Peter J. Brand wrote:
> "Roy Schestowitz" <newsgroups@schestowitz.com> wrote in message
> news:d2husk$1ujn$1@godfrey.mcc.ac.uk...
>>I am not sure this directly answers your question, but...
>>
>> Peter Brand wrote:
>>> I'm using gimp 2.2.4. I need to scan in 48 bit color images. I'm using
>>> an HP Scanjet 4570c and have the twain capture software set for maximum
>>> bit depth (48 bit color). When I capture the image it is only 24 bit...
>>
>> From what you wrote (or maybe the way it is phrased), it is the scanning
>> software that is unable to scan at 48 bit color.
>>
>
> Both scanner and capture software are 48 bit and set for that.
> Unfortunately, the scanned image is only 24 bit in gimp. I'm not sure if
> it's a communication issue (between the scanner software and gimp) or if
> gimp can only handle 24 bit depth images. I have not been able to get any
> image software it capture a 48 bit depth image yet - I was hopeful that
> gimp would be my answer. Why I want this is answered below.
So let me see if I get this right. You save your images having scanned them
as Bitmaps, PNM's or whetever, still with 16 bits for each channel. Can you
confirm that it practically gets saved in that way? Looking at the file
size should be an easy way to check. Then, once opened in the GIMP, you get
no warning message, but the image that appears before you is reduced to 32
bit? One thing you could do, if crucial, is to slice your data into chunks
of 16 bit channels and deal with each separately. The usefullness of this
probably depends on whether you have inter-channel dependencies in your
task. Do you want to just handle red? Scan the file and get the first 16
bits of each 48 bit chunk; then dump these into a file and treat as
greyscale.
>>> ...Can gimp
>>> handle 48 bit color depth images?
>>
>> I don't think 1.2 can and I assume 2.2.4 can't either (I know 2.2 fairly
>> well, but have it installed elsewhere). What do you need such high colour
>> depth for if I may ask?
>>
>
> I am scanning a special type of film used for radiation measurement
> (gafchromic film made by ISP Corp.
> http://www.ispcorp.com/products/dosimetry/index.html). The film's response
> can best be seen in the data from the red channel. With 48 bit color depth
> I get 16 bits of data in the red channel (blue and green too but they are
> not used). With only 24 bit I only get 8 bits in each channel (254
> levels), which is not enough resolution.
I was faced with a similar problem weeks ago... high-colour-depth TIFF
images of protein Gel that a colleague passed on to me for registration. I
of course asked why it was necessary (I mean, most monitors won't even
render over 16 or 32bit) and his reply was that AstraZeneca wanted high
precision.
I think many graphics libraries (and as in the case I mentioned -- vision
libraries) do not handle datatypes that they 'think' are impractical.
Roy
--
Roy S. Schestowitz
http://Schestowitz.com
|
|
