At 12:15 PM -0800 12/18/05, richardbrackin wrote:
I’m still in tinkering around mode but I need some clarification:
The manual says “For the purposes of speed and memory use, large images are down-sampled as they are opened.”
What is considered large? Is it a moving target based on the amount of RAM installed on the machine? Is it set in stone to a particular pixel count?
set in stone at the moment. RGB images that are over 500 pixels on their largest side are down-sampled to 500 on that side (and whatever makes sense on the smaller side)
Example: I opened an 8bit TIF file that is from an 8 megapixel digital camera. Storage size is 24,016.5 KB (disk), 975.6KB (ram). 333 x 500 = 166,500 pixels.
right, that looks to be about in line.
When I convert all colors to list, I get 166,500 colors. When I convert unique colors in this particular image, I get 27,991 colors.
OK, sounds good.
When it counts unique colors, does it look at the full size image or does it take this downsampled version and count colors?
… and does it make a difference counting color from the original vs. counting colors from a downsampled version?
at the moment it looks at the down-sampled version. This is mostly for speed.
I am certainly open to other ideas. I am working either a dynamic down-sampling or a user-selectable level (or both). It is plain to see that downsampling is necessary for images above a certain size as it slows the machine to a crawl. ColorThink is typically saving 3 versions of each image (or each image ‘version’ in a workflow) for each one you see on screen. This means it uses many more times the amount of memory than a normal imaging application like Photoshop might (although PS can certainly get up there with layers, etc).
Creating Custom Targets from image data…
This sounds awesome!! I can’t wait to try this one out. I may need some hand holding, though. I have a few questions about this that I am not sure how to ask yet.
The simple flow is this:
- open an image you are having trouble with
- apply the output profile to it if it is not yet in the output color space.
- select the Target Marquee tool and choose “custom”
- set the number of patches you want selected.
- click-drag on the image to select the range of the image you want sampled.
ColorThink will pull the sample points out into a list.
- then select “use as primary list” from the color list’s popup menu and CT will replace the image with a target of the sampled colors.
that’s it! You can then save the color list device values and formulate them into a target using your favorite target software (ProfileMaker or ColorPort most likely) and you can either use it as a target for sampling or to augment your existing profiling targets (Profilemaker supports this but Monaco Profiler doesn’t).
Hope that helps…
This is the very sort of operation that I can build into a Guide for easier access & learning. If you come across things you do a lot that you would like automated let me know and I’ll see if we can create a Guide for them.
o Steve Upton CHROMiX www.chromix.com
Post generated from email list