Re: [PATCH ] libv4l: skip false Pixart markers

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hans de Goede wrote:
> Hi,
> 
> On 02/01/2010 11:13 PM, Németh Márton wrote:
>> From: Márton Németh<nm127@xxxxxxxxxxx>
>>
>> The byte sequence 0xff, 0xff, 0xff 0xff is not a real marker to skip, instead
>> it is one byte from the image and the following three 0xff bytes might belong
>> to a real marker. Modify pixart_fill_nbits() macro to pass the first 0xff byte
>> as an image data.
>>
> 
> Oh, good catch. I'm still seeing the occasional bad frame though :(

The same at my side, this patch alone does not solve the whole problem complete.
I have the feeling that at least same of the corrupted frames will not come with
this patch, through I haven't verified this with measurement.

On the other hand, in my previous email used a little bit different code: I jumped
over the 1024 and 512 bytes without parsing it. That would be maybe necessary
to add.

By the way, is there any reason why pixart_fill_nbits() is a macro?

> While on the subject of the pac7302. I've been playing around a bit, and I have the
> feeling that if we were to go for a lower auto gain target (set autogain off and
> lower exposure, you can do this ie with v4l2ucp), combined with a gamma correction of
> 1500 (again use ie v4l2ucp), the images is much better (less over exposed, more
> contrast).
> 
> Do you agree ?

Well, my Labtec Webcam 2200 works only with acceptable indoors, when I try to
capture something outdoors under direct sunshine conditions I get overexposed
frames. I found, however, an interesting pointer in two cameras' user's manual,
see the Note column:

  http://linuxtv.org/wiki/index.php/PixArt_PAC7301/PAC7302#Identification

There is a setting indoor/outdoor which is currently not available in gspca_pac7302
driver. Maybe this would be an interesting point to figure out which register
is related to this setting.

Regards,

	Márton Németh

>> Signed-off-by: Márton Németh<nm127@xxxxxxxxxxx>
>> ---
>> diff -r f23c5a878fb1 v4l2-apps/libv4l/libv4lconvert/tinyjpeg.c
>> --- a/v4l2-apps/libv4l/libv4lconvert/tinyjpeg.c	Mon Feb 01 13:32:46 2010 +0100
>> +++ b/v4l2-apps/libv4l/libv4lconvert/tinyjpeg.c	Mon Feb 01 23:05:39 2010 +0100
>> @@ -339,10 +339,15 @@
>>   	    } \
>>   	    break; \
>>   	  case 0xff: \
>> -	    if (stream[1] == 0xff&&  (stream[2]<  7 || stream[2] == 0xff)) { \
>> -	      stream += 3; \
>> -	      c = *stream++; \
>> -	      break; \
>> +	    if (stream[1] == 0xff) { \
>> +		if (stream[2]<  7) { \
>> +		    stream += 3; \
>> +		    c = *stream++; \
>> +		    break; \
>> +		} else if (stream[2] == 0xff) { \
>> +		    /* four 0xff in a row: the first belongs to the image data */ \
>> +		    break; \
>> +		}\
>>   	    } \
>>   	    /* Error fall through */ \
>>   	  default: \
> 
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[Index of Archives]     [Linux Input]     [Video for Linux]     [Gstreamer Embedded]     [Mplayer Users]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]
  Powered by Linux