>On Jun 14, 11:17 am, "cutemonster" <ckh...@hotmail.com> wrote:
>
>> Yes, I tried synchronous and asynchronously to Unblank signal.  There
are
>> two problems.  The first one is the noise itself with the stroke
signal
>> and the second one is the stroke signal being drawn at the second time
>> doesn't being drawn at the exact position.  It's like one pixel
toggling
>> between frames.
>>
>> Do you know which kind of filter fit the best in this situation?
>>
>> thanks for you suggestion- Hide quoted text -
>>
>> - Show quoted text -
>
>Sounds like you capture it as live video or multiple frames.  In such
>case you would need to synchronize or lock the clock to a reference,
>the unblank may be?  Having no idea what the unblank looks like so I
>can't tell how to do it, how often does the unblank recycling?
>
>Having a good synchronized clock with low jitter is, IMHO, much easier
>than doing the filter to minimize the artifact of pixels toggling
>positions between frames
>
>For the spatial noise, If I was you I would try a high order analog
>filter infront of the A2D first, before doing any digital filtering
>
>Regards,
>
>
>
>

thanks for the suggestion. For analog filtering, all I can do now is
putting a circuit in breadboard. The problem is that it's a couple inches
away from ADC. For digital filtering,  are they all very similar?  I mean,
there are many type of digital filtering and I'm confused which one I
should use.  Are they specifically work for certain application?

thanks,

On Jun 14, 11:17 am, "cutemonster" <ckh...@hotmail.com> wrote:

> Yes, I tried synchronous and asynchronously to Unblank signal.  There are
> two problems.  The first one is the noise itself with the stroke signal
> and the second one is the stroke signal being drawn at the second time
> doesn't being drawn at the exact position.  It's like one pixel toggling
> between frames.
>
> Do you know which kind of filter fit the best in this situation?
>
> thanks for you suggestion- Hide quoted text -
>
> - Show quoted text -

Sounds like you capture it as live video or multiple frames.  In such
case you would need to synchronize or lock the clock to a reference,
the unblank may be?  Having no idea what the unblank looks like so I
can't tell how to do it, how often does the unblank recycling?

Having a good synchronized clock with low jitter is, IMHO, much easier
than doing the filter to minimize the artifact of pixels toggling
positions between frames

For the spatial noise, If I was you I would try a high order analog
filter infront of the A2D first, before doing any digital filtering

Regards,

>On Jun 13, 3:09 pm, "cutemonster" <ckh...@hotmail.com> wrote:
>> >On Jun 10, 2:38 pm, "cutemonster" <ckh...@hotmail.com> wrote:
>> >> >So you want to display a stroke signal on monitor?
>> >> >Why you need to sample the X? Is it time dimension? Is it a
constant
>> >> >ramping, or ramping with retrace, or random?
>>
>> >> I have to sample x and y because it doesn't work like raster
signal.
>> It's
>> >> voltage varies in time. There is another signal input called
>> Unblank(TTL).
>> >> It turn on and off of XY signal.  
>>
>> >Have you tried to lock the sampling clocks to the unblank?
>>
>> No, I don't understand how to lock it with sampling clock.  Can you
please
>> explain?
>>
>> thanks
>
>Locking the clock to the unblank meaning that their phase relation is
>fixed.  For some reason I can't escape from the time dimension :)  If
>you just sample for 1 frame then probably you don't need to do
>locking.  Do you sample just 1 image then display it continuously on
>the raster?
>
>If you capture it as live video or multiple frames, then how can you
>tell which pixels belong to frame n, which belong to frames n+1, n
>+2,...
>
>Regards,
>
>
>
>

Yes, I tried synchronous and asynchronously to Unblank signal.  There are
two problems.  The first one is the noise itself with the stroke signal
and the second one is the stroke signal being drawn at the second time
doesn't being drawn at the exact position.  It's like one pixel toggling
between frames. 

Do you know which kind of filter fit the best in this situation?

thanks for you suggestion

On Jun 13, 3:09 pm, "cutemonster" <ckh...@hotmail.com> wrote:
> >On Jun 10, 2:38 pm, "cutemonster" <ckh...@hotmail.com> wrote:
> >> >So you want to display a stroke signal on monitor?
> >> >Why you need to sample the X? Is it time dimension? Is it a constant
> >> >ramping, or ramping with retrace, or random?
>
> >> I have to sample x and y because it doesn't work like raster signal.
> It's
> >> voltage varies in time. There is another signal input called
> Unblank(TTL).
> >> It turn on and off of XY signal.  
>
> >Have you tried to lock the sampling clocks to the unblank?
>
> No, I don't understand how to lock it with sampling clock.  Can you please
> explain?
>
> thanks

Locking the clock to the unblank meaning that their phase relation is
fixed.  For some reason I can't escape from the time dimension :)  If
you just sample for 1 frame then probably you don't need to do
locking.  Do you sample just 1 image then display it continuously on
the raster?

If you capture it as live video or multiple frames, then how can you
tell which pixels belong to frame n, which belong to frames n+1, n
+2,...

Regards,

>On Jun 10, 2:38 pm, "cutemonster" <ckh...@hotmail.com> wrote:
>> >So you want to display a stroke signal on monitor?
>> >Why you need to sample the X? Is it time dimension? Is it a constant
>> >ramping, or ramping with retrace, or random?
>>
>> I have to sample x and y because it doesn't work like raster signal. 
It's
>> voltage varies in time. There is another signal input called
Unblank(TTL).
>> It turn on and off of XY signal.  
>
>Have you tried to lock the sampling clocks to the unblank?
>
>

No, I don't understand how to lock it with sampling clock.  Can you please
explain?

thanks

On Jun 10, 2:38 pm, "cutemonster" <ckh...@hotmail.com> wrote:
> >So you want to display a stroke signal on monitor?
> >Why you need to sample the X? Is it time dimension? Is it a constant
> >ramping, or ramping with retrace, or random?
>
> I have to sample x and y because it doesn't work like raster signal.  It's
> voltage varies in time. There is another signal input called Unblank(TTL).
> It turn on and off of XY signal.  

Have you tried to lock the sampling clocks to the unblank?

Thanks for the info, Lakkos.  There are similar projects.  However, they
transfer the logics into fpga instead of digitizing the analog x and y
inputs.  Dealing with analog signal really give me a headache. 

thanks again.

cutemonster wrote:

> 
> I have to sample x and y because it doesn't work like raster signal.  It's
> voltage varies in time. There is another signal input called Unblank(TTL). 
> It turn on and off of XY signal.  
> 

You are talking here about a hardware Vector to Raster Converter.

There is a nice FPGA based Vector to Raster converter
in the Asteroids Deluxe video-game recreation made by
the fellows of FPGA-ARCADE at :

http://home.freeuk.com/fpgaarcade/ast_main.htm

and by Eric Crabill of FPGA-GAMES at :

http://www.fpga-games.com/astdlx.htm


For a general hardware implementation you can also
check the US States Patent and Trademark Office

http://patft.uspto.gov/netahtml/PTO/srchnum.htm

for Patent Nrs 5969699 and 6496160.



The above references might give you an idea of the kind
of digital filters used during this conversion process.




--==============================================
--               Spiros Lakkos
--  http://cybernetic-engineering.blogspot.com/
--==============================================

>So you want to display a stroke signal on monitor?
>Why you need to sample the X? Is it time dimension? Is it a constant
>ramping, or ramping with retrace, or random?
>
>

I have to sample x and y because it doesn't work like raster signal.  It's
voltage varies in time. There is another signal input called Unblank(TTL). 
It turn on and off of XY signal.  

On Jun 10, 4:44 am, "cutemonster" <ckh...@hotmail.com> wrote:
> >"cutemonster" <ckh...@hotmail.com> wrote in message
> >news:RoGdndnLdPZkYvXb4p2dnAA@giganews.com...
>
> >> My question is, do I need a adaptive filter
>
> >Nobody will tell you what you need unless you explain what you are trying
> to
> >do. What is your application? If you are measuring DC why are you
> sampling
> >at 50 MHz?
>
> >/Mikhail
>
> Hi, thanks for the reply. I'm converting a stroke signal to raster signal.
>  There are actually 2 adc.  one for x channel and the other one for y
> channel.  Imagine like a xy graph, depends on the voltage level of each
> channel, it tells where to draw the pixels. So, for 10 bit, it can display
> 1024 x 768 resolution.  The signal runs at around 10mhz.  I don't know if I
> make my question clear enough, please let me know.  I really appreciate
> it.

So you want to display a stroke signal on monitor?
Why you need to sample the X? Is it time dimension? Is it a constant
ramping, or ramping with retrace, or random?