pwc
[Top] [All Lists]

Re: [pwc] PWC on embedded system

To: Hamed Ahmadi Nejad <>
Subject: Re: [pwc] PWC on embedded system
From: Luc Saillard <>
Date: Tue, 21 Jun 2005 09:58:10 +0200
On Tue, Jun 21, 2005 at 12:08:02AM -0700, Hamed Ahmadi Nejad wrote:

> Ah, thanks!  But one question:  Suppose I use 2 buffers.  The kernel
> is writing to buffer 1, and I am working with buffer 2.  If the image
> processing takes more than the time remaining until the next frame,
> won't the kernel start writing to buffer 2 before the image processing
> is finished? 

No, if the kernel doesn't have any free buffer, the frame is discarded i need
to check but i think this is logical. You can increase the number of frames
used by the kernel at the expense of more memory ... 

> How do you suggest I deal with this, if the image processing is slower than
> the capture rate?  (Since it's just a 300 mhz machine, this could happen
> even at 5 fps.)
 
I've use  this webcam on my celeron 300Mhz, so i think any hardware which can
decode and work on a 640x480  is ok. If you increase the number of
buffer you will not increase your CPU power, because when all buffers will be
fill, you will not have more power to discard old buffer. This is the same
problem when decoding a video, if you want all frames, or only one frame per
two frames.

Using the YUV422p format, you can works only on the Y channel (grayscale) and
you don't have to convert the image in RGB, and you process only 640*480
bytes per frames.

Luc
_______________________________________________
pwc mailing list

http://lists.saillard.org/mailman/listinfo/pwc

<Prev in Thread] Current Thread [Next in Thread>
Admin

Disclaimer: Neither Andrew Taylor nor the University of NSW School of Computer and Engineering take any responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the birding-aus mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU