? The rectilinear nature of the lines suggest it is related to how MP tiles zoomed displays. Since the lines are 1 pixel wide, perhaps it is an simple arithmetic problem associated with sizing and positioning the image tiles. A rounding error, perhaps.

Is it a problem with certain Windows environments? I have XP SP3.

What about display cards? I only have 128mb of display memory on my old ATI graphics card. Could that be the issue?

Does anyone in the forum NOT have this issue? Info about systems where the lines do NOT appear might be helpful.

Please, this needs to be fixed.

[Vapavite123]

Hello,

I'm on XP SP2
2 graphics cards with identical characteristics to yours (with 128 MB ATI and ATI 64MB)

I had the same problem as you and I never spoke about this problem because I did not know how to reproduce it consistently. In addition, for some time I am not affected by this problem and there may be several reasons for this.
I rarely zoom!
'' If I use it, I do CTRL + wheel or a selection in the image!
I do not have the same percentages of zoom you (My screens: 1024X768), but I guess it was already like that before!
I had a problem with a monitor that has failed (dual screen). I replaced an older and do I have reversed their position and thus their dependence with the cards!


Good luck and good for you

Bonjour,

Je suis sur XP SP2
2 cartes graphiques ayant des caractéristiques identiques à la vôtre (ATI avec 128 Mo et ATI avec 64 Mo)

J’ai eu le même problème que vous et je n'étais jamais intervenu sur ce problème car je ne voyais pas comment le reproduire de manière systématique. De plus, depuis un certain temps je ne suis plus affecté par ce problème et il peut y avoir plusieurs raisons à cela.
Je zoom rarement !
Si je l''utilise, je fais CTRL + roulette ou une sélection dans l'image !
Je n'ai pas les mêmes pourcentages de zoom que vous (Mes écrans: 1024X768), mais je suppose que c'était déjà comme cela avant !
J'ai eu un problème avec un moniteur qui est tombé en panne (double écran). Je l'ai remplacé par un plus vieux et pour le faire, j'ai inversé leur position et donc leur dépendance avec les cartes !


Bonne continuation et bien à vous

[oops66]

Hello,
FYI: I also sometimes have this bug under XnViewMP 0.51 Linux, and only into the "fullscreen mode" (right now). So it probably depends of the software screen resolution & the W&H image size/ratio .

Same kind of behavior as here - screenshot (...lines in background color appear over the photo giving a tiling effect... a line of one pixel (witdh)) :

or here
http://newsgroup.xnview.com/viewtopic.p ... 90#p111190

Examples:
(With a LCD Screen :Native resolution, 1280x1024 ... Samsung LCD SyncMaster 913N)

1-) software screen resolution def. = 1080x1024 (5:4)
Here, for example in full screen mode only: ... At zoom level 100%, I have this bug with image.jpg with (W=906 H=1022). and if I resize any image with (W=906 H=1022) I also have this bug. ==> BUG (but if I zoom out & zoom in, into the fullscreen mode(a "refresh") again to have 100% too, it's OK - then no more tiling effect, so no bug )

With an other image.jpg (W=3036 H=1225)... At zoom = 37% => NO BUG


2-) software screen resolution def. = 1152x864 (4:3)
Here, for example in full screen mode only: with an image.jpg (the same as before) (W=906 H=1022)... At zoom level 100% ==> NO MORE BUG

But with an image.jpg (the same as before) (W=3036 H=1225) ... At zoom = 37% ==> BUG

... So it probably depends of the software screen resolution & the W&H image size/ratio ? .

- So maybe, in a first step, a workaround is to refresh first to display the image in fullscreen mode ?
It will also be very interesting to have a "refresh screen" function into the View menu with a new shortcut (or better, with the same F5 shortcut too) for the fullscreen mode and the view mode too, because into the fullscreen mode (or View mode) "refresh (F5)-browser mode" does not work here.

Do you have this bug only into the full screen mode ?

[oops66]

FYI:
Here my graphic card:

Code: Select all

$ lspci | grep VGA
01:00.0 VGA compatible controller: Advanced Micro Devices [AMD] nee ATI Caicos [Radeon HD 7400 Series]

It's an ATI too ...

Code: Select all

$ xrandr
Screen 0: minimum 320 x 200, current 1280 x 1024, maximum 16384 x 16384
HDMI-0 disconnected (normal left inverted right x axis y axis)
DVI-0 connected 1280x1024+0+0 (normal left inverted right x axis y axis) 376mm x 301mm
   1280x1024      60.0 +   75.0* 
   1152x864       75.0  
   1024x768       75.1     70.1     60.0  
   832x624        74.6  
   800x600        72.2     75.0     60.3     56.2  
   640x480        72.8     75.0     66.7     60.0  
   720x400        70.1  

Maybe, I have a way to reproduce this bug (into the full screen mode):

For example:
With a software screen resolution def. = 1080x1024 (5:4)

If I resize the image to: (1078x1078) or (1022x1022) , ratio=1, I have the bug (with 2 lines of 1 pixel - a cross)

So to try to reproduce, for exemple:
resize a picture to = Your W. software screen resolution - 2 pixels , so 1080-2, so 1078x1078 (full screen = zoom 94%)
or
resize a picture to = Your H. software screen resolution - 2 pixels , so 1024-2, so 1022x1022 (full screen = zoom 100%)

-------------------------
... In fact here, after more tests, I also have this bug at software screen resolution def. = 1080x1024 (5:4), for all of these "xrandr resolution -2 pixels" execpted for "*-ok.jpg files" (ratio=1, square)
1022.jpg 1150.jpg 1278.jpg 638-ok.jpg 718-ok.jpg 798.jpg 830.jpg 862.jpg
1022.jpg 398.jpg 478.jpg 598.jpg 766.jpg 862.jpg 622-ok.jpg

Does anyone have this bug with an non-ATI graphic card ?

[oops66]

... Seems to be fixed here with the XnViewMP 0.56 Linux X64 version
http://newsgroup.xnview.com/viewtopic.p ... 07#p104807
Thank you Pierre.

[Matthew Wai]

Does the above bug have to do with this one spotted by me?

[oops66]

Does the above bug have to do with this one spotted by me?

Hello,
I don't think so, ... but I think you have to make a new test to see. (was a round error ... so ???)