XO BW mode (was Re: Guidance sought on collaboration techniques)

[Bert Freudenberg]

On 14.02.2009, at 17:24, Gary C Martin wrote:

> On 14 Feb 2009, at 10:16, Bert Freudenberg wrote:
>
>>
>> On 13.02.2009, at 19:27, Gary C Martin wrote:
>>>
>>> Note the backlight goes through the colour refractive screen magic,
>>> even in BW mode, so it's not as sharp as with the backlight 100% off
>>
>> That explanation is horrendously wrong. Please do not perpetuate  
>> the rumor of "modes" in the display hardware. There are none.
>
> Horrendously wrong? Wow... well Bert you've been with the project  
> for way longer than me so I do trust you to be right, I must have  
> hit a nerve, sorry.
>
>> And, there is no "refractive screen magic", that idea did not make  
>> it into production. The backlight uses regular r/g/b color filters.
>
> Was my 'refractive' faux pas your only reason for using  
> 'horrendously wrong'? I guess the term I should have given there was  
> 'transflective', apologies to optical geeks everywhere. I couldn't  
> see any obvious errors else where.

You attributed the decrease in sharpness to some "colour refractive  
screen magic". There is no magic and it got nothing to do with the  
screen at all.

>> The blur you see in color mode is the DCON performing anti-aliasing  
>> by adding color components from the four neighboring pixels into  
>> each pixel (e.g., for a red pixel it takes the red components of  
>> the surrounding pixels into account, etc).
>>
>> In early DCON drivers it was possible to toggle the anti-aliasing  
>> independently of the backlight intensity, so you could clearly see  
>> its effect. IIUC this capability has now been coupled to the  
>> backlight intensity, which is fine in regular use, but takes away  
>> the possibility to easily demonstrate what's happening.
>
> You do appear to be wrong here.

I'd be glad to be wrong, but that's not the flag I meant. That one did  
not disable color, but only the anti-aliasing. That is, you could see  
colors, but it was crisp (leading to colored fringes so that's why the  
setting is not available independently anymore). The fact that you  
can't change the AA independently of the color-to-grayscale conversion  
makes it appear as if the two were intrinsically coupled.

> Have access to three XOs here (two MP and one B4) that all happily  
> respond to the below command:
>
>      su
>      echo 1 > /sys/devices/platform/dcon/output
>
> Disables colour while leaving the backlight setting alone, allowing  
> a slightly crisper BW display with backlight on if you want it  
> (close look at text edges reminds me of the little colour pixels you  
> get with sub-pixel resolution tricks on conventional screens), and:
>
>      su
>      echo 0 > /sys/devices/platform/dcon/output
>
> Enables colour while leaving the backlight setting alone, allowing a  
> 'colour display' mode when the backlight is potentially off (not a  
> very useful combination as it's the backlight that goes through the  
> colour layer so you're left looking at a slightly soft BW image from  
> just reflected light).

If the backlight is on and there is not much ambient light in the room  
you see *no* reflected light, simply because there is no light to be  
reflected. If you are in the sun there is so much reflected light you  
see no colored background light anymore.

> I can try and take some pixel macro photos of the effect if it makes  
> you feel better? :-)
>
> 	http://wiki.laptop.org/go/Display#DCON_screen_driver_chip


Thanks, I do feel fine ;)

I am however slightly frustrated (and I'm not even one of the OLPC  
engineers) about all the misinformation spread about the XO. Didn't  
mean to be offensive though, please chalk this up to English being not  
my mother tongue.

- Bert -