Re: 16 bit per sample support

Ewald R. de Wit (ewald@pobox.com)
Tue, 6 Apr 1999 19:47:14 +0200

Oliver Rauch (oliver.rauch@Wolfsburg.DE) wrote:
> If the image depth is greater than 8 bits per sample most of the scanners
> (e.g all UMAX, as far as I know microtek too) do send the data padded to
> the most significant bits.

Just as a little scan trivia: the HP format is LSB-justified.

> The only problem I see is that the frontend does not now the real bit depth,
> because
> the backend says 16 bits in all this cases!

Well then lets define 'depth' to be the native depth. If depth>8 then
the frontend should know that one sample is send MSB justified and stretched
(the way Andy described) to 16 bits. How would that sound?

> > Putting it in all the backends would mean more (duplicate) work.
>
> No, no duplicate work, because the backend or the frontend has to do this.

There are more backends then frontends :-) But I agree that was a moot
point.

-- 
  --  Ewald







--
Source code, list archive, and docs: http://www.mostang.com/sane/
To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com