Color STN LCD controller

[Anders Hellerup Madsen]

Hi,

I'm trying to design a controller for a Hitachi color STN display, but
I'm running in to all kinds of problems here. I think most of it comes
from the fact that I really don't know anything about how LCD displays
work, so I was hoping someone in this group could enlighten me a little.

From what I've gathered from the displays datasheet, it seems the
display is controlled by two different clock signals, one for vertical
and one for horizontal lines. Once these clock signals are set
correctly, it should just be a matter of clocking the pixel data in right?

There seems to be three bits for each color component, red, blue and
green, and by my calculation this gives 8 different colors. However the
display is listed as beeing able to show up to 65,000 different colors.
How exactly is this possible?

I have thought about various dithering schemes, but i think most of them
implies very dramatical losses in refresh speeds, resolution and image
quality. What is the normal way of producing colors on a CSTN display?

Regards, Anders

 

[Lasse Langwadt Christense]

Anders Hellerup Madsen wrote:

Hi,

I'm trying to design a controller for a Hitachi color STN display, but
I'm running in to all kinds of problems here. I think most of it comes
from the fact that I really don't know anything about how LCD displays
work, so I was hoping someone in this group could enlighten me a little.

From what I've gathered from the displays datasheet, it seems the
display is controlled by two different clock signals, one for vertical
and one for horizontal lines. Once these clock signals are set
correctly, it should just be a matter of clocking the pixel data in right?

kinda, you would normally have something like a clk to clock in the
pixels, a sync for each line and a sync for each frame.

There seems to be three bits for each color component, red, blue and
green, and by my calculation this gives 8 different colors. However the
display is listed as beeing able to show up to 65,000 different colors.
How exactly is this possible?

shouldn't that be, ... 8*8*8 = 512 colors

I have thought about various dithering schemes, but i think most of them
implies very dramatical losses in refresh speeds, resolution and image
quality. What is the normal way of producing colors on a CSTN display?

Regards, Anders

if you stick to some form of toggling between different intensities of
each color to get apparent intensities in between, shouldn't only the
resulting refresh rate go down?

-Lasse

 

[Gerd B.]

"Anders Hellerup Madsen" <anders@hellerup-madsen.dk> schrieb im Newsbeitrag
news:bp29um$b6s$1@news.net.uni-c.dk...

Hi,

I'm trying to design a controller for a Hitachi color STN display, but
I'm running in to all kinds of problems here. I think most of it comes
from the fact that I really don't know anything about how LCD displays
work, so I was hoping someone in this group could enlighten me a little.

From what I've gathered from the displays datasheet, it seems the
display is controlled by two different clock signals, one for vertical
and one for horizontal lines. Once these clock signals are set
correctly, it should just be a matter of clocking the pixel data in right?

There seems to be three bits for each color component, red, blue and
green, and by my calculation this gives 8 different colors. However the
display is listed as beeing able to show up to 65,000 different colors.
How exactly is this possible?

I have thought about various dithering schemes, but i think most of them
implies very dramatical losses in refresh speeds, resolution and image
quality. What is the normal way of producing colors on a CSTN display?

Regards, Anders

In the last weeks I've designed a TFT Controller für Sharp TFT Panels.
The Panel ist 640x480 Pixel, 262144 Colors. EBay: 40$
The TFT Controller has an Video + SVHS Input. FPGA: Altera Cyclone EP1C3 (
about 20$).
The "source-Code" is an Altera Quartus "graphical block diagram." Size:
about 1 printed page.

You see: Its very easy to design a controller for TFT displays.

Gerd

 

[Anders Hellerup Madsen]

Lasse Langwadt Christensen wrote:

Anders Hellerup Madsen wrote:
From what I've gathered from the displays datasheet, it seems the
display is controlled by two different clock signals, one for vertical
and one for horizontal lines. Once these clock signals are set
correctly, it should just be a matter of clocking the pixel data in
right?

kinda, you would normally have something like a clk to clock in the
pixels, a sync for each line and a sync for each frame.

Ok, this is like I thought then.

There seems to be three bits for each color component, red, blue and
green, and by my calculation this gives 8 different colors. However
the display is listed as beeing able to show up to 65,000 different
colors. How exactly is this possible?

shouldn't that be, ... 8*8*8 = 512 colors

I'm sorry, I was unclear at this point. There is three bits per pixel,
one for each color component, so it is only eight colors.

I have thought about various dithering schemes, but i think most of
them implies very dramatical losses in refresh speeds, resolution and
image quality. What is the normal way of producing colors on a CSTN
display?

if you stick to some form of toggling between different intensities of
each color to get apparent intensities in between, shouldn't only the
resulting refresh rate go down?

Yep, and apperently this is the way things are expected to be done. I
was also considering a kind of dithering where I would use, fx. four
adjectant pixels of different colors, to give the illusion that the four
pixels together were a mixed shade of another color. But this would
halve the resolution and generally look very ugly.

But still I am very puzzled about this display. It's a Hitachi SX16H003
display which can display up to 65000 different colors according to this
site:

http://www.hitachi-displays.com/catalog/skg-225-e.html

However according to the datasheet it seems there is only 3 bits per
pixels. I am still puzzled. The datasheet can be found here:

http://www.hitachi-displays-eu.com/pdf/sx16h003.pdf

Thank you for your answer,

Regards Anders

 

[Iwo Mergler]

Anders Hellerup Madsen wrote:

Yep, and apperently this is the way things are expected to be done. I
was also considering a kind of dithering where I would use, fx. four
adjectant pixels of different colors, to give the illusion that the four
pixels together were a mixed shade of another color. But this would
halve the resolution and generally look very ugly.

But still I am very puzzled about this display. It's a Hitachi SX16H003
display which can display up to 65000 different colors according to this
site:

http://www.hitachi-displays.com/catalog/skg-225-e.html

However according to the datasheet it seems there is only 3 bits per
pixels. I am still puzzled. The datasheet can be found here:

http://www.hitachi-displays-eu.com/pdf/sx16h003.pdf

Remember that you have time as a third dimension for dithering.
STN displays are slow to respond, so dithering in time, between
frames, can give you more color resolution.

Comercial STN display chips use a combination between spacial and
temporal dither to achieve the claimed 65000 colors. Look up
how floyd-steinberg dithering is done and extend it to three
dimensions. It's a starting point.

Kind regards,

Iwo