Bug 35160

Summary: [RADEON:KMS:RV200:M7:DDX] inferior output using external VGA with radeon mobility m7 laptop
Product: xorg Reporter: Clint Adams <clint>
Component: Driver/RadeonAssignee: xf86-video-ati maintainers <xorg-driver-ati>
Status: RESOLVED INVALID QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium CC: kibi
Version: unspecified   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
uname output
none
dmesg output
none
Xorg log
none
photo of screen
none
photo of screen
none
photo of screen
none
photo of screen
none
photo of screen
none
photo of screen
none
xrandr --verbose none

Description Clint Adams 2011-03-09 13:48:15 UTC
On various laptops with radeon cards, the LVDS display is fine, but the font rendering varies from blurry to spastic on an external display, and sometimes the color balance is affected.

One card is

01:00.0 VGA compatible controller: ATI Technologies Inc Radeon Mobility M7 LW [Radeon Mobility 7500]

and the Debian revision of xserver-xorg-video-radeon is 1:6.13.1-2+squeeze1.

No large distortions in the style of #22175 are experienced.

This happens with an empty/non-existent config.  The effects are not entirely consistent, but some form of the problem consistently occurs when windows are placed on the external monitor.
Comment 1 Clint Adams 2011-03-09 14:12:02 UTC
Created attachment 44284 [details]
uname output
Comment 2 Clint Adams 2011-03-09 14:12:32 UTC
Created attachment 44285 [details]
dmesg output
Comment 3 Clint Adams 2011-03-09 14:13:57 UTC
Created attachment 44286 [details]
Xorg log
Comment 4 Michel Dänzer 2011-03-10 00:40:38 UTC
(In reply to comment #4)
> On various laptops with radeon cards, the LVDS display is fine, but the font
> rendering varies from blurry to spastic on an external display, and sometimes
> the color balance is affected.

Can you attach photos highlighting the problem?

Is the external display an LCD as well, or a good old fashioned CRT?
Comment 5 Michel Dänzer 2011-03-10 00:42:01 UTC
Oh, and can you attach the output of xrandr --verbose?
Comment 6 Jerome Glisse 2011-03-10 09:37:31 UTC
Also please list GPU you tested with and kind of video output (VGA,DVI,HDMI,DISPLAY PORT)
Comment 7 Clint Adams 2011-03-10 12:10:27 UTC
Created attachment 44325 [details]
photo of screen
Comment 8 Clint Adams 2011-03-10 12:10:54 UTC
Created attachment 44326 [details]
photo of screen
Comment 9 Clint Adams 2011-03-10 12:11:19 UTC
Created attachment 44327 [details]
photo of screen
Comment 10 Clint Adams 2011-03-10 12:11:49 UTC
Created attachment 44328 [details]
photo of screen
Comment 11 Clint Adams 2011-03-10 12:12:21 UTC
Created attachment 44329 [details]
photo of screen
Comment 12 Clint Adams 2011-03-10 12:12:54 UTC
Created attachment 44330 [details]
photo of screen
Comment 13 Clint Adams 2011-03-10 12:29:31 UTC
Created attachment 44331 [details]
xrandr --verbose
Comment 14 Clint Adams 2011-03-10 12:31:06 UTC
Attached 3 pairs of photos comparing the good LVDS (LCD) screen versus the external LCD monitor connected to the VGA port on the laptop, as well as xrandr --verbose.  I don't understand the GPU question.
Comment 15 Clint Adams 2011-03-10 12:36:29 UTC
In addition to the card I listed in the first submission, the problem also occurs with a laptop containing

ATI Technologies Inc Radeon RV250 [Mobility FireGL 9000]
Comment 16 Clint Adams 2011-03-10 12:39:48 UTC
This problem occurs with a number of laptop and external monitor combinations (though all are LCD).  It does not occur when using the driver from lenny with the same kernel, hardware, and monitor.
Comment 17 Michel Dänzer 2011-03-11 03:11:46 UTC
It's kinda hard to tell from the photos, but I think the text might be using sub-pixel anti-aliasing. Does disabling that help?
Comment 18 Clint Adams 2011-03-11 08:04:49 UTC
(In reply to comment #17)
> It's kinda hard to tell from the photos, but I think the text might be using
> sub-pixel anti-aliasing. Does disabling that help?

Unless it's the default, it is not enabled except for the Korean fonts in ttf-unfonts-core and ttf-unfonts-extra.  Removing this does change the font rendering on the LVDS, but appears to have no effect on the VGA.

If it's the default, then I probably don't understand fontconfig well enough to ensure that it's disabled.
Comment 19 Michel Dänzer 2011-03-11 08:38:32 UTC
You can tell by inspecting the aliased pixels of black text on white background. If any of them aren't gray (i.e. not all of their colour components are equal), that's sub-pixel anti-aliasing.
Comment 20 Clint Adams 2011-03-11 10:33:42 UTC
(In reply to comment #19)
> You can tell by inspecting the aliased pixels of black text on white
> background. If any of them aren't gray (i.e. not all of their colour components
> are equal), that's sub-pixel anti-aliasing.

There are varying shades of pixels along the edges of the letters.  If I disable "Smoothing" and "Hinting" in gnome-appearance-properties, it becomes strictly black and white, and looks much worse on both screens.
Comment 21 Michel Dänzer 2011-03-12 10:27:35 UTC
(In reply to comment #20)
> If I disable "Smoothing" and "Hinting" in gnome-appearance-properties, it
> becomes strictly black and white, and looks much worse on both screens.

Disabling 'smoothing' disables anti-aliasing. If it's set to 'grayscale' as opposed to 'Subpixel (LCDs)', sub-pixel anti-aliasing should be disabled. Or you could just compare the example text in the dialogue to see the effect of each setting.

When sub-pixel anti-aliasing is enabled, you may also want to play with the subpixel orders at the bottom of that dialogue.
Comment 22 Clint Adams 2011-03-12 10:46:44 UTC
> Disabling 'smoothing' disables anti-aliasing. If it's set to 'grayscale' as
> opposed to 'Subpixel (LCDs)', sub-pixel anti-aliasing should be disabled. Or
> you could just compare the example text in the dialogue to see the effect of
> each setting.

In that case it was not enabled; I believe the setting was called "Best shape" or
something like that.
 
> When sub-pixel anti-aliasing is enabled, you may also want to play with the
> subpixel orders at the bottom of that dialogue.

We'll give that a try on Monday.  Is there anything else that will help isolate the bug?
Comment 23 Clint Adams 2011-03-16 10:38:38 UTC
> When sub-pixel anti-aliasing is enabled, you may also want to play with the
> subpixel orders at the bottom of that dialogue.

I cannot detect any variation between these.
Comment 24 Adam Jackson 2018-06-12 19:06:46 UTC
Mass closure: This bug has been untouched for more than six years, and is not
obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.