Re: [libdc1394-devel] 10-bit modes

[Johann S. <j.s...@ir...>]

Damien Douxchamps wrote:
> On Thu, 2007-05-31 at 16:39 +1200, Johann Schoonees wrote:
>>
>>Just to make sure I understand:
>>
>>Encode the image in one of the available encodings: MONO, YUV411, 
>>YUV422, YUV444, or RGB, using 10, 12 or 14 bits per component as 
>>specified by a data depth register.  Then transmit the image by 
>>sending the 8 MSBs of each component as specified in the current 
>>standard for 8-bit data (including any padding), while collecting the 
>>MSBs (2, 4 or 6 bits per component) in a buffer in the same order as 
>>the MSBs.  When the MSB data has been transmitted, transmit the LSB 
>>data as if it is another image, except that the data will be packed 
>>and not correspond byte-wise to an exising format.  Padding bits are 
>>added to the last byte and padding bytes are added to the last packet 
>>as required.
> 
> 
> My idea is slightly different. In any 8bit mode, you can send more than
> 8bit not by sending another frame but by adding the extra bits in the
> padding of the frame itself. Thus:
> 
>  ___________
> |           |
> |           |
> |           |
> |           |
> | 8bit data |
> |   (MSB)   |
> |           |
> |           |
> |___________|
> |           |
> |  Padding  |
> |   (LSB)   |
> |___________|
> 
> The camera simply has to specify a TOTAL_BYTES value that is the sum of
> the two fields.
> 
> So all informations such as framerate etc... are kept identical, but
> those who read the TOTAL_BYTES register and a new register (or new field
> in a current register) will know that the large padding is meaningful.

Oh, in the padding.  That sounds good.

The big advantage I see in this scheme is that it would be possible to 
discard the padding and view or otherwise process the image as 8-bit 
data, using the standard formats.

The MSB padding is a non-standard extension which may eventually 
become well-supported if it is adopted by the 1394TA.

Standards are important. :-)

Johann

This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.